1
|
Yeung C, Ungi T, Hu Z, Jamzad A, Kaufmann M, Walker R, Merchant S, Engel CJ, Jabs D, Rudan J, Mousavi P, Fichtinger G. From quantitative metrics to clinical success: assessing the utility of deep learning for tumor segmentation in breast surgery. Int J Comput Assist Radiol Surg 2024:10.1007/s11548-024-03133-y. [PMID: 38642296 DOI: 10.1007/s11548-024-03133-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/19/2024] [Accepted: 03/28/2024] [Indexed: 04/22/2024]
Abstract
PURPOSE Preventing positive margins is essential for ensuring favorable patient outcomes following breast-conserving surgery (BCS). Deep learning has the potential to enable this by automatically contouring the tumor and guiding resection in real time. However, evaluation of such models with respect to pathology outcomes is necessary for their successful translation into clinical practice. METHODS Sixteen deep learning models based on established architectures in the literature are trained on 7318 ultrasound images from 33 patients. Models are ranked by an expert based on their contours generated from images in our test set. Generated contours from each model are also analyzed using recorded cautery trajectories of five navigated BCS cases to predict margin status. Predicted margins are compared with pathology reports. RESULTS The best-performing model using both quantitative evaluation and our visual ranking framework achieved a mean Dice score of 0.959. Quantitative metrics are positively associated with expert visual rankings. However, the predictive value of generated contours was limited with a sensitivity of 0.750 and a specificity of 0.433 when tested against pathology reports. CONCLUSION We present a clinical evaluation of deep learning models trained for intraoperative tumor segmentation in breast-conserving surgery. We demonstrate that automatic contouring is limited in predicting pathology margins despite achieving high performance on quantitative metrics.
Collapse
Affiliation(s)
- Chris Yeung
- School of Computing, Queen's University, Kingston, ON, Canada.
| | - Tamas Ungi
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Zoe Hu
- School of Medicine, Queen's University, Kingston, ON, Canada
| | - Amoon Jamzad
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Martin Kaufmann
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Ross Walker
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Shaila Merchant
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Cecil Jay Engel
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Doris Jabs
- Department of Radiology, Queen's University, Kingston, ON, Canada
| | - John Rudan
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, ON, Canada
| | | |
Collapse
|
2
|
Connolly L, Fooladgar F, Jamzad A, Kaufmann M, Syeda A, Ren K, Abolmaesumi P, Rudan JF, McKay D, Fichtinger G, Mousavi P. ImSpect: Image-driven self-supervised learning for surgical margin evaluation with mass spectrometry. Int J Comput Assist Radiol Surg 2024:10.1007/s11548-024-03106-1. [PMID: 38600411 DOI: 10.1007/s11548-024-03106-1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/07/2024] [Accepted: 03/08/2024] [Indexed: 04/12/2024]
Abstract
PURPOSE Real-time assessment of surgical margins is critical for favorable outcomes in cancer patients. The iKnife is a mass spectrometry device that has demonstrated potential for margin detection in cancer surgery. Previous studies have shown that using deep learning on iKnife data can facilitate real-time tissue characterization. However, none of the existing literature on the iKnife facilitate the use of publicly available, state-of-the-art pretrained networks or datasets that have been used in computer vision and other domains. METHODS In a new framework we call ImSpect, we convert 1D iKnife data, captured during basal cell carcinoma (BCC) surgery, into 2D images in order to capitalize on state-of-the-art image classification networks. We also use self-supervision to leverage large amounts of unlabeled, intraoperative data to accommodate the data requirements of these networks. RESULTS Through extensive ablation studies, we show that we can surpass previous benchmarks of margin evaluation in BCC surgery using iKnife data, achieving an area under the receiver operating characteristic curve (AUC) of 81%. We also depict the attention maps of the developed DL models to evaluate the biological relevance of the embedding space CONCLUSIONS: We propose a new method for characterizing tissue at the surgical margins, using mass spectrometry data from cancer surgery.
Collapse
Affiliation(s)
| | | | | | | | | | - Kevin Ren
- Queen's University, Kingston, ON, Canada
| | | | | | - Doug McKay
- Queen's University, Kingston, ON, Canada
| | | | | |
Collapse
|
3
|
To MNN, Fooladgar F, Wilson P, Harmanani M, Gilany M, Sojoudi S, Jamzad A, Chang S, Black P, Mousavi P, Abolmaesumi P. LensePro: label noise-tolerant prototype-based network for improving cancer detection in prostate ultrasound with limited annotations. Int J Comput Assist Radiol Surg 2024:10.1007/s11548-024-03104-3. [PMID: 38598142 DOI: 10.1007/s11548-024-03104-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/03/2024] [Accepted: 03/04/2024] [Indexed: 04/11/2024]
Abstract
PURPOSE The standard of care for prostate cancer (PCa) diagnosis is the histopathological analysis of tissue samples obtained via transrectal ultrasound (TRUS) guided biopsy. Models built with deep neural networks (DNNs) hold the potential for direct PCa detection from TRUS, which allows targeted biopsy and subsequently enhances outcomes. Yet, there are ongoing challenges with training robust models, stemming from issues such as noisy labels, out-of-distribution (OOD) data, and limited labeled data. METHODS This study presents LensePro, a unified method that not only excels in label efficiency but also demonstrates robustness against label noise and OOD data. LensePro comprises two key stages: first, self-supervised learning to extract high-quality feature representations from abundant unlabeled TRUS data and, second, label noise-tolerant prototype-based learning to classify the extracted features. RESULTS Using data from 124 patients who underwent systematic prostate biopsy, LensePro achieves an AUROC, sensitivity, and specificity of 77.9%, 85.9%, and 57.5%, respectively, for detecting PCa in ultrasound. Our model shows it is effective for detecting OOD data in test time, critical for clinical deployment. Ablation studies demonstrate that each component of our method improves PCa detection by addressing one of the three challenges, reinforcing the benefits of a unified approach. CONCLUSION Through comprehensive experiments, LensePro demonstrates its state-of-the-art performance for TRUS-based PCa detection. Although further research is necessary to confirm its clinical applicability, LensePro marks a notable advancement in enhancing automated computer-aided systems for detecting prostate cancer in ultrasound.
Collapse
Affiliation(s)
- Minh Nguyen Nhat To
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada.
| | - Fahimeh Fooladgar
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Paul Wilson
- School of Computing, Queen's University, Kingston, Canada
| | | | - Mahdi Gilany
- School of Computing, Queen's University, Kingston, Canada
| | - Samira Sojoudi
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Amoon Jamzad
- School of Computing, Queen's University, Kingston, Canada
| | - Silvia Chang
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Peter Black
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, Canada.
| | - Purang Abolmaesumi
- Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada.
| |
Collapse
|
4
|
Groves LA, Keita M, Talla S, Kikinis R, Fichtinger G, Mousavi P, Camara M. A Review of Low-Cost Ultrasound Compatible Phantoms. IEEE Trans Biomed Eng 2023; 70:3436-3448. [PMID: 37339047 DOI: 10.1109/tbme.2023.3288071] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 06/22/2023]
Abstract
Ultrasound-compatible phantoms are used to develop novel US-based systems and train simulated medical interventions. The price difference between lab-made and commercially available ultrasound-compatible phantoms lead to the publication of many papers categorized as low-cost in the literature. The aim of this review was to improve the phantom selection process by summarizing the pertinent literature. We compiled papers on US-compatible spine, prostate, vascular, breast, kidney, and li ver phantoms. We reviewed papers for cost and accessibility, providing an overview of the materials, construction time, shelf life, needle insertion limits, and manufacturing and evaluation methods. This information was summarized by anatomy. The clinical application associated with each phantom was also reported for those interested in a particular intervention. Techniques and common practices for building low-cost phantoms were provided. Overall, this article aims to summarize a breadth of ultrasound-compatible phantom research to enable informed phantom methods selection.
Collapse
|
5
|
Tomalty D, Giovannetti O, Velikonja L, Munday J, Kaufmann M, Iaboni N, Jamzad A, Rubino R, Fichtinger G, Mousavi P, Nicol CJB, Rudan JF, Adams MA. Molecular characterization of human peripheral nerves using desorption electrospray ionization mass spectrometry imaging. J Anat 2023; 243:758-769. [PMID: 37264225 PMCID: PMC10557387 DOI: 10.1111/joa.13909] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/15/2023] [Revised: 05/11/2023] [Accepted: 05/20/2023] [Indexed: 06/03/2023] Open
Abstract
Desorption electrospray ionization mass spectrometry imaging (DESI-MSI) is a molecular imaging method that can be used to elucidate the small-molecule composition of tissues and map their spatial information using two-dimensional ion images. This technique has been used to investigate the molecular profiles of variety of tissues, including within the central nervous system, specifically the brain and spinal cord. To our knowledge, this technique has yet to be applied to tissues of the peripheral nervous system (PNS). Data generated from such analyses are expected to advance the characterization of these structures. The study aimed to: (i) establish whether DESI-MSI can discriminate the molecular characteristics of peripheral nerves and distinguish them from surrounding tissues and (ii) assess whether different peripheral nerve subtypes are characterized by unique molecular profiles. Four different nerves for which are known to carry various nerve fiber types were harvested from a fresh cadaveric donor: mixed, motor and sensory (sciatic and femoral); cutaneous, sensory (sural); and autonomic (vagus). Tissue samples were harvested to include the nerve bundles in addition to surrounding connective tissue. Samples were flash-frozen, embedded in optimal cutting temperature compound in cross-section, and sectioned at 14 μm. Following DESI-MSI analysis, identical tissue sections were stained with hematoxylin and eosin. In this proof-of-concept study, a combination of multivariate and univariate statistical methods was used to evaluate molecular differences between the nerve and adjacent tissue and between nerve subtypes. The acquired mass spectral profiles of the peripheral nerve samples presented trends in ion abundances that seemed to be characteristic of nerve tissue and spatially corresponded to the associated histology of the tissue sections. Principal component analysis (PCA) supported the separation of the samples into distinct nerve and adjacent tissue classes. This classification was further supported by the K-means clustering analysis, which showed separation of the nerve and background ions. Differences in ion expression were confirmed using ANOVA which identified statistically significant differences in ion expression between the nerve subtypes. The PCA plot suggested some separation of the nerve subtypes into four classes which corresponded with the nerve types. This was supported by the K-means clustering. Some overlap in classes was noted in these two clustering analyses. This study provides emerging evidence that DESI-MSI is an effective tool for metabolomic profiling of peripheral nerves. Our results suggest that peripheral nerves have molecular profiles that are distinct from the surrounding connective tissues and that DESI-MSI may be able to discriminate between nerve subtypes. DESI-MSI of peripheral nerves may be a valuable technique that could be used to improve our understanding of peripheral nerve anatomy and physiology. The ability to utilize ambient mass spectrometry techniques in real time could also provide an unprecedented advantage for surgical decision making, including in nerve-sparing procedures in the future.
Collapse
Affiliation(s)
- Diane Tomalty
- Department of Biomedical and Molecular SciencesQueen's UniversityKingstonOntarioCanada
| | - Olivia Giovannetti
- Department of Biomedical and Molecular SciencesQueen's UniversityKingstonOntarioCanada
| | - Leah Velikonja
- Department of Biomedical and Molecular SciencesQueen's UniversityKingstonOntarioCanada
| | - Jasica Munday
- Department of Biomedical and Molecular SciencesQueen's UniversityKingstonOntarioCanada
| | - Martin Kaufmann
- Department of SurgeryQueen's UniversityKingstonOntarioCanada
- Gastrointestinal Diseases Research UnitKingston Health Sciences CenterKingstonOntarioCanada
| | - Natasha Iaboni
- Department of Pathology and Molecular MedicineQueen's UniversityKingstonOntarioCanada
| | - Amoon Jamzad
- School of ComputingQueen's UniversityKingstonOntarioCanada
| | - Rachel Rubino
- Division of Cancer Biology and GeneticsQueen's Cancer Research InstituteKingstonOntarioCanada
| | | | - Parvin Mousavi
- School of ComputingQueen's UniversityKingstonOntarioCanada
| | - Christopher J. B. Nicol
- Department of Pathology and Molecular MedicineQueen's UniversityKingstonOntarioCanada
- Division of Cancer Biology and GeneticsQueen's Cancer Research InstituteKingstonOntarioCanada
| | - John F. Rudan
- Department of SurgeryQueen's UniversityKingstonOntarioCanada
| | - Michael A. Adams
- Department of Biomedical and Molecular SciencesQueen's UniversityKingstonOntarioCanada
| |
Collapse
|
6
|
Wilson PFR, Gilany M, Jamzad A, Fooladgar F, To MNN, Wodlinger B, Abolmaesumi P, Mousavi P. Self-Supervised Learning With Limited Labeled Data for Prostate Cancer Detection in High-Frequency Ultrasound. IEEE Trans Ultrason Ferroelectr Freq Control 2023; 70:1073-1083. [PMID: 37478033 DOI: 10.1109/tuffc.2023.3297840] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 07/23/2023]
Abstract
Deep learning-based analysis of high-frequency, high-resolution micro-ultrasound data shows great promise for prostate cancer (PCa) detection. Previous approaches to analysis of ultrasound data largely follow a supervised learning (SL) paradigm. Ground truth labels for ultrasound images used for training deep networks often include coarse annotations generated from the histopathological analysis of tissue samples obtained via biopsy. This creates inherent limitations on the availability and quality of labeled data, posing major challenges to the success of SL methods. However, unlabeled prostate ultrasound data are more abundant. In this work, we successfully apply self-supervised representation learning to micro-ultrasound data. Using ultrasound data from 1028 biopsy cores of 391 subjects obtained in two clinical centers, we demonstrate that feature representations learned with this method can be used to classify cancer from noncancer tissue, obtaining an AUROC score of 91% on an independent test set. To the best of our knowledge, this is the first successful end-to-end self-SL (SSL) approach for PCa detection using ultrasound data. Our method outperforms baseline SL approaches, generalizes well between different data centers, and scales well in performance as more unlabeled data are added, making it a promising approach for future research using large volumes of unlabeled data. Our code is publicly available at https://www.github.com/MahdiGilany/SSL_micro_ultrasound.
Collapse
|
7
|
Gilany M, Wilson P, Perera-Ortega A, Jamzad A, To MNN, Fooladgar F, Wodlinger B, Abolmaesumi P, Mousavi P. TRUSformer: improving prostate cancer detection from micro-ultrasound using attention and self-supervision. Int J Comput Assist Radiol Surg 2023:10.1007/s11548-023-02949-4. [PMID: 37217768 DOI: 10.1007/s11548-023-02949-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/15/2023] [Accepted: 05/02/2023] [Indexed: 05/24/2023]
Abstract
PURPOSE A large body of previous machine learning methods for ultrasound-based prostate cancer detection classify small regions of interest (ROIs) of ultrasound signals that lie within a larger needle trace corresponding to a prostate tissue biopsy (called biopsy core). These ROI-scale models suffer from weak labeling as histopathology results available for biopsy cores only approximate the distribution of cancer in the ROIs. ROI-scale models do not take advantage of contextual information that are normally considered by pathologists, i.e., they do not consider information about surrounding tissue and larger-scale trends when identifying cancer. We aim to improve cancer detection by taking a multi-scale, i.e., ROI-scale and biopsy core-scale, approach. METHODS Our multi-scale approach combines (i) an "ROI-scale" model trained using self-supervised learning to extract features from small ROIs and (ii) a "core-scale" transformer model that processes a collection of extracted features from multiple ROIs in the needle trace region to predict the tissue type of the corresponding core. Attention maps, as a by-product, allow us to localize cancer at the ROI scale. RESULTS We analyze this method using a dataset of micro-ultrasound acquired from 578 patients who underwent prostate biopsy, and compare our model to baseline models and other large-scale studies in the literature. Our model shows consistent and substantial performance improvements compared to ROI-scale-only models. It achieves [Formula: see text] AUROC, a statistically significant improvement over ROI-scale classification. We also compare our method to large studies on prostate cancer detection, using other imaging modalities. CONCLUSIONS Taking a multi-scale approach that leverages contextual information improves prostate cancer detection compared to ROI-scale-only models. The proposed model achieves a statistically significant improvement in performance and outperforms other large-scale studies in the literature. Our code is publicly available at www.github.com/med-i-lab/TRUSFormer .
Collapse
Affiliation(s)
- Mahdi Gilany
- School of Computing, Queen's University, Kingston, Canada.
| | - Paul Wilson
- School of Computing, Queen's University, Kingston, Canada
| | | | - Amoon Jamzad
- School of Computing, Queen's University, Kingston, Canada
| | - Minh Nguyen Nhat To
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Fahimeh Fooladgar
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | | | - Purang Abolmaesumi
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, Canada
| |
Collapse
|
8
|
Kaufmann M, Iaboni N, Jamzad A, Hurlbut D, Ren KYM, Rudan JF, Mousavi P, Fichtinger G, Varma S, Caycedo-Marulanda A, Nicol CJB. Metabolically Active Zones Involving Fatty Acid Elongation Delineated by DESI-MSI Correlate with Pathological and Prognostic Features of Colorectal Cancer. Metabolites 2023; 13:metabo13040508. [PMID: 37110166 PMCID: PMC10141897 DOI: 10.3390/metabo13040508] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/25/2023] [Revised: 03/18/2023] [Accepted: 03/20/2023] [Indexed: 04/03/2023] Open
Abstract
Colorectal cancer (CRC) is the second leading cause of cancer deaths. Despite recent advances, five-year survival rates remain largely unchanged. Desorption electrospray ionization mass spectrometry imaging (DESI) is an emerging nondestructive metabolomics-based method that retains the spatial orientation of small-molecule profiles on tissue sections, which may be validated by ‘gold standard’ histopathology. In this study, CRC samples were analyzed by DESI from 10 patients undergoing surgery at Kingston Health Sciences Center. The spatial correlation of the mass spectral profiles was compared with histopathological annotations and prognostic biomarkers. Fresh frozen sections of representative colorectal cross sections and simulated endoscopic biopsy samples containing tumour and non-neoplastic mucosa for each patient were generated and analyzed by DESI in a blinded fashion. Sections were then hematoxylin and eosin (H and E) stained, annotated by two independent pathologists, and analyzed. Using PCA/LDA-based models, DESI profiles of the cross sections and biopsies achieved 97% and 75% accuracies in identifying the presence of adenocarcinoma, using leave-one-patient-out cross validation. Among the m/z ratios exhibiting the greatest differential abundance in adenocarcinoma were a series of eight long-chain or very-long-chain fatty acids, consistent with molecular and targeted metabolomics indicators of de novo lipogenesis in CRC tissue. Sample stratification based on the presence of lympovascular invasion (LVI), a poor CRC prognostic indicator, revealed the abundance of oxidized phospholipids, suggestive of pro-apoptotic mechanisms, was increased in LVI-negative compared to LVI-positive patients. This study provides evidence of the potential clinical utility of spatially-resolved DESI profiles to enhance the information available to clinicians for CRC diagnosis and prognosis.
Collapse
Affiliation(s)
- Martin Kaufmann
- Department of Surgery, Kingston Health Sciences Centre, Kingston, ON K7L 2V7, Canada
- Gastrointestinal Diseases Research Unit, Kingston Health Sciences Center, Kingston, ON K7L 2V7, Canada
| | - Natasha Iaboni
- Department of Pathology and Molecular Medicine, Queen’s University and Kingston Health Sciences Centre, Kingston, ON K7L 3N6, Canada
| | - Amoon Jamzad
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada
| | - David Hurlbut
- Department of Pathology and Molecular Medicine, Queen’s University and Kingston Health Sciences Centre, Kingston, ON K7L 3N6, Canada
| | - Kevin Yi Mi Ren
- Department of Pathology and Molecular Medicine, Queen’s University and Kingston Health Sciences Centre, Kingston, ON K7L 3N6, Canada
| | - John F. Rudan
- Department of Surgery, Kingston Health Sciences Centre, Kingston, ON K7L 2V7, Canada
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada
| | - Gabor Fichtinger
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada
| | - Sonal Varma
- Department of Pathology and Molecular Medicine, Queen’s University and Kingston Health Sciences Centre, Kingston, ON K7L 3N6, Canada
| | - Antonio Caycedo-Marulanda
- Department of Surgery, Kingston Health Sciences Centre, Kingston, ON K7L 2V7, Canada
- Orlando Health Colon and Rectal Institute, Orlando, FL 32806, USA
| | - Christopher J. B. Nicol
- Department of Pathology and Molecular Medicine, Queen’s University and Kingston Health Sciences Centre, Kingston, ON K7L 3N6, Canada
- Queen’s Cancer Research Institute, Division of Cancer Biology and Genetics, Kingston, ON K7L 3N6, Canada
| |
Collapse
|
9
|
Chen B, Maslove DM, Curran JD, Hamilton A, Laird PR, Mousavi P, Sibley S. A deep learning model for the classification of atrial fibrillation in critically ill patients. Intensive Care Med Exp 2023; 11:2. [PMID: 36635373 PMCID: PMC9837355 DOI: 10.1186/s40635-022-00490-3] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/11/2022] [Accepted: 12/27/2022] [Indexed: 01/14/2023] Open
Abstract
BACKGROUND Atrial fibrillation (AF) is the most common cardiac arrhythmia in the intensive care unit and is associated with increased morbidity and mortality. New-onset atrial fibrillation (NOAF) is often initially paroxysmal and fleeting, making it difficult to diagnose, and therefore difficult to understand the true burden of disease. Automated algorithms to detect AF in the ICU have been advocated as a means to better quantify its true burden. RESULTS We used a publicly available 12-lead ECG dataset to train a deep learning model for the classification of AF. We then conducted an external independent validation of the model using continuous telemetry data from 984 critically ill patients collected in our institutional database. Performance metrics were stratified by signal quality, classified as either clean or noisy. The deep learning model was able to classify AF with an overall sensitivity of 84%, specificity of 89%, positive predictive value (PPV) of 55%, and negative predictive value of 97%. Performance was improved in clean data as compared to noisy data, most notably with respect to PPV and specificity. CONCLUSIONS This model demonstrates that computational detection of AF is currently feasible and effective. This approach stands to improve the efficiency of retrospective and prospective research into AF in the ICU by automating AF detection, and enabling precise quantification of overall AF burden.
Collapse
Affiliation(s)
- Brian Chen
- grid.410356.50000 0004 1936 8331School of Computing, Queen’s University, Kingston, Canada
| | - David M. Maslove
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, 76 Stuart Street, Kingston, ON K7L 2V7 Canada
| | - Jeffrey D. Curran
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, 76 Stuart Street, Kingston, ON K7L 2V7 Canada
| | - Alexander Hamilton
- grid.410356.50000 0004 1936 8331Centre for Health Innovation, Queen’s University, Kingston, Canada
| | - Philip R. Laird
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, 76 Stuart Street, Kingston, ON K7L 2V7 Canada
| | - Parvin Mousavi
- grid.410356.50000 0004 1936 8331School of Computing, Queen’s University, Kingston, Canada
| | - Stephanie Sibley
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, 76 Stuart Street, Kingston, ON K7L 2V7 Canada
| |
Collapse
|
10
|
Fooladgar F, Nguyen Nhat to M, Javadi G, Sojoudi S, Eshumani W, Chang S, Black P, Mousavi P, Abolmaesumi P. Semi-supervised learning from coarse histopathology labels. Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization 2022. [DOI: 10.1080/21681163.2022.2154275] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 12/23/2022]
Affiliation(s)
- Fahimeh Fooladgar
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada
| | - Minh Nguyen Nhat to
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada
| | - Golara Javadi
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada
| | - Samira Sojoudi
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada
| | | | - Silvia Chang
- Vancouver General Hospital, Vancouver, BC, Canada
| | - Peter Black
- Vancouver General Hospital, Vancouver, BC, Canada
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON, Canada
| | - Purang Abolmaesumi
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada
| |
Collapse
|
11
|
Selim Y, Di Lena É, Abu-Omar N, Baig Z, Verhoeff K, La J, Purich K, Albacete S, Valji R, Purich K, Safar A, Schellenberg M, Schellenberg M, Schellenberg M, Schellenberg M, Schellenberg M, Schellenberg M, Daza J, Glass LT, Verhoeff K, Johnson G, Guidolin K, Glass LT, Balvardi S, Gawad N, McKechnie T, McKechnie T, Purich K, Henley J, Imbert E, Li C, Skinner S, Lenet T, Lenet T, Metz J, Ahn H(S, Do U, Rouhi A, Greenberg B, Muaddi H, Park L, Vogt K, Bradley N, Deng SX, Murphy P, Alhabboubi M, Lie J, Laplante S, Lie J, Drung J, Nixon T, Allard-Coutu A, Mansouri S, Lee A, Tweedy J, D’Elia MA, Hopkins B, Srivastava A, Alibhai K, Lee C, Moon J(J, How N, Spoyalo K, Lalande A, Baig Z, Schweitzer C, Keogh J, Huo B, Patel YS, Patel YS, Jogiat U, McGuire AL, Jogiat U, Lee Y, Barber E, Akhtar-Danesh GG, Bondzi-Simpson A, Bowker R, Ahmadi N, Abdul SA, Patel P, Harrison L, Shi G, Shi G, Alaichi JA, Kidane B, Qu LC, Alaichi J, Mackay E, Lee J, Purich K, Castelo M, Caycedo-Marulanda A, Caycedo-Marulanda A, Brennan K, Brennan K, Brennan K, Garfinkle R, Sharma S, Candy S, Patel S, LeGal G, Spadafora S, Maclellan S, Trottier D, Jonker D, Asmis T, Mallick R, Ramsay T, Carrier M, McKechnie T, Shojaei D, Motamedi A, Ghuman A, Karimuddin A, Raval M, Brown C, Shojaei D, Wang H, Buie D, Wang H, McKechnie T, Buie D, Al Busaidi N, Rajabiyazdi F, Demian M, Boutros M, Farooq A, Brown C, Phang T, Ghuman A, Karimuddin A, McKechnie T, Raval M, Udwadia F, Marinescu D, Alqahtani M, Pang A, Vasilevsky CA, Boutros M, Oviedo SC, McFadden N, Spence R, Lee L, Hirsch G, Neumann K, Neumann K, Spence R, Johnson G, Singh H, Helewa R, Yilbas A, Netto FS, Katz J, Robitaille S, Sharma B, Khan U, Selzner M, Mocanu V, Dang J, Wilson H, Switzer N, Birch D, Karmali S, Mocanu V, Robitaille S, Jogiat U, Forbes H, Switzer N, Birch D, Karmali S, Verhoeff K, Mocanu V, Kung J, Purich K, Switzer N, Sadri H, Birch D, Karmali S, Tassé N, Tchernof A, Nadeau M, Dawe P, Beckett A, Biertho L, Lin A, Verhoeff K, Selznick S, Mocanu V, Kung JY, Birch DW, Karmali S, Switzer NJ, Fowler-Woods M, Fowler-Woods A, Shingoose G, Hatala A, Daeninck F, Wiseman V, Vergis A, Hardy K, Clouston K, Debru E, Sun W, Dang J, Switzer N, Birch D, Karmali S, De Gara C, Wiseman V, Halasz J, Dang J, Switzer N, Kanji A, Birch D, Modi R, Karmali S, Gu J, Jarrar A, Kolozsvari N, Wiseman V, Samarasinghe Y, Chen L, Hapugall A, Javidan A, McKechnie T, Doumouras A, Hong D, Laplante S, Stogryn S, Maeda A, Brennan K, Jackson T, Okrainec A, Birch D, Karmali S, Kanji A, Switzer N, Balas M, Gee D, Hutter M, Meireles O, Baker L, Jung J, Vergis A, Hardy K, Boudreau V, Hong D, Anvari M, Iranmanesh P, Barlow K, Cookson T, Bolis R, Ichhpuniani S, Shanthanna H, Shiroky J, Deghan S, Zevin B, Cloutier Z, Cookson T, Barlow K, Boudreau V, Anvari M, Brodie J, Johnson G, O’Brien E, Tedman-Aucoin K, Lawlor D, Murphy R, Twells L, Pace D, Ellsmere J, Evans B, Zhang T, Deehan E, Zhang L, Kao D, Hotte N, Birch D, Karmali S, Samarasinghe K, Walter J, Madsen K, Williams E, Kong W, Fundytus A, Holden J, Booth C, Patel S, Caycedo-Marulanda A, Chung W, Nanji S, Merchant S, Hansen B, Paszat L, Baxter N, Scheer A, Moon J(J, Ghezeljeh TN, Lapointe-Gagner M, Nguyen-Powanda P, Elhaj H, Rajabiyazdi F, Lee L, Feldman LS, Fiore J, Sutradhar R, Haas B, Moon J(J, Guttman M, Eskander A, Coburn N, Chesney T, Tillman B, Zuk V, Mahar A, Hsu A, Chan W, Hallet J, Moon J(J, Guidolin K, Servidio-Italiano F, Quereshy F, Sapisochin G, Prisman E, Mitmaker E, Walker R, Wu J, Nguyen A, Wiseman S, Moon J(J, Hong D, Gupta M, Pasternak J, Vergara-Fernandez O, Brar A, Conner J, Kirsch R, Brar M, Kennedy E, Govindarajan A, Paradis T, Gryfe R, MacRae H, Cohen Z, McLeod R, Swallow C, Ghasemi F, Delisle M, Stacey D, Abou-Khalil J, Balaa F, Bayat Z, Bertens K, Dingley B, Martel G, McAlpine K, Nessim C, Tadros S, Carrier M, Auer R, Lim D, Eskander A, Patel S, Coburn N, Sutradhar R, Chan W, Hallet J, Pickard F, Mathieson A, Hogan M, Seal M, Acidi B, Johnston W, Maalouf MF, Callegaro D, Brar S, Gladdy R, Chung P, Catton C, Khalili K, Honore C, Swallow C, Fu N, Kazazian K, Wachtel A, Pacholczyk K, Ng D, Swett-Cosentino J, Savage P, Shibahara Y, Kalimuthu S, Espin-Garcia O, Conner J, Yeung J, Darling G, Baronas V, Swallow C, Mercier F, Barabash V, Law C, Coburn N, Singh S, Myrehaug S, Chan W, Hallet J, Auer RC, Wiseman V, Muhn N, Heller B, Lovrics P, Ng D, Koerber D, Al-Sukhni E, Cyr D, Kazazian K, Swallow C, Apte S, Lie J, Nessim C, Ban J, Chen L, McKevitt E, Warburton R, Pao JS, Dingee C, Kuusk U, Bazzarelli A, Hong NL, Hajjar R, Gagliardi A, Nyhof B, Marfil-Garza B, Sandha G, Cooper D, Dajani K, Bigam D, Anderson B, Kin T, Lam A, Ehlebracht A, O’Gorman D, Senior P, Ricordi C, Shapiro AJ, Moser M, Lam H, Flemming J, Jalink D, Nanji S, Bennett S, Caminsky N, Zhang B, Tywonek K, Meyers B, Serrano P, Lenet T, Shorr R, Abou-Khalil J, Bertens K, Balaa F, Martel G, Caminsky N, Jayaraman S, Wei A, Mahar A, Kaliwal Y, Martel G, Coburn N, Hallet J, Williams P, Claasen M, Ivanics T, Gilbert R, Englesakis M, Gallinger S, Hansen B, Sapisochin G, Ivanics T, Claasen M, Gallinger S, Hansen B, Sapisochin G, Lenet T, Morin G, Abou-Khalil J, Balaa F, Martel G, Brind’Amour A, Bertens K, Balaa F, Bertens K, Martel G, Abou-Khalil J, Collin Y, Auer RC, Ivanics T, Toso C, Adam R, Ijzermans J, Sapisochin G, Polak W, Léveillé M, Lawson C, Collin Y, Tai LH, Phang T, Greene B, Jayaraman S, Tsang M, Al-Arnawoot A, Rajendran L, Lamb T, Turner A, Reid M, Rekman J, Mimeault R, Hopkins J, Abou-Khalil J, Bertens K, Martel G, Balaa F, Zhang C, Lemke M, Glinka J, Leslie K, Skaro A, Tang E, Hopkins J, Greene B, Tsang M, Jayaraman S, Bubis L, Jayaraman S, Tsang ME, Ganescu O, Vanounou T, Pelletier JS, Greene B, Levin Y, Tsang M, Jayaraman S, Ganescu O, Pelletier JS, Vanounou T, Choi WJ, Muaddi H, Ivanics T, Classen MP, Sapisochin G, Alam A, Caminsky N, Mansouri S, Lagace P, Lagace P, El-Kefraoui C, Mainprize M, Melland-Smith M, Verhoeff K, Verhoeff K, Nasser K, Mailloux O, Purich K, Whyte M, Li T, Ahmad MS, Sun W, Ahn H(S, Lee Y, Roach E, Chow A, Trac J, He W, Ramji K, Kouzmina E, Koziak C, Hossain I, Mocanu V, Hanna N, Castelo M, Pook M, Zuckerman J, Choi WJ, Watanabe A, Saravana-Bawan B, Cyr D, Brackstone M, Ivankovic V, Nair AG, Hirpara D, Stockley C, Ng D, Luu S, Meloche-Dumas L, D’Elia MA, Eom A, Tang K, Khan S, Schmitz E, Chen KT, Newman-Bremang J, Verhoeff K, Jette N, Mir Z, Griffiths C, Rajendran L, Zuckerman J, Choi WJ, Choi WJ, Gilbert R, Lenet T, Amhis N, Claasen M, Mansouri S, Workewych A, Lee A, Waugh E, Zhu A, Nabavian H, Roldan J, Lagrotteria A, Roldan J, Rajendran L, Safa N, Rahman S, Kaneva P, Feldman L, Baig Z, Ginther N, Gill D, Sarwar Z, Verdiales C, Moser M, Mocanu V, Fang B, Dang J, Sun W, Switzer N, Birch D, Karmali S, Alqaydi A, Wei X, Digby G, Brogly S, Merchant S, Verhoeff K, Miles A, Kung JY, Shapiro AJ, Bigam DL, Matkin A, Dumestre D, Peiris L, Turner S, Verhoeff K, Mador B, McLennan S, Jastaniah A, Owattanapanich N, Grigorian A, Lam L, Nahmias J, Inaba K, Liasidis P, Inaba K, Demetriades D, Benjamin ER, Cowan S, Owattanapanich N, Wong MD, Inaba K, Demetriades D, Owattanapanich N, Emigh B, Karavites L, Clark DH, Lam L, Inaba K, Benjamin ER, Owattanapanich N, Inaba K, Demetriades D, Anderson GA, Owattanapanich N, Cheng V, Lam L, Inaba K, Myers L, Cuthbertson B, Myles P, Shulman M, Wijeysundera D, Murphy PB, Allen L, Minkhorst K, Bowker D, Tang ES, Leslie K, Hawel JE, Ma O, Purich K, Skinner S, Dhaliwal R, Strickland M, Park J, Vergis A, Gillman L, Rivard J, Lin J, Zorigtbaatar A, Nadeem M, Ibrahim T, Neilson Z, Kim KY(P, Rajendran L, Chadi S, Quereshy F, Davidson M, Friedrich E, Champion C, Semsar-Kazerooni K, Kaneva P, Mueller C, Vassiliou M, Al Mahroos M, Fiore J, Schwartzman K, Feldman L, Purich K, Verhoeff K, Anderson B, Daniel R, Kruse C, Levin M, Lee Y, Doumouras A, Hong D, Eskicioglu C, Povolo C, Lee J, Lee Y, Park L, Doumouras A, Hong D, Bhandari M, Eskicioglu C, Lee E, Verhoeff K, Sydora D, Turner S, Strickland M, Boone D, Roberts S, McGrouther D, Manuel P, Dykstra M, Wang H, Snelgrove R, Verhoeff K, Purich K, Perry T, Strickland M, Dhaliwal R, Skanes S, Tropiano J, McIsaac D, Tinmouth A, Hallet J, Nicholls S, Fergusson D, Martel G, Tropiano J, Skanes S, Ivankovic V, McIsaac D, Tinmouth A, Patey A, Fergusson D, Martel G, Naqvi R, Noppens R, Hawel J, Elnahas A, Schlachta C, Alkhamesi N, Lenet T, Gilbert R, Mallick R, Shaw J, McIsaac D, Martel G, Pook M, Najafi T, Rajabiyazdi F, El-Kefraoui C, Balvardi S, Barone N, Elhaj H, Nguyen-Powanda P, Lee L, Baldini G, Feldman L, Fiore J, Purich K, Jogiat U, Mapiour D, Kim M, Nadler A, Stukel T, De Mestral C, Nathens A, Pautler S, Shayegan B, Hanna W, Schlachta C, Breau R, Hopkins L, Jackson T, Karanicolas P, Griffiths C, Ali S, Archer V, Cloutier Z, Choi D, McKechnie T, Serrano P, McClure JA, Jones P, Mrkobrada M, Flier S, Welk B, Dubois L, Khwaja K, Allen L, Tung L, Hameed M, Spoyalo K, Lampron J, Garcia-Ochoa C, Jastaniah A, Engels P, Talwar G, Cloutier Z, Minor S, Moffat B, Parry N, Vogt K, Ebeye T, Sharma BT, Samman A, Zulfiqar A, Purzner RH, Greene B, Tsang M, Jayaraman S, Allen L, Ball CG, Hameed M, Engels P, Nenshi R, Widder S, Minor S, Parry N, Bradley N, Moffat B, Vogt K, Liberman S, Wong E, Chughtai T, Turner J, Wiseman S, Li J, Mak N, Namazi B, Kiani P, Hashimoto D, Alseidi A, Pasten M, Brunt LM, Gill S, Davis B, Bloom M, Pernar L, Okrainec A, Madani A, Hamilton T, Nabata K, Zhang JW, Zhao D, Dawe P, Hameed M, Allen L, Leslie K, Drung J, Purich K, Verhoeff K, Skinner S, Dhaliwal R, Strickland M, Heller B, Wright V, Cho W, Wong M, Singh K, Comeau É, Collin Y, Oviedo SC, Rajendran L, Lamb T, Reid M, Turner A, Bertens K, Rekman J, Balaa F, Neumann K, Porter G, Teo I, Forse CL, Musselman R, Grozman L, Alshawan N, Fraser S, Bergman S, Pelletier JS, Vanounou T, Wong EG, Nolan B, Qiao L, Jung J, Raîche I, McDonald H, Gawad N, Cheifetz R, Hawley P, Rehany Z, Bakhshi M, Razek T, Grushka J, Boulanger N, Watt L, Vivier A, Clark G, Fata P, Deckelbaum D, Khwaja K, Jastaniah A, Wong E, Ren K, Qiu Y, Hamann K, Lalande A, MacNeill A, Park J, Alexis S, Spoyalo K, Patterson K, Gadhari N, Zhao J, MacNeill A, Ahmed R, Farooq A, Karimuddin A, Wiseman S, Chhor A, Begum H, Akhtar-Danesh N, Finley C, Manos D, Xu Z, Matheson K, Chun S, Fris J, Wallace A, French D, Baste JM, Shargall Y, Waddell TK, Yasufuku K, Machuca TN, Xie F, Thabane L, Hanna WC, Beauchamp MK, Wald J, Mbuagbaw L, Key BL, Green SM, Hanna WC, Behzadi A, Donahoe L, Nasralla A, Poon J, Safieddine N, Seyednejad N, Tavakoli I, Turner S, Sharma S, Hilzenrat RA, McConechy M, Frank I, Hughsman C, Yip S, Choi JJ, Yee J, Sasewich H, Turner S, Baracos V, Eurich D, Filafilo H, Bedard E, Samarasinghe Y, Patel J, Khondker A, McKechnie T, Samarasinghe N, Finley C, Hanna W, Shargall Y, Agzarian J, Whidden M, Aguirre F, Graham A, Chhor A, Begum H, Ewusie J, Mbuagbaw L, Yaron S, Keshishi M, Ademe Y, Tizazu A, Rose M, Uddin S, Ko M, Ddamba J, French D, Wallace A, Safieddine N, Gazala S, Simone C, Wright F, Finley C, Villeneuve PJ, Gilbert S, Sundaresan S, Seely AJ, Maziak DE, Manos D, Wallace A, Plourde M, Turner D, Ringash J, Manuel D, Kidane B, Gupta V, Decker K, Davis L, Darling G, Compton C, Coburn N, Boyes R, Pugliese M, Mahar A, Hsu A, Kokavec A, Ho E, Waterman R, Wang E, Harnden K, Malthaner R, Nayak R, Qiabi M, Malthaner R, Nayak R, Qiabi M, Patel Y, Farrokhyar F, Simunovic M, Hanna WC, Gowing S, Liu R, Srinathan S, Tan L, Malthaner R, Malik P, Patel Y, Hanna W, Fahim U, Hafizi A, Ashrafi AS, McKechnie T, Amin N, Doumouras A, Hong D, Eskicioglu C, Streu C, Wong C, Schiller D, Sue-Chue-Lam C, Paszat L, Scheer A, Hansen B, Kishibe T, Baxter N, Neary E, Ibrahim T, Verschoor C, Patel S, Merchant S, Chadi S, Romagnolo L, Verschoor C, Brown C, Karimuddin A, Raval M, Phang T, Vikis E, Melich G, Patel S, Patel S, Caycedo-Marulanda A, Merchant S, Booth C, Kong W, McClintock C, Bankhead C, Heneghan C, Patel S, McClintock C, Merchant S, Caycedo-Marulanda A, Booth C, Bankhead C, Heneghan C, Zhang L, Flemming J, Djerboua M, Nanji S, Caycedo-Marulanda A, Merchant S, Patel S, Demian M, Sabboobeh S, Moon J, Hulme-Moir M, Liberman AS, Feinberg S, Hayden DM, Chadi SA, Demyttenaere S, Samuel L, Hotakorzian N, Quintin L, Morin N, Ghitulescu G, Faria J, Vasilevsky CA, Boutros M, Mckechnie T, Khamar J, Ichhpuniani S, Eskicioglu C, Patel S, Merchant S, Caycedo-Marulanda A, Bankhead C, Heneghan C, Govind S, Lee J, Lee Y, Hong D, Eskicioglu C, Lu J, Khamar J, Lee Y, Amin N, Hong D, Eskicioglu C, Cardenas L, Schep D, Doumouras A, Hong D, Wong R, Levine O, Eskicioglu C, Mueller C, Stein B, Charlebois P, Liberman S, Fried G, Feldman L, Wang A, Liberman S, Charlebois P, Stein B, Fiore JF, Feldman L, Lee L, Wang A, Liberman S, Charlebois P, Stein B, Fiore JF, Feldman L, Lee L, Barkun A, Levy J, Bogdan R, Hawel J, Elnahas A, Alkhamesi NA, Schlachta CM, Caycedo-Marulanda A, Iaboni N, Hurlbut D, Kaufmann M, Ren KYM, Jamzad A, Mousavi P, Fichtinger G, Nicol CJ, Rudan JF, Brennan K, Caycedo-Marulanda A, Merchant S, McClintock C, Patel SV, McClintock C, Bankhead C, Merchant S, Caycedo-Marulanda A, Booth C, Heneghan C, Zhang L, Huo B, Donaldson A, Flemming J, Nanji S, Caycedo-Marulanda A, Merchant S, Brogly S, Patel S, Lenet T, Park L, Murthy S, Musselman R, McKechnie T, Lee J, Biro J, Lee Y, Park L, Doumouras A, Hong D, Eskicioglu C, Singh H, Helewa R, Reynolds K, Sibley K, Doupe M, Brennan K, Flemming J, Nanji S, Merchant S, Djerboua M, Caycedo-Marulanda A, Patel S, Johnson G, Hochman D, Helewa R, Garfinkle R, Dell’Aniello S, Zelkowitz P, Vasilevsky CA, Brassard P, Boutros M, Zoughlami A, Abibula W, Amar A, Ghitulescu G, Vasilevsky CA, Brassard P, Boutros M, Araji T, Pang A, Vasilevsky CA, Boutros M, Ehlebracht A, Faria J, Ghitulescu G, Morin N, Pang A, Vasilevsky CA, Boutros M, Robitaille S, Oliver M, Charlebois P, Stein B, Liberman S, Feldman LS, Lee L, Kennedy E, Victor C, Govindarajan A, Zhang L, Brennan K, Djerboua M, Nanji S, Merchant SJ, Caycedo-Marulanda A, Flemming J, Robitaille S, Penta R, Pook M, Fiore JF, Feldman L, Lee L, Wong-Chong N, Marinescu D, Bhatnagar S, Morin N, Ghitulescu G, Vasilevsky CA, Faria J, Boutros M, Arif A, Ladua G, Bhang E, Brown C, Donellan F, Stuart H, Loree J, Patel S, Zhang L, MacDonald PH, Merchant S, Barnett KW, Caycedo-Marulanda A, Brown C, Karimuddin A, Stuart H, Ghuman A, Phang T, Raval M, Yoon HM, Fragoso G, Oliero M, Calvé A, Rendos HV, Gonzalez E, Brereton NJ, Cuisiniere T, Gerkins C, Djediai S, Annabi B, Diop K, Routy B, Laplante P, Cailhier JF, Taleb N, Alratrout H, Dagbert F, Loungnarath R, Sebajang H, Schwenter F, Wassef R, Ratelle R, Debroux E, Richard C, Santos MM, Hamad D, Alsulaim H, Monton O, Marinescu D, Pang A, Vasilevsky CA, Boutros M, Marinescu D, Alqahtani M, Pang A, Ghitulescu G, Vasilevsky CA, Boutros M, Marinescu D, Garfinkle R, Boutros M, Zwiep T, Greenberg J, Lenet T, Musselman R, Williams L, Raiche I, McIsaac D, Thavorn K, Fergusson D, Moloo H, Charbonneau J, Paré X, Frigault J, Letarte F, Ott M, Karanicolas P, Brackstone M, Ashmalla S, Weaver J, Tagalakis V, Boutros M, Stotland P, Caycedo-Marulanda A, Moloo H, Jayaraman S. 2022 Canadian Surgery Forum Sept. 15–17, 202201. Operative classification of ventral abdominal hernias: new and practical classification02. Watchful waiting for large primary splenic cysts03. Transversus abdominis plane (TAP) blocks with and without dexamethasone in colorectal surgery04. What factors determine publication of resident research day projects?05. Characterization of near-infrared imaging and indocyanine green use amongst general surgeons06. Variation in opioid prescribing after outpatient breast surgery: Time for a streamlined approach?07. Trends in graduate degree types and research output for Canadian academic general surgeons08. Would you prefer to undergo breast-conserving therapy or a mastectomy for early breast cancer? Comparison of perceptions of general and plastic surgeons09. Lack of representation of women and BIPOC individuals in Canadian academic surgery10. Medical student interest and perspectives on pursuing surgical careers: a multicentre survey evaluating 5-year trends11. Difficult cholecystectomy with cholecystogastric fistula12. Surviving nonsurvivable injuries: patients who elude the “lethal” Abbreviated Injury Scale (AIS) score of six13. Gunshot wounds sustained during legal intervention v. those inflicted by civilians: a comparative analysis14. The impact of delayed time to first head CT on functional outcomes after blunt head trauma with moderately depressed GCS15. Contemporary utility of diagnostic peritoneal aspiration in trauma16. Impact of delayed time to first head CT in traumatic brain injury17. Radiologic predictors of in-hospital mortality after traumatic craniocervical dissociation18. Measurement properties of a patient-reported instrument to evaluate functional status after major surgery19. The safety of venous thromboembolism chemoprophylaxis use in endoscopic retrograde cholangiopancreatography20. Characterizing Canadian rural surgeons: trends over time and 10-year replacement needs21. Contextual interference for skills development and transfer in laparoscopic surgery: a randomized controlled trial22. Evaluating the accuracy and design of visual backgrounds in academic surgical journals23. Defining rural surgery in Canada24. Validity of video-based general and procedure-specific self-assessment tools for surgical trainees in laparoscopic cholecystectomy25. Examining the equity and diversity characteristics of academic general surgeons in Canada26. Video-based coaching for surgical residents: a systematic review and meta-analysis27. Very-low-energy diets prior to nonbariatric surgery: a systematic review and meta-analysis28. Factors associated with resident research success: a descriptive analysis of Canadian general surgery trainees29. Global surgery pilot curriculum in Canadian undergraduate medical education: a novel approach30. How useful is ultrasound in predicting surgical findings of “mild cholecystitis”?31. Implementing a colorectal surgery “virtual hospital”: description of a novel outpatient care pathway to advance surgical care32. Trends in training and workforce planning for Canadian pediatric surgeons: a 10-year model33. Patient perspectives on intraoperative blood transfusion: results of semistructured interviews with perioperative patients34. Understanding intraoperative transfusion decision-making variability: a qualitative study using the Theoretical Domains Framework35. Effectiveness of preoperative oscillating positive expiratory pressure (OPEP) therapy in reduction of postoperative respiratory morbidity in patients undergoing surgery: a systematic review37. Accuracy of point-of-care testing devices for hemoglobin in the operating room: a systematic review and meta-analysis38. Opioid-free analgesia after outpatient general surgery: a qualitative study focused on the perspectives of patients and clinicians involved in a pilot trial39. The impact of the COVID-19 pandemic on general surgery residency: an analysis of operative volumes by residents at a Canadian general surgery residency program40. Postoperative care protocols for elderly emergency surgical patients: a quality improvement initiative42. Adverse events following robotic compared to laparoscopic and open surgery: a population-based analysis43. Is accrual higher for patients randomized to pragmatic v. exploratory randomized clinical trials? A systematic review and meta-analysis44. Effect of preoperative proton-pump inhibitor use on postoperative infectious and renal complications after elective general surgery45. The early burden of COVID-19 in emergency general surgery care across Canada46. Laparoscopic subtotal cholecystectomy for the difficult gallbladder: evolution of technique at a single teaching hospital and retrospective review47. The demand for emergency general surgery in Canada: a public health crisis48. Attitudes of Canadian general surgery staff and residents toward point-of-care ultrasound49. Psychological impact of COVID-19 on Canadian surgical residents50. Validation of an artificial intelligence platform for the guidance of safe laparoscopic cholecystectomy51. Predictors of recurrent appendicitis after nonoperative management: a prospective cohort study52. The effect of the first wave of the COVID-19 pandemic on colorectal and hepatobiliary oncologic outcomes at a tertiary care centre53. Trends in training and workforce representation for Canadian general surgeons working in critical care: a descriptive study54. White presentation: teaching safe opioid prescription and opioid use disorder management in Canadian universities56. How bad is really bad, eh? Impact of the first wave of the COVID-19 pandemic on residents’ operative volume: the experience of a Canadian general surgery program57. Surgeon-specific encounters within a multidisciplinary care pathway: Is there a role for shared care models in surgery?59. A pan-Canadian analysis of approach to treatment for acute appendicitis60. Appendix neoplasm stratified by age: understanding the best treatment for appendicitis61. Predicting acute cholecystitis on final pathology to prioritize surgical urgency: an evaluation of the Tokyo criteria and development of a novel predictive score62. Obesity is an independent predictor of acute renal failure after surgery64. Validation of a clinical decision-making assessment tool in general surgery65. Moral distress in the provision of palliative care delivery for surgical patients in British Columbia: lessons learned from the perspectives of general surgeons66. Delays in presentation and severity of illness predict adverse surgical outcomes among patients transferred from rural Indigenous communities for acute care surgery67. Remote video-based suturing education with smartphones (REVISE): a randomized controlled trial68. Modified Delphi consensus on appropriate use of laboratory investigations in acute care surgery patients72. Impacts of inpatient food at a tertiary care centre on patient satisfaction, nutrition and planetary health73. Racial disparities in health outcomes for oncological surgery in Canada75. Risk of recurrent laryngeal nerve injury from thyroidectomy is lower when intraoperative nerve monitoring (IONM) is used: an analysis of 17 688 patients from the NSQIP database01. The impact of the COVID-19 pandemic on non-smallcell lung cancer pathologic stage and presentation02. Screening criteria evaluation for expansion in pulmonary neoplasias (screen)03. Robotic-assisted lobectomy for early-stage lung cancer provides better patient-reported quality of life than video-assisted lobectomy: early results of the RAVAL trial04. Breathe Anew: designing and testing the feasibility of a novel intervention for lung cancer survivorship05. Learning objectives for thoracic surgery: developing a national standard for undergraduate medical education06. Plasma cell-free DNA as a point-of-care well-being biomarker for early-stage non-small-cell lung cancer patients07. Sarcopenia determined by skeletal muscle index predicts overall survival, disease-free survival and postoperative complications in resectable esophageal cancer: a systematic review and meta-analysis08. The short- and long-term effects of open v. minimally invasive thymectomy in myasthenia gravis patients: a systematic review and meta-analysis09. Optimizing opioid prescribing practices following minimally invasive lung resections through a structured quality improvement process10. Effects of virtual postoperative postdischarge care in patients undergoing lung resection during the COVID-19 pandemic11. Initiating Ethiopia’s first minimally invasive surgery program: a novel approach for collaborations in global surgical education12. Patient outcomes following salvage lung cancer surgery after definitive chemotherapy or radiation13. Replacing chest X-rays after chest tube removal with clinical assessment in postoperative thoracic surgery patients14. Updating the practice of thoracic surgery in Canada: a survey of the Canadian Association of Thoracic Surgeons15. The impact of COVID-19 on the diagnosis and treatment of lung cancer16. Development of a prediction model for survival time in esophageal cancer patients treated with resection17. The development and validation of a mixed reality thoracic surgical anatomy atlas18. Routine placement of feeding tubes should be avoided in esophageal cancer patients undergoing surgery19. Nodal count is no different during robotic segmentectomy compared with robotic lobectomy20. Point-of-care ultrasound-guided percutaneous biopsy of solid masses in the thoracic outpatient clinic: a safe, high-yield procedure to accelerate tissue diagnosis for patients with advanced thoracic malignancy21. Sarcopenia and modified frailty index are not associated with adverse outcomes after esophagectomy for esophageal cancer: a retrospective cohort study22. Near-infrared-guided segmental resection for lung cancer: an analysis of the learning curve23. Routine use of feeding jejunostomy tubes in patients undergoing esophagectomy for esophageal malignancy is safe and associated with low complication rates01. Ghost ileostomy versus loop ileostomy following total mesorectal excision for rectal cancer: a systematic review and meta-analysis02. Analysis of 100 consecutive colorectal cancers presenting at a Canadian tertiary care centre: delayed diagnosis and advanced disease03. Clinical delays and comparative outcomes in younger and older adults with colorectal cancer: a systematic review04. Recurrence rates of rectal cancer after transanal total mesorectal excision (taTME): a systematic review and meta-analysis05. Transanal total mesorectal excision for abdominoperineal resection (taTME-APR) is associated with poor oncological outcomes in rectal cancer patients: a word of caution from a multicentric Canadian cohort study06. Association between survival and receipt of recommended and timely treatment in locally advanced rectal cancer: a population-based study07. Trends and the impact of incomplete preoperative staging in rectal cancer08. Postoperative outcomes after elective colorectal surgery in patients with cirrhosis09. Bowel stimulation before loop ileostomy closure to reduce postoperative ileus: a multicentre, single-blinded, randomized controlled trial10. Recurrence following perineal rectosigmoidectomy ( Altemeier) with levatorplasty: a systematic review and meta-analysis11. Nonmodifiable risk factors and receipt of surveillance investigations following treatment of rectal cancer12. Safety and effectiveness of endoscopic full-thickness resection for the management of colorectal lesions: a systematic review and meta-analysis13. Impact of preoperative carbohydrate loading before colectomy: a systematic review and meta-analysis of randomized controlled trials14. Statin therapy in patients undergoing short-course neoadjuvant radiotherapy for rectal cancer15. Feasibility of targeted lymphadenectomy during complete mesocolic excision for colon cancer using indocyanine green immunofluorescence lymphatic mapping16. Feasibility of expanding an ambulatory colectomy protocol: a retrospective analysis of early discharge following minimally invasive colectomy in an enhanced recovery pathway17. Impact of rectal cancer on bowel dysfunction before treatment and its relationship with post-treatment function18. Canadian cost–utility analysis of artificial-intelligence-assisted colonoscopy for adenoma detection in fecal immunochemical-based colorectal cancer screening19. A comparison of outcomes following intracorporeal and extracorporeal anastomotic techniques in laparoscopic right colectomies20. Assessment of metabolic signatures using desorption electrospray ionization mass spectrometry (DESI) and rapid evaporative ionization mass spectrometry (REIMS) of rectal cancer samples to assist in determining treatment response21. The association between hospital characteristics and minimally invasive rectal cancer surgery: a population-based study22. Cancer centre level designation and the impact on treatment and outcomes in those with rectal cancer: a population-based study23. Oncological outcomes after colorectal cancer in patients with liver cirrhosis: a systematic review and meta-analysis24. Optimal preoperative nutrition for penetrating Crohn disease: a systematic review and meta-analysis25. Lymph node ratio as a predictor of survival for colon cancer: a systematic review and meta-analysis26. Barriers and facilitators for use of new recommendations for optimal endoscopic localization of colorectal neoplasms according to gastroenterologists and surgeons27. Emergency colorectal surgery in patients with cirrhosis: a population-based descriptive study28. Local recurrence rates and associated risk factors after transanal endoscopic microsurgery for benign polyps and adenocarcinomas29. Bowel dysfunction impacts mental health after restorative proctectomy for rectal cancer30. Evolution of psychological morbidity following restorative proctectomy for rectal cancer: a systematic review and meta-analysis31. Frailty predicts LARS and quality of life in rectal cancer survivors after restorative proctectomy32. Low anterior resection syndrome in a reference North American population: prevalence and predictive factors33. The evolution of enhanced recovery: same day discharge after laparoscopic colectomy34. Effect of ERAS protocols on length of stay after colorectal surgery: an interrupted time series analysis35. Practice patterns and outcomes in individuals with cirrhosis and colorectal cancer: a population-based study36. Understanding the impact of bowel function on quality of life after rectal cancer surgery37. Right-sided colectomies for diverticulitis have worse outcomes compared with left-sided colectomies38. Symptom burden and time from symptom onset to cancer diagnosis in patients with early-onset colorectal cancer39. The impact of access to robotic rectal surgery at a tertiary care centre: a Canadian perspective40. Management of rectal neuroendocrine tumours by transanal endoscopic microsurgery41. The gut microbiota modulates colorectal anastomotic healing in patients undergoing surgery for colorectal cancer42. Is there added risk of complications for concomitant procedures during an ileocolic resection for Crohn disease?43. Cost of stoma-related hospital readmissions for rectal cancer patients following restorative proctectomy with a diverting loop ileostomy: a nationwide readmissions database analysis44. Older age associated with quality of rectal cancer care: an ACS-NSQIP database study45. Outcomes of patients undergoing elective bowel resection before and after implementation of an anemia screening and treatment program47. Loop ileostomy closure as a 23-hour stay procedure: a randomized controlled trial48. Extended duration perioperative thromboprophylaxis with low-molecular-weight heparin to improve disease-free survival following surgical resection of colorectal cancer: a multicentre randomized controlled trial (PERIOP-01 Trial)49. Three-stage versus modified 2-stage ileal pouch anal anastomosis: perioperative outcomes, function and quality of life50. Compliance with extended venous thromboembolism prophylaxis in rectal cancer51. Extended-duration venous thromboembolism prophylaxis after diversion in rectal cancer52. Financial and occupational impact of low anterior resection syndrome: a qualitative study53. Nonoperative management for rectal cancer: patient perspectives54. Trends in ileostomy-related emergency department visits for rectal cancer patients55. Long-term implications of treatment of fecal incontinence: a single Canadian centre’s retrospective cohort study: a 17-year follow-up56. Externally benchmarking colorectal resection outcomes in our province against the ACS NSQIP risk calculator: identifying opportunities for improvement57. Externally benchmarking our provincial colectomy outcomes against the ACS NSQIP using the Codman Score: to identify possible opportunities for improvement of outcomes58. Rural v. urban documentation of recommended practices for optimal endoscopic colorectal lesion localization01. Incidence of in-hospital opioid use and pain after inguinal hernia repair02. Ventral hernia repair following liver transplantation: outcome of repair techniques and risk factors for recurrence01. Impact of the COVID-19 pandemic on bariatric surgery in North America: a retrospective analysis of 834 647 patients02. Patient selection and 30-day outcomes of SADI-S compared to RYGB: a retrospective cohort study of 47 375 patients03. New persistent opioid use following bariatric surgery: a systematic review and pooled proportion meta-analysis04. Bariatric surgery should be offered to active-duty military personnel: a retrospective study of the Canadian Armed Forces experience05. Opioid prescribing practices and use following bariatric surgery: a systematic review and pooled summary of data06. Sacred sharing circles: urban Indigenous Manitobans’ experiences with bariatric surgery07. Gastrogastric hernia after laparoscopic gastric great curve plication: a video presentation08. Characterization of comorbidities predictive of bariatric surgery09. Efficacy of preoperative high-dose liraglutide in patients with superobesity10. The effect of linear stapled gastrojejunostomy size in Roux-en-Y gastric bypass11. Fragility of statistically significant outcomes in randomized trials comparing bariatric surgeries12. Weight loss outcomes for patients undergoing conversion to Roux-en-Y gastric bypass after sleeve gastrectomy13. Are long waiting lists for bariatric surgery detrimental to patients? A single-centre experience14. Does upper gastrointestinal swallow study after bariatric surgery lead to earlier detection of leak?15. Pharmaceutical utilization before and after bariatric surgery16. Same-day discharge Roux-en-Y gastric bypass at a Canadian bariatric centre: pathway implementation and early experiences17. Safety and efficiency of performing primary bariatric surgery at an ambulatory site of a tertiary care hospital: a 5-year experience18. Impact of psychiatric diagnosis on weight loss outcomes 3 years after bariatric surgery19. Ursodeoxycholic acid (UDCA) for prevention of gallstone disease after laparoscopic sleeve gastrectomy (LSG): an Atlantic Canada perspective20. Fecal microbial transplantation and fibre supplementation in patients with severe obesity and metabolic syndrome: a randomized double-blind, placebo-controlled phase 2 trial01. Incidence, timing and outcomes of venous thromboembolism in patients undergoing surgery for esophagogastric cancer: a population-based cohort study04. Omission of axillary staging and survival in elderly women with early-stage breast cancer: a population-based cohort study05. Patients’ experiences receiving cancer surgery during the COVID-19 pandemic: a qualitative study06. Cancer surgery outcomes are better at high-volume centres07. Attitudes of Canadian colorectal cancer care providers toward liver transplantation for colorectal liver metastases: a national survey08. Quality of narrative central and lateral neck dissection reports for thyroid cancer treatment suggests need for a national standardized synoptic operative template09. Transoral endoscopic thyroidectomy vestibular approach (TOETVA): indications and technique10. Temporal trends in lymph node assessment as a quality indicator in colorectal cancer patients treated at a high-volume Canadian centre11. Molecular landscape of early-stage breast cancer with nodal metastasis12. Beta testing of a risk-stratified patient decision aid to facilitate shared decision making for postoperative extended thromboprophylaxis in patients undergoing major abdominal surgery for cancer13. Breast reconstruction use and impact on oncologic outcomes among inflammatory breast cancer patients: a systematic review14. Association between patient-reported symptoms and health care resource utilization: a first step to develop patient-centred value measures in cancer care15. Complications after colorectal liver metastases resection in Newfoundland and Labrador16. Why do patients with nonmetastatic primary retroperitoneal sarcoma not undergo resection?17. Loss of FAM46Cexpression predicts inferior postresection survival and induces ion channelopathy in gastric adenocarcinoma18. Liver-directed therapy of neuroendocrine liver metastases19. Neoadjuvant pembrolizumab use in microsatellite instability high (MSI-H) rectal cancer: benefits of its use in lynch syndrome20. MOLLI for excision of nonpalpable breast lesions: a case series22. Patients awaiting mastectomy report increased depression, anxiety, and decreased quality of life compared with patients awaiting lumpectomy for treatment of breast cancer23. Is microscopic margin status important in retroperitoneal sarcoma (RPS) resection? A systematic review and meta-analysis24. Absence of benefit of routine surveillance in very-low-risk and low-risk gastric gastrointestinal stromal tumors25. Effect of intraoperative in-room specimen radiography on margin status in breast-conserving surgery26. Active surveillance for DCIS of the breast: qualitative interviews with patients and physicians01 Outcomes following extrahepatic and intraportal pancreatic islet transplantation: a comparative cohort study02. Cholang-funga-gitis03. Evaluating the effect of a low-calorie prehepatectomy diet on perioperative outcomes: a systematic review and meta-analysis04. Toxicity profiles of systemic therapy for advanced hepatocellular carcinoma: a systematic review to guide neoadjuvant trials05. Should cell salvage be used in liver resection and transplantation? A systematic review and meta-analysis06. The association between surgeon and hospital variation in use of laparoscopic liver resection and short-term outcomes07. Systematic review and meta-analysis of prognostic factors for early recurrence in intrahepatic cholangiocarcinoma after curative-intent resection08. Impact of neoadjuvant chemotherapy on postoperative outcomes of patients undergoing hepatectomy for intrahepatic cholangiocarcinoma: ACS-NSQIP propensity-matched analysis09. The impact of prophylactic negative pressure wound therapy on surgical site infections in pancreatic resection: a systematic review and meta-analysis10. Does hepatic pedicle clamping increase the risk of colonic anastomotic leak after combined hepatectomy and colectomy? Analysis of the ACS NSQIP database11. Development of a culture process to grow a full-liver tissue substitute12. Liver transplantation for fibrolamellar hepatocellular carcinoma: an analysis of the European Liver Transplant Registry13. Arming beneficial viruses to treat pancreatic cancer14. Hepaticoduodenostomy versus hepaticojenunostomy for biliary reconstruction: a retrospective review of a single-centre experience15. Feasibility and safety of a “shared care” model in complex hepatopancreatobiliary surgery: a 5-year analysis of pancreaticoduodenectomy16. Laparoscopic v. open pancreaticoduodenectomy: initial institutional experience and NSQIP-matched analysis17. Laparoscopic spleen-preserving distal pancreatectomy: Why not do a Warshaw?18. The impact of COVID-19 on pancreaticoduodenectomy outcomes in a high-volume hepatopancreatobiliary centre19. Transitioning from open to minimally invasive pancreaticoduodenectomy: the learning curve factor in an academic centre20. Closed-incision negative-pressure wound therapy following pancreaticoduodenectomy for prevention of surgical site infections in high-risk patients21. Robotic Appleby procedure for recurrent pancreatic cancer22. The influence of viral hepatitis status on posthepatectomy complications in patients with hepatocellular carcinoma: a NSQIP analysis. Can J Surg 2022. [DOI: 10.1503/cjs.014322] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/05/2022] Open
|
12
|
Chen B, Javadi G, Hamilton A, Sibley S, Laird P, Abolmaesumi P, Maslove D, Mousavi P. Quantifying deep neural network uncertainty for atrial fibrillation detection with limited labels. Sci Rep 2022; 12:20140. [PMID: 36418604 PMCID: PMC9684456 DOI: 10.1038/s41598-022-24574-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2022] [Accepted: 11/17/2022] [Indexed: 11/25/2022] Open
Abstract
Atrial fibrillation (AF) is the most common arrhythmia found in the intensive care unit (ICU), and is associated with many adverse outcomes. Effective handling of AF and similar arrhythmias is a vital part of modern critical care, but obtaining knowledge about both disease burden and effective interventions often requires costly clinical trials. A wealth of continuous, high frequency physiological data such as the waveforms derived from electrocardiogram telemetry are promising sources for enriching clinical research. Automated detection using machine learning and in particular deep learning has been explored as a solution for processing these data. However, a lack of labels, increased presence of noise, and inability to assess the quality and trustworthiness of many machine learning model predictions pose challenges to interpretation. In this work, we propose an approach for training deep AF models on limited, noisy data and report uncertainty in their predictions. Using techniques from the fields of weakly supervised learning, we leverage a surrogate model trained on non-ICU data to create imperfect labels for a large ICU telemetry dataset. We combine these weak labels with techniques to estimate model uncertainty without the need for extensive human data annotation. AF detection models trained using this process demonstrated higher classification performance (0.64-0.67 F1 score) and improved calibration (0.05-0.07 expected calibration error).
Collapse
Affiliation(s)
- Brian Chen
- grid.410356.50000 0004 1936 8331School of Computing, Queen’s University, Kingston, ON Canada
| | - Golara Javadi
- grid.17091.3e0000 0001 2288 9830Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC Canada
| | - Alexander Hamilton
- grid.410356.50000 0004 1936 8331School of Computing, Queen’s University, Kingston, ON Canada
| | - Stephanie Sibley
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, Kingston, ON Canada
| | - Philip Laird
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, Kingston, ON Canada
| | - Purang Abolmaesumi
- grid.17091.3e0000 0001 2288 9830Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC Canada
| | - David Maslove
- grid.410356.50000 0004 1936 8331Department of Critical Care Medicine, Queen’s University, Kingston, ON Canada
| | - Parvin Mousavi
- grid.410356.50000 0004 1936 8331School of Computing, Queen’s University, Kingston, ON Canada
| |
Collapse
|
13
|
Kitner N, Rodgers JR, Ungi T, Olding T, Joshi C, Mousavi P, Fichtinger G, Korzeniowski M. 49: Automated Catheter Tracking in 3D Ultrasound Images from High-Dose-Rate Prostate Brachytherapy Using Deep Learning and Feature Extraction. Radiother Oncol 2022. [DOI: 10.1016/s0167-8140(22)04328-6] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/14/2022]
|
14
|
Gerolami J, Wong JJM, Zhang R, Chen T, Imtiaz T, Smith M, Jamaspishvili T, Koti M, Glasgow JI, Mousavi P, Renwick N, Tyryshkin K. A Computational Approach to Identification of Candidate Biomarkers in High-Dimensional Molecular Data. Diagnostics (Basel) 2022; 12:diagnostics12081997. [PMID: 36010347 PMCID: PMC9407361 DOI: 10.3390/diagnostics12081997] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/05/2022] [Revised: 08/16/2022] [Accepted: 08/17/2022] [Indexed: 12/13/2022] Open
Abstract
Complex high-dimensional datasets that are challenging to analyze are frequently produced through ‘-omics’ profiling. Typically, these datasets contain more genomic features than samples, limiting the use of multivariable statistical and machine learning-based approaches to analysis. Therefore, effective alternative approaches are urgently needed to identify features-of-interest in ‘-omics’ data. In this study, we present the molecular feature selection tool, a novel, ensemble-based, feature selection application for identifying candidate biomarkers in ‘-omics’ data. As proof-of-principle, we applied the molecular feature selection tool to identify a small set of immune-related genes as potential biomarkers of three prostate adenocarcinoma subtypes. Furthermore, we tested the selected genes in a model to classify the three subtypes and compared the results to models built using all genes and all differentially expressed genes. Genes identified with the molecular feature selection tool performed better than the other models in this study in all comparison metrics: accuracy, precision, recall, and F1-score using a significantly smaller set of genes. In addition, we developed a simple graphical user interface for the molecular feature selection tool, which is available for free download. This user-friendly interface is a valuable tool for the identification of potential biomarkers in gene expression datasets and is an asset for biomarker discovery studies.
Collapse
Affiliation(s)
- Justin Gerolami
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Justin Jong Mun Wong
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Ricky Zhang
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Tong Chen
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Tashifa Imtiaz
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Miranda Smith
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Tamara Jamaspishvili
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada
- Department of Pathology & Laboratory Medicine, SUNY Upstate Medical University, Syracuse, NY 13210, USA
| | - Madhuri Koti
- Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON K7L 3N6, Canada
| | | | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Neil Renwick
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada
| | - Kathrin Tyryshkin
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada
- Correspondence: ; Tel.: +1-613-533-2345
| |
Collapse
|
15
|
Nezamabadi K, Mayfield J, Li P, Greenland GV, Rodriguez S, Simsek B, Mousavi P, Shatkay H, Abraham MR. Toward ECG-based analysis of hypertrophic cardiomyopathy: a novel ECG segmentation method for handling abnormalities. J Am Med Inform Assoc 2022; 29:1879-1889. [PMID: 35923089 PMCID: PMC9552290 DOI: 10.1093/jamia/ocac122] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/08/2022] [Revised: 06/22/2022] [Accepted: 07/21/2022] [Indexed: 11/13/2022] Open
Abstract
OBJECTIVE Abnormalities in impulse propagation and cardiac repolarization are frequent in hypertrophic cardiomyopathy (HCM), leading to abnormalities in 12-lead electrocardiograms (ECGs). Computational ECG analysis can identify electrophysiological and structural remodeling and predict arrhythmias. This requires accurate ECG segmentation. It is unknown whether current segmentation methods developed using datasets containing annotations for mostly normal heartbeats perform well in HCM. Here, we present a segmentation method to effectively identify ECG waves across 12-lead HCM ECGs. METHODS We develop (1) a web-based tool that permits manual annotations of P, P', QRS, R', S', T, T', U, J, epsilon waves, QRS complex slurring, and atrial fibrillation by 3 experts and (2) an easy-to-implement segmentation method that effectively identifies ECG waves in normal and abnormal heartbeats. Our method was tested on 131 12-lead HCM ECGs and 2 public ECG sets to evaluate its performance in non-HCM ECGs. RESULTS Over the HCM dataset, our method obtained a sensitivity of 99.2% and 98.1% and a positive predictive value of 92% and 95.3% when detecting QRS complex and T-offset, respectively, significantly outperforming a state-of-the-art segmentation method previously employed for HCM analysis. Over public ECG sets, it significantly outperformed 3 state-of-the-art methods when detecting P-onset and peak, T-offset, and QRS-onset and peak regarding the positive predictive value and segmentation error. It performed at a level similar to other methods in other tasks. CONCLUSION Our method accurately identified ECG waves in the HCM dataset, outperforming a state-of-the-art method, and demonstrated similar good performance as other methods in normal/non-HCM ECG sets.
Collapse
Affiliation(s)
- Kasra Nezamabadi
- Computational Biomedicine Lab, Computer and Information Sciences, University of Delaware, Newark, Delaware, USA
| | - Jacob Mayfield
- Division of Cardiology, Hypertrophic Cardiomyopathy Center of Excellence, University of California San Francisco, San Francisco, California, USA
| | - Pengyuan Li
- Computational Biomedicine Lab, Computer and Information Sciences, University of Delaware, Newark, Delaware, USA
| | - Gabriela V Greenland
- Division of Cardiology, Hypertrophic Cardiomyopathy Center of Excellence, University of California San Francisco, San Francisco, California, USA
| | - Sebastian Rodriguez
- Division of Cardiology, Hypertrophic Cardiomyopathy Center of Excellence, University of California San Francisco, San Francisco, California, USA
| | - Bahadir Simsek
- Division of Cardiology, Hypertrophic Cardiomyopathy Center of Excellence, University of California San Francisco, San Francisco, California, USA
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, Ontario, Canada
| | - Hagit Shatkay
- Computational Biomedicine Lab, Computer and Information Sciences, University of Delaware, Newark, Delaware, USA
| | - M Roselle Abraham
- Hypertrophic Cardiomyopathy Center of Excellence, Division of Cardiology, University of California San Francisco, San Francisco, USA
| |
Collapse
|
16
|
Ehrlich J, Jamzad A, Asselin M, Rodgers JR, Kaufmann M, Haidegger T, Rudan J, Mousavi P, Fichtinger G, Ungi T. Sensor-Based Automated Detection of Electrosurgical Cautery States. Sensors 2022; 22:s22155808. [PMID: 35957364 PMCID: PMC9371045 DOI: 10.3390/s22155808] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 07/07/2022] [Revised: 07/30/2022] [Accepted: 08/01/2022] [Indexed: 02/04/2023]
Abstract
In computer-assisted surgery, it is typically required to detect when the tool comes into contact with the patient. In activated electrosurgery, this is known as the energy event. By continuously tracking the electrosurgical tools’ location using a navigation system, energy events can help determine locations of sensor-classified tissues. Our objective was to detect the energy event and determine the settings of electrosurgical cautery—robustly and automatically based on sensor data. This study aims to demonstrate the feasibility of using the cautery state to detect surgical incisions, without disrupting the surgical workflow. We detected current changes in the wires of the cautery device and grounding pad using non-invasive current sensors and an oscilloscope. An open-source software was implemented to apply machine learning on sensor data to detect energy events and cautery settings. Our methods classified each cautery state at an average accuracy of 95.56% across different tissue types and energy level parameters altered by surgeons during an operation. Our results demonstrate the feasibility of automatically identifying energy events during surgical incisions, which could be an important safety feature in robotic and computer-integrated surgery. This study provides a key step towards locating tissue classifications during breast cancer operations and reducing the rate of positive margins.
Collapse
Affiliation(s)
- Josh Ehrlich
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Amoon Jamzad
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Mark Asselin
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Jessica Robin Rodgers
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Martin Kaufmann
- Department of Surgery, Kingston Health Sciences Centre, Kingston, ON K7L 2V7, Canada; (M.K.); (J.R.)
| | - Tamas Haidegger
- University Research and Innovation Center (EKIK), Óbuda University, 1034 Budapest, Hungary
- Correspondence: (T.H.); (T.U.)
| | - John Rudan
- Department of Surgery, Kingston Health Sciences Centre, Kingston, ON K7L 2V7, Canada; (M.K.); (J.R.)
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Gabor Fichtinger
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
| | - Tamas Ungi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (J.E.); (A.J.); (M.A.); (J.R.R.); (P.M.); (G.F.)
- Correspondence: (T.H.); (T.U.)
| |
Collapse
|
17
|
Connolly L, Deguet A, Leonard S, Tokuda J, Ungi T, Krieger A, Kazanzides P, Mousavi P, Fichtinger G, Taylor RH. Bridging 3D Slicer and ROS2 for Image-Guided Robotic Interventions. Sensors (Basel) 2022; 22:5336. [PMID: 35891016 PMCID: PMC9324680 DOI: 10.3390/s22145336] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 06/20/2022] [Revised: 07/13/2022] [Accepted: 07/14/2022] [Indexed: 06/15/2023]
Abstract
Developing image-guided robotic systems requires access to flexible, open-source software. For image guidance, the open-source medical imaging platform 3D Slicer is one of the most adopted tools that can be used for research and prototyping. Similarly, for robotics, the open-source middleware suite robot operating system (ROS) is the standard development framework. In the past, there have been several "ad hoc" attempts made to bridge both tools; however, they are all reliant on middleware and custom interfaces. Additionally, none of these attempts have been successful in bridging access to the full suite of tools provided by ROS or 3D Slicer. Therefore, in this paper, we present the SlicerROS2 module, which was designed for the direct use of ROS2 packages and libraries within 3D Slicer. The module was developed to enable real-time visualization of robots, accommodate different robot configurations, and facilitate data transfer in both directions (between ROS and Slicer). We demonstrate the system on multiple robots with different configurations, evaluate the system performance and discuss an image-guided robotic intervention that can be prototyped with this module. This module can serve as a starting point for clinical system development that reduces the need for custom interfaces and time-intensive platform setup.
Collapse
Affiliation(s)
- Laura Connolly
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (T.U.); (P.M.); (G.F.)
| | - Anton Deguet
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
| | - Simon Leonard
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
| | | | - Tamas Ungi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (T.U.); (P.M.); (G.F.)
| | - Axel Krieger
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
| | - Peter Kazanzides
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (T.U.); (P.M.); (G.F.)
| | - Gabor Fichtinger
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (T.U.); (P.M.); (G.F.)
| | - Russell H. Taylor
- Whiting School of Engineering, Johns Hopkins University, Baltimore, MD 21218, USA; (A.D.); (S.L.); (A.K.); (P.K.); (R.H.T.)
| |
Collapse
|
18
|
Thirumal S, Jamzad A, Cotechini T, Siemens DR, Mousavi P. Automated Cell Phenotyping for Imaging Mass Cytometry. Annu Int Conf IEEE Eng Med Biol Soc 2022; 2022:426-429. [PMID: 36085862 DOI: 10.1109/embc48229.2022.9871071] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/15/2023]
Abstract
Imaging mass cytometry (IMC) is a new advancement in tissue imaging that is quickly gaining wider usage since its recent launch. It improves upon current tissue imaging methods by allowing for a significantly higher number of proteins to be imaged at once on a single tissue slide. For most analyses of IMC data, determining the phenotype of each cell is a crucial step. Current methods of phenotyping require sufficient biological knowledge regarding the protein expression profile of the various cell types. Here, we develop a deep convolutional autoencoder-classifier to automate the cell phenotyping process into four basic cell types. Biopsy tissue from bladder cancer patients is used to evaluate the efficacy of the classification. The model is evaluated and validated through feature importance, confirming that the significant features are biologically relevant. Our results demonstrate the potential of deep learning to automate the task of cell phenotyping for high-dimensional IMC data.
Collapse
|
19
|
Hu Z, Nasute Fauerbach PV, Yeung C, Ungi T, Rudan J, Engel CJ, Mousavi P, Fichtinger G, Jabs D. Real-time automatic tumor segmentation for ultrasound-guided breast-conserving surgery navigation. Int J Comput Assist Radiol Surg 2022; 17:1663-1672. [PMID: 35588339 DOI: 10.1007/s11548-022-02658-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/10/2022] [Accepted: 04/22/2022] [Indexed: 11/05/2022]
Abstract
PURPOSE Ultrasound-based navigation is a promising method in breast-conserving surgery, but tumor contouring often requires a radiologist at the time of surgery. Our goal is to develop a real-time automatic neural network-based tumor contouring process for intraoperative guidance. Segmentation accuracy is evaluated by both pixel-based metrics and expert visual rating. METHODS This retrospective study includes 7318 intraoperative ultrasound images acquired from 33 breast cancer patients, randomly split between 80:20 for training and testing. We implement a u-net architecture to label each pixel on ultrasound images as either tumor or healthy breast tissue. Quantitative metrics are calculated to evaluate the model's accuracy. Contour quality and usability are also assessed by fellowship-trained breast radiologists and surgical oncologists. Additionally, the viability of using our u-net model in an existing surgical navigation system is evaluated by measuring the segmentation frame rate. RESULTS The mean dice similarity coefficient of our u-net model is 0.78, with an area under the receiver-operating characteristics curve of 0.94, sensitivity of 0.95, and specificity of 0.67. Expert visual ratings are positive, with 93% of responses rating tumor contour quality at or above 7/10, and 75% of responses rating contour quality at or above 8/10. Real-time tumor segmentation achieved a frame rate of 16 frames-per-second, sufficient for clinical use. CONCLUSION Neural networks trained with intraoperative ultrasound images provide consistent tumor segmentations that are well received by clinicians. These findings suggest that neural networks are a promising adjunct to alleviate radiologist workload as well as improving efficiency in breast-conserving surgery navigation systems.
Collapse
Affiliation(s)
- Zoe Hu
- School of Medicine, Queen's University, 88 Stuart Street, Kingston, ON, K7L 3N6, Canada.
| | | | - Chris Yeung
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Tamas Ungi
- School of Computing, Queen's University, Kingston, ON, Canada
| | - John Rudan
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Cecil Jay Engel
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, ON, Canada
| | | | - Doris Jabs
- Department of Radiology, Queen's University, Kingston, ON, Canada
| |
Collapse
|
20
|
Thirumal S, Jamzad A, Cotechini T, Hindmarch CT, Graham CH, Siemens DR, Mousavi P. TITAN
: An
End‐to‐End
Data Analysis Environment for the
Hyperion
Imaging System. Cytometry A 2022; 101:423-433. [DOI: 10.1002/cyto.a.24535] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/10/2021] [Revised: 12/30/2021] [Accepted: 01/11/2022] [Indexed: 11/10/2022]
Affiliation(s)
| | - Amoon Jamzad
- Queen's University, School of Computing Kingston Canada
| | - Tiziana Cotechini
- Department of Biomedical and Molecular Sciences Queen's University Kington Canada
| | | | - Charles H. Graham
- Department of Biomedical and Molecular Sciences Queen's University Kington Canada
| | - D. Robert Siemens
- Department of Biomedical and Molecular Sciences Queen's University Kington Canada
- Department of Urology Queen's University Kingston Canada
| | | |
Collapse
|
21
|
Kaczmarek E, Nanayakkara J, Sedghi A, Pesteie M, Tuschl T, Renwick N, Mousavi P. Topology preserving stratification of tissue neoplasticity using Deep Neural Maps and microRNA signatures. BMC Bioinformatics 2022; 23:38. [PMID: 35026982 PMCID: PMC8756719 DOI: 10.1186/s12859-022-04559-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/05/2021] [Accepted: 12/30/2021] [Indexed: 11/14/2022] Open
Abstract
Background Accurate cancer classification is essential for correct treatment selection and better prognostication. microRNAs (miRNAs) are small RNA molecules that negatively regulate gene expression, and their dyresgulation is a common disease mechanism in many cancers. Through a clearer understanding of miRNA dysregulation in cancer, improved mechanistic knowledge and better treatments can be sought. Results We present a topology-preserving deep learning framework to study miRNA dysregulation in cancer. Our study comprises miRNA expression profiles from 3685 cancer and non-cancer tissue samples and hierarchical annotations on organ and neoplasticity status. Using unsupervised learning, a two-dimensional topological map is trained to cluster similar tissue samples. Labelled samples are used after training to identify clustering accuracy in terms of tissue-of-origin and neoplasticity status. In addition, an approach using activation gradients is developed to determine the attention of the networks to miRNAs that drive the clustering. Using this deep learning framework, we classify the neoplasticity status of held-out test samples with an accuracy of 91.07%, the tissue-of-origin with 86.36%, and combined neoplasticity status and tissue-of-origin with an accuracy of 84.28%. The topological maps display the ability of miRNAs to recognize tissue types and neoplasticity status. Importantly, when our approach identifies samples that do not cluster well with their respective classes, activation gradients provide further insight in cancer subtypes or grades. Conclusions An unsupervised deep learning approach is developed for cancer classification and interpretation. This work provides an intuitive approach for understanding molecular properties of cancer and has significant potential for cancer classification and treatment selection.
Collapse
|
22
|
Santilli AML, Ren K, Oleschuk R, Kaufmann M, Rudan J, Fichtinger G, Mousavi P. Application of Intraoperative Mass Spectrometry and Data Analytics for Oncological Margin Detection, A Review. IEEE Trans Biomed Eng 2022; 69:2220-2232. [PMID: 34982670 DOI: 10.1109/tbme.2021.3139992] [Citation(s) in RCA: 5] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/08/2022]
Abstract
OBJECTIVE A common phase of early-stage oncological treatment is the surgical resection of cancerous tissue. The presence of cancer cells on the resection margin, referred to as positive margin, is correlated with the recurrence of cancer and may require re-operation, negatively impacting many facets of patient outcomes. There exists a significant gap in the surgeons ability to intraoperatively delineate between tissues. Mass spectrometry methods have shown considerable promise as intraoperative tissue profiling tools that can assist with the complete resection of cancer. To do so, the vastness of the information collected through these modalities must be digested, relying on robust and efficient extraction of insights through data analysis pipelines. METHODS We review clinical mass spectrometry literature and prioritize intraoperatively applied modalities. We also survey the data analysis methods employed in these studies. RESULTS Our review outlines the advantages and shortcomings of mass spectrometry imaging and point-based tissue probing methods. For each modality, we identify statistical, linear transformation and machine learning techniques that demonstrate high performance in classifying cancerous tissues across several organ systems. A limited number of studies presented results captured intraoperatively. CONCLUSION Through continued research of data centric techniques, like mass spectrometry, and the development of robust analysis approaches, intraoperative margin assessment is becoming feasible. SIGNIFICANCE By establishing the relatively short history of mass spectrometry techniques applied to surgical studies, we hope to inform future applications and aid in the selection of suitable data analysis frameworks for the development of intraoperative margin detection technologies.
Collapse
|
23
|
Kaczmarek E, Jamzad A, Imtiaz T, Nanayakkara J, Renwick N, Mousavi P. Multi-Omic Graph Transformers for Cancer Classification and Interpretation. Pac Symp Biocomput 2022; 27:373-384. [PMID: 34890164] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Subscribe] [Scholar Register] [Indexed: 06/13/2023]
Abstract
Next-generation sequencing has provided rapid collection and quantification of 'big' biological data. In particular, multi-omics and integration of different molecular data such as miRNA and mRNA can provide important insights to disease classification and processes. There is a need for computational methods that can correctly model and interpret these relationships, and handle the difficulties of large-scale data. In this study, we develop a novel method of representing miRNA-mRNA interactions to classify cancer. Specifically, graphs are designed to account for the interactions and biological communication between miRNAs and mRNAs, using message-passing and attention mechanisms. Patient-matched miRNA and mRNA expression data is obtained from The Cancer Genome Atlas for 12 cancers, and targeting information is incorporated from TargetScan. A Graph Transformer Network (GTN) is selected to provide high interpretability of classification through self-attention mechanisms. The GTN is able to classify the 12 different cancers with an accuracy of 93.56% and is compared to a Graph Convolutional Network, Random Forest, Support Vector Machine, and Multilayer Perceptron. While the GTN does not outperform all of the other classifiers in terms of accuracy, it allows high interpretation of results. Multi-omics models are compared and generally outperform their respective single-omics performance. Extensive analysis of attention identifies important targeting pathways and molecular biomarkers based on integrated miRNA and mRNA expression.
Collapse
Affiliation(s)
- Emily Kaczmarek
- Medical Informatics Laboratory, School of Computing, Queen's University, Kingston, K7L 3N6, Canada,
| | | | | | | | | | | |
Collapse
|
24
|
Kaczmarek E, Pyman B, Nanayakkara J, Tuschl T, Tyryshkin K, Renwick N, Mousavi P. Discriminating Neoplastic from Nonneoplastic Tissues Using an miRNA-Based Deep Cancer Classifier. Am J Pathol 2021; 192:344-352. [PMID: 34774515 DOI: 10.1016/j.ajpath.2021.10.012] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Received: 08/15/2021] [Revised: 10/07/2021] [Accepted: 10/13/2021] [Indexed: 10/19/2022]
Abstract
Next-generation sequencing has enabled the collection of large biological data sets, allowing novel molecular-based classification methods to be developed for increased understanding of disease. miRNAs are small regulatory RNA molecules that can be quantified using next-generation sequencing and are excellent classificatory markers. Herein, we adapt a deep cancer classifier (DCC) to differentiate neoplastic from nonneoplastic samples using comprehensive miRNA expression profiles from 1031 human breast and skin tissue samples. The classifier was fine-tuned and evaluated using 750 neoplastic and 281 nonneoplastic breast and skin tissue samples. Performance of the DCC was compared with two machine-learning classifiers: support vector machine and random forests. In addition, performance of feature extraction through the DCC was also compared with a developed feature selection algorithm, cancer specificity. The DCC had the highest performance of area under the receiver operating curve and high performance in both sensitivity and specificity, unlike machine-learning and feature selection models, which often performed well in one metric compared with the other. In particular, deep learning was shown to have noticeable advantages with highly heterogeneous data sets. In addition, our cancer specificity algorithm identified candidate biomarkers for differentiating neoplastic and nonneoplastic tissue samples (eg, miR-144 and miR-375 in breast cancer and miR-375 and miR-451 in skin cancer).
Collapse
Affiliation(s)
- Emily Kaczmarek
- Medical Informatics Laboratory, School of Computing, Queen's University, Kingston, Ontario, Canada.
| | - Blake Pyman
- Medical Informatics Laboratory, School of Computing, Queen's University, Kingston, Ontario, Canada
| | - Jina Nanayakkara
- Laboratory of Translational RNA Biology, Department of Pathology and Molecular Medicine, Queen's University, Kingston, Ontario, Canada
| | - Thomas Tuschl
- Laboratory of RNA Molecular Biology, Rockefeller University, New York, New York
| | - Kathrin Tyryshkin
- Laboratory of Translational RNA Biology, Department of Pathology and Molecular Medicine, Queen's University, Kingston, Ontario, Canada
| | - Neil Renwick
- Laboratory of Translational RNA Biology, Department of Pathology and Molecular Medicine, Queen's University, Kingston, Ontario, Canada.
| | - Parvin Mousavi
- Medical Informatics Laboratory, School of Computing, Queen's University, Kingston, Ontario, Canada
| |
Collapse
|
25
|
Aghaeeaval M, Bendahan N, Shivji Z, McInnis C, Jamzad A, Lomax LB, Shukla G, Mousavi P, Winston GP. Prediction of patient survival following postanoxic coma using EEG data and clinical features. Annu Int Conf IEEE Eng Med Biol Soc 2021; 2021:997-1000. [PMID: 34891456 DOI: 10.1109/embc46164.2021.9629946] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/13/2023]
Abstract
Electroencephalography (EEG) is an effective and non-invasive technique commonly used to monitor brain activity and assist in outcome prediction for comatose patients post cardiac arrest. EEG data may demonstrate patterns associated with poor neurological outcome for patients with hypoxic injury. Thus, both quantitative EEG (qEEG) and clinical data contain prognostic information for patient outcome. In this study we use machine learning (ML) techniques, random forest (RF) and support vector machine (SVM) to classify patient outcome post cardiac arrest using qEEG and clinical feature sets, individually and combined. Our ML experiments show RF and SVM perform better using the joint feature set. In addition, we extend our work by implementing a convolutional neural network (CNN) based on time-frequency images derived from EEG to compare with our qEEG ML models. The results demonstrate significant performance improvement in outcome prediction using non-feature based CNN compared to our feature based ML models. Implementation of ML and DL methods in clinical practice have the potential to improve reliability of traditional qualitative assessments for postanoxic coma patients.
Collapse
|
26
|
Connolly L, Jamzad A, Kaufmann M, Farquharson CE, Ren K, Rudan JF, Fichtinger G, Mousavi P. Combined Mass Spectrometry and Histopathology Imaging for Perioperative Tissue Assessment in Cancer Surgery. J Imaging 2021; 7:203. [PMID: 34677289 PMCID: PMC8539093 DOI: 10.3390/jimaging7100203] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/09/2021] [Revised: 09/28/2021] [Accepted: 09/30/2021] [Indexed: 12/16/2022] Open
Abstract
Mass spectrometry is an effective imaging tool for evaluating biological tissue to detect cancer. With the assistance of deep learning, this technology can be used as a perioperative tissue assessment tool that will facilitate informed surgical decisions. To achieve such a system requires the development of a database of mass spectrometry signals and their corresponding pathology labels. Assigning correct labels, in turn, necessitates precise spatial registration of histopathology and mass spectrometry data. This is a challenging task due to the domain differences and noisy nature of images. In this study, we create a registration framework for mass spectrometry and pathology images as a contribution to the development of perioperative tissue assessment. In doing so, we explore two opportunities in deep learning for medical image registration, namely, unsupervised, multi-modal deformable image registration and evaluation of the registration. We test this system on prostate needle biopsy cores that were imaged with desorption electrospray ionization mass spectrometry (DESI) and show that we can successfully register DESI and histology images to achieve accurate alignment and, consequently, labelling for future training. This automation is expected to improve the efficiency and development of a deep learning architecture that will benefit the use of mass spectrometry imaging for cancer diagnosis.
Collapse
Affiliation(s)
- Laura Connolly
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (A.J.); (C.E.F.); (G.F.); (P.M.)
| | - Amoon Jamzad
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (A.J.); (C.E.F.); (G.F.); (P.M.)
| | - Martin Kaufmann
- Department of Surgery, Queen’s University, Kingston, ON K7L 3N6, Canada; (M.K.); (J.F.R.)
| | - Catriona E. Farquharson
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (A.J.); (C.E.F.); (G.F.); (P.M.)
| | - Kevin Ren
- Department of Pathology and Molecular Medicine, Queen’s University, Kingston, ON K7L 3N6, Canada;
| | - John F. Rudan
- Department of Surgery, Queen’s University, Kingston, ON K7L 3N6, Canada; (M.K.); (J.F.R.)
| | - Gabor Fichtinger
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (A.J.); (C.E.F.); (G.F.); (P.M.)
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 3N6, Canada; (A.J.); (C.E.F.); (G.F.); (P.M.)
| |
Collapse
|
27
|
Isen J, Perera-Ortega A, Vos SB, Rodionov R, Kanber B, Chowdhury FA, Duncan JS, Mousavi P, Winston GP. Non-parametric combination of multimodal MRI for lesion detection in focal epilepsy. Neuroimage Clin 2021; 32:102837. [PMID: 34619650 PMCID: PMC8503566 DOI: 10.1016/j.nicl.2021.102837] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 05/06/2021] [Revised: 09/10/2021] [Accepted: 09/20/2021] [Indexed: 12/21/2022]
Abstract
Multivariate voxel-based analysis useful for lesion detection in focal epilepsy. Non-parametric combination algorithm used to combine data from various MR sequences. Successful lesion detection demonstrated in MRI-positive and MRI-negative patients. Multimodal analysis detected abnormalities from diverse epileptogenic pathologies. Sensitivity of multivariate analysis notably higher than univariate analyses.
One third of patients with medically refractory focal epilepsy have normal-appearing MRI scans. This poses a problem as identification of the epileptogenic region is required for surgical treatment. This study performs a multimodal voxel-based analysis (VBA) to identify brain abnormalities in MRI-negative focal epilepsy. Data was collected from 69 focal epilepsy patients (42 with discrete lesions on MRI scans, 27 with no visible findings on scans), and 62 healthy controls. MR images comprised T1-weighted, fluid-attenuated inversion recovery (FLAIR), fractional anisotropy (FA) and mean diffusivity (MD) from diffusion tensor imaging, and neurite density index (NDI) from neurite orientation dispersion and density imaging. These multimodal images were coregistered to T1-weighted scans, normalized to a standard space, and smoothed with 8 mm FWHM. Initial analysis performed voxel-wise one-tailed t-tests separately on grey matter concentration (GMC), FLAIR, FA, MD, and NDI, comparing patients with epilepsy to controls. A multimodal non-parametric combination (NPC) analysis was also performed simultaneously on FLAIR, FA, MD, and NDI. Resulting p-maps were family-wise error rate corrected, threshold-free cluster enhanced, and thresholded at p < 0.05. Sensitivity was established through visual comparison of results to manually drawn lesion masks or seizure onset zone (SOZ) from stereoelectroencephalography. A leave-one-out cross-validation with the same analysis protocols was performed on controls to determine specificity. NDI was the best performing individual modality, detecting focal abnormalities in 38% of patients with normal MRI and conclusive SOZ. GMC demonstrated the lowest sensitivity at 19%. NPC provided superior performance to univariate analyses with 50% sensitivity. Specificity in controls ranged between 96 and 100% for all analyses. This study demonstrated the utility of a multimodal VBA utilizing NPC for detecting epileptogenic lesions in MRI-negative focal epilepsy. Future work will apply this approach to datasets from other centres and will experiment with different combinations of MR sequences.
Collapse
Affiliation(s)
- Jonah Isen
- School of Computing, Queen's University, Kingston, Canada
| | | | - Sjoerd B Vos
- Centre for Medical Image Computing, University College London, London, UK; Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; MRI Unit, Epilepsy Society, Chalfont St Peter, UK; National Institute for Health Research Biomedical Research Centre at University College London and University College London NHS Foundation Trust, London, UK; Neuroradiological Academic Unit, UCL Queen Square Institute of Neurology, University College London, London, UK
| | - Roman Rodionov
- Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; MRI Unit, Epilepsy Society, Chalfont St Peter, UK
| | - Baris Kanber
- Centre for Medical Image Computing, University College London, London, UK; Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; MRI Unit, Epilepsy Society, Chalfont St Peter, UK; National Institute for Health Research Biomedical Research Centre at University College London and University College London NHS Foundation Trust, London, UK
| | - Fahmida A Chowdhury
- Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; National Institute for Health Research Biomedical Research Centre at University College London and University College London NHS Foundation Trust, London, UK
| | - John S Duncan
- Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; MRI Unit, Epilepsy Society, Chalfont St Peter, UK; National Institute for Health Research Biomedical Research Centre at University College London and University College London NHS Foundation Trust, London, UK
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, Canada
| | - Gavin P Winston
- School of Computing, Queen's University, Kingston, Canada; Department of Clinical and Experimental Epilepsy, UCL Queen Square Institute of Neurology, University College London, London, UK; MRI Unit, Epilepsy Society, Chalfont St Peter, UK; National Institute for Health Research Biomedical Research Centre at University College London and University College London NHS Foundation Trust, London, UK; Department of Medicine, Division of Neurology & Centre for Neuroscience Studies, Queen's University, Kingston, Canada.
| |
Collapse
|
28
|
Fichtinger G, Mousavi P, Ungi T, Fenster A, Abolmaesumi P, Kronreif G, Ruiz-Alzola J, Ndoye A, Diao B, Kikinis R. Design of an Ultrasound-Navigated Prostate Cancer Biopsy System for Nationwide Implementation in Senegal. J Imaging 2021; 7:154. [PMID: 34460790 PMCID: PMC8404908 DOI: 10.3390/jimaging7080154] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/27/2021] [Revised: 08/04/2021] [Accepted: 08/07/2021] [Indexed: 12/05/2022] Open
Abstract
This paper presents the design of NaviPBx, an ultrasound-navigated prostate cancer biopsy system. NaviPBx is designed to support an affordable and sustainable national healthcare program in Senegal. It uses spatiotemporal navigation and multiparametric transrectal ultrasound to guide biopsies. NaviPBx integrates concepts and methods that have been independently validated previously in clinical feasibility studies and deploys them together in a practical prostate cancer biopsy system. NaviPBx is based entirely on free open-source software and will be shared as a free open-source program with no restriction on its use. NaviPBx is set to be deployed and sustained nationwide through the Senegalese Military Health Service. This paper reports on the results of the design process of NaviPBx. Our approach concentrates on "frugal technology", intended to be affordable for low-middle income (LMIC) countries. Our project promises the wide-scale application of prostate biopsy and will foster time-efficient development and programmatic implementation of ultrasound-guided diagnostic and therapeutic interventions in Senegal and beyond.
Collapse
Affiliation(s)
- Gabor Fichtinger
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada; (P.M.); (T.U.)
| | - Parvin Mousavi
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada; (P.M.); (T.U.)
| | - Tamas Ungi
- School of Computing, Queen’s University, Kingston, ON K7L 2N8, Canada; (P.M.); (T.U.)
| | - Aaron Fenster
- Department of Medical Biophysics, Schulich School of Medicine & Dentistry, Western University, London, ON N6A 5B7, Canada;
| | - Purang Abolmaesumi
- Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC V6T 1Z4, Canada;
| | - Gernot Kronreif
- Austrian Center for Medical Innovation and Technology, 2700 Wiener Neustadt, Austria;
| | - Juan Ruiz-Alzola
- Departamento de Señales y Comunicaciones, University of Las Palmas de Gran Canaria, 35001 Las Palmas, Spain;
| | - Alain Ndoye
- Department of Urology, Hôpital Aristide Le Dantec, Cheikh Anta Diop University, Dakar 10700, Senegal; (A.N.); (B.D.)
| | - Babacar Diao
- Department of Urology, Hôpital Aristide Le Dantec, Cheikh Anta Diop University, Dakar 10700, Senegal; (A.N.); (B.D.)
- Department of Urology, Ouakam Military Hospital, Dakar BP 5321, Senegal
| | - Ron Kikinis
- Harvard Medical School, Brigham and Women’s Hospital, Boston, MA 02115, USA;
| |
Collapse
|
29
|
Santilli AML, Jamzad A, Sedghi A, Kaufmann M, Logan K, Wallis J, Ren KYM, Janssen N, Merchant S, Engel J, McKay D, Varma S, Wang A, Fichtinger G, Rudan JF, Mousavi P. Domain adaptation and self-supervised learning for surgical margin detection. Int J Comput Assist Radiol Surg 2021; 16:861-869. [PMID: 33956307 DOI: 10.1007/s11548-021-02381-6] [Citation(s) in RCA: 8] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/17/2021] [Accepted: 04/13/2021] [Indexed: 01/02/2023]
Abstract
PURPOSE One in five women who undergo breast conserving surgery will need a second revision surgery due to remaining tumor. The iKnife is a mass spectrometry modality that produces real-time margin information based on the metabolite signatures in surgical smoke. Using this modality and real-time tissue classification, surgeons could remove all cancerous tissue during the initial surgery, improving many facets of patient outcomes. An obstacle in developing a iKnife breast cancer recognition model is the destructive, time-consuming and sensitive nature of the data collection that limits the size of the datasets. METHODS We address these challenges by first, building a self-supervised learning model from limited, weakly labeled data. By doing so, the model can learn to contextualize the general features of iKnife data from a more accessible cancer type. Second, the trained model can then be applied to a cancer classification task on breast data. This domain adaptation allows for the transfer of learnt weights from models of one tissue type to another. RESULTS Our datasets contained 320 skin burns (129 tumor burns, 191 normal burns) from 51 patients and 144 breast tissue burns (41 tumor and 103 normal) from 11 patients. We investigate the effect of different hyper-parameters on the performance of the final classifier. The proposed two-step method performed statistically significantly better than a baseline model (p-value < 0.0001), by achieving an accuracy, sensitivity and specificity of 92%, 88% and 92%, respectively. CONCLUSION This is the first application of domain transfer for iKnife REIMS data. We showed that having a limited number of breast data samples for training a classifier can be compensated by self-supervised learning and domain adaption on a set of unlabeled skin data. We plan to confirm this performance by collecting new breast samples and extending it to incorporate other cancer tissues.
Collapse
Affiliation(s)
| | - Amoon Jamzad
- School of Computing, Queen's University, Ontario, Canada
| | - Alireza Sedghi
- School of Computing, Queen's University, Ontario, Canada
| | | | - Kathryn Logan
- Department of Pathology and Molecular Medicine, Queen's University, Ontario, Canada
| | - Julie Wallis
- Department of Pathology and Molecular Medicine, Queen's University, Ontario, Canada
| | - Kevin Y M Ren
- Department of Pathology and Molecular Medicine, Queen's University, Ontario, Canada
| | | | | | - Jay Engel
- Department of Surgery, Queen's University, Ontario, Canada
| | - Doug McKay
- Department of Surgery, Queen's University, Ontario, Canada
| | - Sonal Varma
- Department of Pathology and Molecular Medicine, Queen's University, Ontario, Canada
| | - Ami Wang
- Department of Pathology and Molecular Medicine, Queen's University, Ontario, Canada
| | | | - John F Rudan
- Department of Surgery, Queen's University, Ontario, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Ontario, Canada
| |
Collapse
|
30
|
Sedghi A, O'Donnell LJ, Kapur T, Learned-Miller E, Mousavi P, Wells WM. Image registration: Maximum likelihood, minimum entropy and deep learning. Med Image Anal 2021; 69:101939. [PMID: 33388458 PMCID: PMC8046343 DOI: 10.1016/j.media.2020.101939] [Citation(s) in RCA: 6] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/02/2020] [Revised: 11/28/2020] [Accepted: 12/07/2020] [Indexed: 01/12/2023]
Abstract
In this work, we propose a theoretical framework based on maximum profile likelihood for pairwise and groupwise registration. By an asymptotic analysis, we demonstrate that maximum profile likelihood registration minimizes an upper bound on the joint entropy of the distribution that generates the joint image data. Further, we derive the congealing method for groupwise registration by optimizing the profile likelihood in closed form, and using coordinate ascent, or iterative model refinement. We also describe a method for feature based registration in the same framework and demonstrate it on groupwise tractographic registration. In the second part of the article, we propose an approach to deep metric registration that implements maximum likelihood registration using deep discriminative classifiers. We show further that this approach can be used for maximum profile likelihood registration to discharge the need for well-registered training data, using iterative model refinement. We demonstrate that the method succeeds on a challenging registration problem where the standard mutual information approach does not perform well.
Collapse
Affiliation(s)
- Alireza Sedghi
- Medical Informatics Laboratory, Queen's University, Kingston, Canada.
| | - Lauren J O'Donnell
- Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, USA
| | - Tina Kapur
- Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, USA
| | - Erik Learned-Miller
- College of Information and Computer Sciences, University of Massachusetts, Amherst, USA
| | - Parvin Mousavi
- Medical Informatics Laboratory, Queen's University, Kingston, Canada
| | - William M Wells
- Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, USA; Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Boston, USA
| |
Collapse
|
31
|
Jamzad A, Jamaspishvili T, Iseman R, Kaufmann M, Berman D, Mousavi P. Abstract PO-005: An efficient digitized annotation platform for pathology-oriented dataset generation in AI research. Clin Cancer Res 2021. [DOI: 10.1158/1557-3265.adi21-po-005] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/16/2022]
Abstract
Abstract
BACKGROUND: In recent years AI and deep learning have transformed the ability to use large amounts of medical data to augment diagnosis and prognosis processes for cancer. For developing AI methodology, histopathologic assessment serves as the gold standard “labels”, enabling investigators to finely map (or annotate) biologically and clinically important features. Yet correlating high dimensional data (radiomic, morphometric, genomic, metabolomic, etc.) with expert histopathologic diagnosis for dataset generation remains a major challenge. CHALLENGE: Traditionally, labels have been extracted from a snapshot that contains all of the annotation layers overlaid on the original tissue through image processing techniques. This implies the use of distinct colors for annotation, which severely constrain the number of possible labels. Particularly, this is most noticeable for heterogeneous tissues like prostate that require complex annotation. Furthermore, the resolution with which the labels can be mapped is limited by the area of the extracted region. OBJECTIVE: Here we present a workflow for pathology-oriented dataset generation for AI studies that is compatible with standard annotation platforms, and addresses these limitations. We introduce a detailed multi-grade and multi-scale annotation protocol for prostate biopsies. The proposed method is capable of exporting labels as independent layers (representing specific grades of the pathology), and resampling them to the desired resolution. METHODS: A collection of 38 prostate biopsy sections from 19 patients fixed on slides were used. The proposed grading annotation protocol is based on the spatial distribution of cancer cells. Nine layers of annotation were considered depicting stroma, benign tissue, low grade (Gleason pattern 3) and high grade cancer (Gleason patterns 4 and 5), two mixed cancer patterns, prostatic intraepithelial neoplasia (PIN), intraductal carcinoma (IDC), and artifact. The coordinates of the annotation boundaries are post-processed and combined into a label image containing all 9 pathological classes. The metabolomic profiles of the prostate biopsies acquired by desorption electrospray ionization (DESI) is considered for data features in this study. The generated image labels are therefore spatially registered to corresponding DESI data of each slide. RESULTS: The generated dataset through proposed method is used in the application of prostate cancer detection. The dataset is validated through qualitative visualization and quantitative analysis. High correlation is observed between label images of the slides and unsupervised linear representation of corresponding DESI spectra. The pixel-based supervised identification of tissue types based on the DESI also shows high accuracy. CONCLUSION: The proposed digitized pathology annotation protocol and dataset generation workflow is compatible with AI oriented cancer research and is capable of handling large number of pathological classes and high dimensional imaging modalities.
Citation Format: Amoon Jamzad, Tamara Jamaspishvili, Rachael Iseman, Martin Kaufmann, David Berman, Parvin Mousavi. An efficient digitized annotation platform for pathology-oriented dataset generation in AI research [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-005.
Collapse
|
32
|
Rodrigues C, Visram K, Sedghi A, Mousavi P, Siemens DR. Attitudes and experience of urology trainees in interpreting prostate magnetic resonance imaging. Can Urol Assoc J 2020; 15:E293-E298. [PMID: 33119496 DOI: 10.5489/cuaj.6614] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/19/2022]
Abstract
INTRODUCTION Multiparametric magnetic resonance imaging (mpMRI) has resulted in accurate prostate cancer localization and image-guided targeted sampling for biopsy. Despite its more recent uptake, knowledge gaps in interpretation and reporting exist. Our objective was to determine the need for an educational intervention among urology residents working with mpMRIs. METHODS We administered an anonymous, cross-sectional, self-report questionnaire to a convenience sample of urology residents in U.S. and Canadian training programs. The survey included both open- and closed-ended questions employing a five-point Likert scale. It was designed to assess familiarity, exposure, experience, and comfort with interpretation of mpMRI. RESULTS Fifty-three surveys were completed by residents in postgraduate years (PGY) 1-5 and of these, only 12 (23%) reported any formal training in mpMRI interpretation. Most residents' responses demonstrated significant experience with prostate biopsies, as well as familiarity with reviewing mpMRI for these patients. However, mean (± standard deviation [SD]) Likert responses suggested a relatively poor understanding of the components of Prostate Imaging-Reporting and Data System (PI-RADS) v2 scoring for T2-weighted films (2.45±1.01), diffusion-weighted imaging (DWI) films (2.26±0.90), and dynamic contrast-enhanced (DCE) films (2.21±0.99). Similar disagreement scores were observed for questions around interpretation of the different functional techniques of MRI images. Residents reported strong interest (4.21±0.91) in learning opportunities to enhance their ability to interpret mpMRI. CONCLUSIONS While mpMRI of the prostate is a tool frequently used by care teams in teaching centers to identify suspicious prostate cancer lesions, there remain knowledge gaps in the ability of trainees to interpret images and understand PI-RADS v2 scoring. Online modules were suggested to balance the needs of trainee education with the residency workflow.
Collapse
Affiliation(s)
- Craig Rodrigues
- Department of Urology, Queen's University, Kingston, ON, Canada
| | - Kash Visram
- Department of Urology, Queen's University, Kingston, ON, Canada
| | - Alireza Sedghi
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, ON, Canada
| | | |
Collapse
|
33
|
Gerolami J, Wu V, Fauerbach PN, Jabs D, Engel CJ, Rudan J, Merchant S, Walker R, Anas EMA, Abolmaesumi P, Fichtinger G, Ungi T, Mousavi P. An End-to-End Solution for Automatic Contouring of Tumor Region in Intraoperative Images of Breast Lumpectomy. Annu Int Conf IEEE Eng Med Biol Soc 2020; 2020:2003-2006. [PMID: 33018396 DOI: 10.1109/embc44109.2020.9176505] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/09/2022]
Abstract
Breast-conserving surgery, also known as lumpectomy, is an early stage breast cancer treatment that aims to spare as much healthy breast tissue as possible. A risk associated with lumpectomy is the presence of cancer positive margins post operation. Surgical navigation has been shown to reduce cancer positive margins but requires manual segmentation of the tumor intraoperatively. In this paper, we propose an end-to-end solution for automatic contouring of breast tumor from intraoperative ultrasound images using two convolutional neural network architectures, the U-Net and residual U-Net. The networks are trained on annotated intraoperative breast ultrasound images and evaluated on the quality of predicted segmentations. This work brings us one step closer to providing surgeons with an automated surgical navigation system that helps reduce cancer-positive margins during lumpectomy.
Collapse
|
34
|
Gerolami J, Jamzad A, Li SJ, Bayat S, Abolmaesumi P, Mousavi P. Soft Tissue Characterization with Temporal Enhanced Ultrasound through Periodic Manipulation of Point Spread Function: A Feasibility Study. Annu Int Conf IEEE Eng Med Biol Soc 2020; 2020:78-81. [PMID: 33017935 DOI: 10.1109/embc44109.2020.9175991] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/11/2023]
Abstract
Temporal enhanced ultrasound (TeUS) is a tissue characterization approach based on analysis of a temporal series of US data. Previously we demonstrated that intrinsic or external micro-motions of scatterers in the tissue contribute towards the tissue classification properties of TeUS. This property is beneficial to detect early stage cancer, for example, where changes in nuclei configuration (scatteres) dominate tissue properties. In this study, we propose an analytical derivation and experiments to acquire TeUS through manipulation of US imaging parameters, which may be simpler to translate to clinical applications. The feasibility of the proposed method is demonstrated on tissue-mimicking phantoms. Using an autoencoder classifier, we are able to classify phantoms of varying elasticities and scattering sizes.
Collapse
|
35
|
Janssen NNY, Kaufmann M, Santilli A, Jamzad A, Vanderbeck K, Ren KYM, Ungi T, Mousavi P, Rudan JF, McKay D, Wang A, Fichtinger G. Navigated tissue characterization during skin cancer surgery. Int J Comput Assist Radiol Surg 2020; 15:1665-1672. [PMID: 32476078 DOI: 10.1007/s11548-020-02200-4] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/11/2020] [Accepted: 05/18/2020] [Indexed: 10/24/2022]
Abstract
PURPOSE Basal cell carcinoma (BCC) is the most commonly diagnosed skin cancer and is treated by surgical resection. Incomplete tumor removal requires surgical revision, leading to significant healthcare costs and impaired cosmesis. We investigated the clinical feasibility of a surgical navigation system for BCC surgery, based on molecular tissue characterization using rapid evaporative ionization mass spectrometry (REIMS). METHODS REIMS enables direct tissue characterization by analysis of cell-specific molecules present within surgical smoke, produced during electrocautery tissue resection. A tissue characterization model was built by acquiring REIMS spectra of BCC, healthy skin and fat from ex vivo skin cancer specimens. This model was used for tissue characterization during navigated skin cancer surgery. Navigation was enabled by optical tracking and real-time visualization of the cautery relative to a contoured resection volume. The surgical smoke was aspirated into a mass spectrometer and directly analyzed with REIMS. Classified BCC was annotated at the real-time position of the cautery. Feasibility of the navigation system, and tissue classification accuracy for ex vivo and intraoperative surgery were evaluated. RESULTS Fifty-four fresh excision specimens were used to build the ex vivo model of BCC, normal skin and fat, with 92% accuracy. While 3 surgeries were successfully navigated without breach of sterility, the intraoperative performance of the ex vivo model was low (< 50%). Hypotheses are: (1) the model was trained on heterogeneous mass spectra that did not originate from a single tissue type, (2) during surgery mixed tissue types were resected and thus presented to the model, and (3) the mass spectra were not validated by pathology. CONCLUSION REIMS-navigated skin cancer surgery has the potential to detect and localize remaining tumor intraoperatively. Future work will be focused on improving our model by using a precise pencil cautery tip for burning localized tissue types, and having pathology-validated mass spectra.
Collapse
Affiliation(s)
| | - Martin Kaufmann
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Alice Santilli
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Amoon Jamzad
- School of Computing, Queen's University, Kingston, ON, Canada
| | | | - Kevin Yi Mi Ren
- Department of Pathology, Queen's University, Kingston, ON, Canada
| | - Tamas Ungi
- School of Computing, Queen's University, Kingston, ON, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, Kingston, ON, Canada
| | - John F Rudan
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Doug McKay
- Department of Surgery, Queen's University, Kingston, ON, Canada
| | - Ami Wang
- Department of Pathology, Queen's University, Kingston, ON, Canada
| | | |
Collapse
|
36
|
Sedghi A, Mehrtash A, Jamzad A, Amalou A, Wells WM, Kapur T, Kwak JT, Turkbey B, Choyke P, Pinto P, Wood B, Xu S, Abolmaesumi P, Mousavi P. Improving detection of prostate cancer foci via information fusion of MRI and temporal enhanced ultrasound. Int J Comput Assist Radiol Surg 2020; 15:1215-1223. [PMID: 32372384 DOI: 10.1007/s11548-020-02172-5] [Citation(s) in RCA: 16] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/18/2019] [Accepted: 04/16/2020] [Indexed: 11/28/2022]
Abstract
PURPOSE The detection of clinically significant prostate cancer (PCa) is shown to greatly benefit from MRI-ultrasound fusion biopsy, which involves overlaying pre-biopsy MRI volumes (or targets) with real-time ultrasound images. In previous literature, machine learning models trained on either MRI or ultrasound data have been proposed to improve biopsy guidance and PCa detection. However, quantitative fusion of information from MRI and ultrasound has not been explored in depth in a large study. This paper investigates information fusion approaches between MRI and ultrasound to improve targeting of PCa foci in biopsies. METHODS We build models of fully convolutional networks (FCN) using data from a newly proposed ultrasound modality, temporal enhanced ultrasound (TeUS), and apparent diffusion coefficient (ADC) from 107 patients with 145 biopsy cores. The architecture of our models is based on U-Net and U-Net with attention gates. Models are built using joint training through intermediate and late fusion of the data. We also build models with data from each modality, separately, to use as baseline. The performance is evaluated based on the area under the curve (AUC) for predicting clinically significant PCa. RESULTS Using our proposed deep learning framework and intermediate fusion, integration of TeUS and ADC outperforms the individual modalities for cancer detection. We achieve an AUC of 0.76 for detection of all PCa foci, and 0.89 for PCa with larger foci. Results indicate a shared representation between multiple modalities outperforms the average unimodal predictions. CONCLUSION We demonstrate the significant potential of multimodality integration of information from MRI and TeUS to improve PCa detection, which is essential for accurate targeting of cancer foci during biopsy. By using FCNs as the architecture of choice, we are able to predict the presence of clinically significant PCa in entire imaging planes immediately, without the need for region-based analysis. This reduces the overall computational time and enables future intra-operative deployment of this technology.
Collapse
Affiliation(s)
| | - Alireza Mehrtash
- The University of British Columbia, Vancouver, BC, Canada.,Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA
| | | | - Amel Amalou
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | - William M Wells
- Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA
| | - Tina Kapur
- Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA
| | | | - Baris Turkbey
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | - Peter Choyke
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | - Peter Pinto
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | - Bradford Wood
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | - Sheng Xu
- The National Institutes of Health Research Center, Baltimore, MD, USA
| | | | | |
Collapse
|
37
|
Javadi G, Samadi S, Bayat S, Pesteie M, Jafari MH, Sojoudi S, Kesch C, Hurtado A, Chang S, Mousavi P, Black P, Abolmaesumi P. Multiple instance learning combined with label invariant synthetic data for guiding systematic prostate biopsy: a feasibility study. Int J Comput Assist Radiol Surg 2020; 15:1023-1031. [PMID: 32356095 DOI: 10.1007/s11548-020-02168-1] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/18/2019] [Accepted: 04/10/2020] [Indexed: 10/24/2022]
Abstract
PURPOSE Ultrasound imaging is routinely used in prostate biopsy, which involves obtaining prostate tissue samples using a systematic, yet, non-targeted approach. This approach is blinded to individual patient intraprostatic pathology, and unfortunately, has a high rate of false negatives. METHODS In this paper, we propose a deep network for improved detection of prostate cancer in systematic biopsy. We address several challenges associated with training such network: (1) Statistical labels: Since biopsy core's pathology report only represents a statistical distribution of cancer within the core, we use multiple instance learning (MIL) networks to enable learning from ultrasound image regions associated with those data; (2) Limited labels: The number of biopsy cores are limited to at most 12 per patient. As a result, the number of samples available for training a deep network is limited. We alleviate this issue by effectively combining Independent Conditional Variational Auto Encoders (ICVAE) with MIL. We train ICVAE to learn label-invariant features of RF data, which is subsequently used to generate synthetic data for improved training of the MIL network. RESULTS Our in vivo study includes data from 339 prostate biopsy cores of 70 patients. We achieve an area under the curve, sensitivity, specificity, and balanced accuracy of 0.68, 0.77, 0.55 and 0.66, respectively. CONCLUSION The proposed approach is generic and can be applied to several other scenarios where unlabeled data and noisy labels in training samples are present.
Collapse
Affiliation(s)
- Golara Javadi
- The University of British Columbia, Vancouver, BC, Canada.
| | - Samareh Samadi
- The University of British Columbia, Vancouver, BC, Canada
| | - Sharareh Bayat
- The University of British Columbia, Vancouver, BC, Canada
| | - Mehran Pesteie
- The University of British Columbia, Vancouver, BC, Canada
| | | | - Samira Sojoudi
- The University of British Columbia, Vancouver, BC, Canada
| | | | | | - Silvia Chang
- Vancouver General Hospital, Vancouver, BC, Canada
| | | | - Peter Black
- Vancouver General Hospital, Vancouver, BC, Canada
| | | |
Collapse
|
38
|
Chan B, Rudan JF, Mousavi P, Kunz M. Intraoperative integration of structured light scanning for automatic tissue classification: a feasibility study. Int J Comput Assist Radiol Surg 2020; 15:641-649. [PMID: 32144629 DOI: 10.1007/s11548-020-02129-8] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/21/2019] [Accepted: 02/17/2020] [Indexed: 11/26/2022]
Abstract
PURPOSE Structured light scanning is a promising inexpensive and accurate intraoperative imaging modality. Integration of these scanners in surgical workflows has the potential to enable rapid registration and augment preoperative imaging, in a practical and timely manner in the operating theatre. Previously, we have demonstrated the intraoperative feasibility of such scanners to capture anatomical surface information with high accuracy. The purpose of this study was to investigate the feasibility of automatically characterizing anatomical tissues from textural and spatial information captured by such scanners using machine learning. Assisted or automatic identification of relevant components of a captured scan is essential for effective integration of the technology in surgical workflow. METHODS During a clinical study, 3D surface scans for seven total knee arthroplasty patients were collected, and textural and spatial features for cartilage, bone, and ligament tissue were collected and annotated. These features were used to train and evaluate machine learning models. As part of our preliminary preparation, three fresh-frozen knee cadaver specimens were also used where 3D surface scans with texture information were collected during different dissection stages. The resulting models were manually segmented to isolate texture information for muscles, tendon, cartilage, and bone. This information, and detailed labels from dissections, provided an in-depth, finely annotated dataset for building machine learning classifiers. RESULTS For characterizing bone, cartilage, and ligament in the intraoperative surface models, random forest and neural network-based models achieved an accuracy of close to 80%, whereas an accuracy of close to 90% was obtained when only characterizing bone and cartilage. Average accuracy of 76-82% was reached for cadaver data in two-, three-, and four-class tissue separation. CONCLUSIONS The results of this project demonstrate the feasibility of machine learning methods to accurately classify multiple types of anatomical tissue. The ability to automatically characterize tissues in intraoperatively collected surface models would streamline the surgical workflow of using structured light scanners-paving the way to applications such as 3D documentation of surgery in addition to rapid registration and augmentation of preoperative imaging.
Collapse
Affiliation(s)
- Brandon Chan
- School of Computing, Queen's University, 557 Goodwin Hall, Kingston, ON, K7L 2N8, Canada
| | - John F Rudan
- Department of Surgery, Kingston Health Sciences Centre, Queen's University, 76 Stuart Street, Kingston, ON, K7L 2V7, Canada
| | - Parvin Mousavi
- School of Computing, Queen's University, 557 Goodwin Hall, Kingston, ON, K7L 2N8, Canada.
| | - Manuela Kunz
- School of Computing, Queen's University, 557 Goodwin Hall, Kingston, ON, K7L 2N8, Canada.
- National Research Council Canada, 1200 Montreal Rd, Building M-50, Ottawa, ON, K1A 0R6, Canada.
| |
Collapse
|
39
|
Herz C, MacNeil K, Behringer PA, Tokuda J, Mehrtash A, Mousavi P, Kikinis R, Fennessy FM, Tempany CM, Tuncali K, Fedorov A. Open Source Platform for Transperineal In-Bore MRI-Guided Targeted Prostate Biopsy. IEEE Trans Biomed Eng 2020; 67:565-576. [PMID: 31135342 PMCID: PMC6874712 DOI: 10.1109/tbme.2019.2918731] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/10/2022]
Abstract
OBJECTIVE Accurate biopsy sampling of the suspected lesions is critical for the diagnosis and clinical management of prostate cancer. Transperineal in-bore MRI-guided prostate biopsy (tpMRgBx) is a targeted biopsy technique that was shown to be safe, efficient, and accurate. Our goal was to develop an open source software platform to support evaluation, refinement, and translation of this biopsy approach. METHODS We developed SliceTracker, a 3D Slicer extension to support tpMRgBx. We followed modular design of the implementation to enable customization of the interface and interchange of image segmentation and registration components to assess their effect on the processing time, precision, and accuracy of the biopsy needle placement. The platform and supporting documentation were developed to enable the use of software by an operator with minimal technical training to facilitate translation. Retrospective evaluation studied registration accuracy, effect of the prostate segmentation approach, and re-identification time of biopsy targets. Prospective evaluation focused on the total procedure time and biopsy targeting error (BTE). RESULTS Evaluation utilized data from 73 retrospective and ten prospective tpMRgBx cases. Mean landmark registration error for retrospective evaluation was 1.88 ± 2.63 mm, and was not sensitive to the approach used for prostate gland segmentation. Prospectively, we observed target re-identification time of 4.60 ± 2.40 min and BTE of 2.40 ± 0.98 mm. CONCLUSION SliceTracker is modular and extensible open source platform for supporting image processing aspects of the tpMRgBx procedure. It has been successfully utilized to support clinical research procedures at our site.
Collapse
|
40
|
Chan B, Sedghi A, Laird P, Maslove D, Mousavi P. Prediction of Patient-specific Acute Hypotensive Episodes in ICU Using Deep Models. Annu Int Conf IEEE Eng Med Biol Soc 2019; 2019:566-569. [PMID: 31945962 DOI: 10.1109/embc.2019.8856985] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/10/2023]
Abstract
Forecasting acute hypotensive episodes (AHE) in intensive care patients has been of recent interest to researchers in the healthcare domain. Advance warning of an impending AHE may give care providers additional information to help mitigate the negative clinical impact of a serious event such as an AHE or prompt a search for an evolving disease process. However, the currently accepted definition of AHE is restrictive does not account for inter-patient variability. In this paper, we propose a novel definition of an AHE based on patient-specific features of blood pressure recordings. Next, we utilize a deep learning-based method to predict the onset of an AHE from multiple physiological readings for different definitions of the prediction task including variable input and gap lengths. Using a cohort of 538 patients, our model was able to successfully predict the onset of an AHE with an accuracy and AUC score of 0.80 and 0.87 respectively. Compared to a baseline logistic regression model, our model outperforms the baseline in most of the definitions of the prediction task.
Collapse
|
41
|
Chan B, Sedghi A, Laird P, Maslove D, Mousavi P. Predictive Modeling using Intensive Care Unit Data: Considerations for Data Pre-processing and Analysis. Annu Int Conf IEEE Eng Med Biol Soc 2019; 2019:3429-3432. [PMID: 31946616 DOI: 10.1109/embc.2019.8857564] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/10/2023]
Abstract
The vast quantities of data generated and collected in the Intensive Care Unit (ICU) have given rise to large retrospective datasets that are frequently used for observational studies. The temporal nature and fine granularity of much of the data collected in the ICU enable the pursuit of predictive modeling, an increasingly common topic in ICU literature. Since patient conditions can rapidly change in the ICU, predicting the onset of events that are indicative of deteriorating patient state has potential clinical utility. However, the development of predictive modeling applications using ICU data requires a number of considerations to maximize prospective performance and clinical utility. In this study, we discuss the challenges encountered and considerations taken by using the prediction of acute hypotensive episodes as an example.
Collapse
|
42
|
Pyman B, Sedghi A, Azizi S, Tyryshkin K, Renwick N, Mousavi P. Exploring microRNA Regulation of Cancer with Context-Aware Deep Cancer Classifier. Pac Symp Biocomput 2019; 24:160-171. [PMID: 30864319] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Subscribe] [Scholar Register] [Indexed: 06/09/2023]
Abstract
BACKGROUND MicroRNAs (miRNAs) are small, non-coding RNA that regulate gene expression through post-transcriptional silencing. Differential expression observed in miRNAs, combined with advancements in deep learning (DL), have the potential to improve cancer classification by modelling non-linear miRNA-phenotype associations. We propose a novel miRNA-based deep cancer classifier (DCC) incorporating genomic and hierarchical tissue annotation, capable of accurately predicting the presence of cancer in wide range of human tissues. METHODS miRNA expression profiles were analyzed for 1746 neoplastic and 3871 normal samples, across 26 types of cancer involving six organ sub-structures and 68 cell types. miRNAs were ranked and filtered using a specificity score representing their information content in relation to neoplasticity, incorporating 3 levels of hierarchical biological annotation. A DL architecture composed of stacked autoencoders (AE) and a multi-layer perceptron (MLP) was trained to predict neoplasticity using 497 abundant and informative miRNAs. Additional DCCs were trained using expression of miRNA cistrons and sequence families, and combined as a diagnostic ensemble. Important miRNAs were identified using backpropagation, and analyzed in Cytoscape using iCTNet and BiNGO. RESULTS Nested four-fold cross-validation was used to assess the performance of the DL model. The model achieved an accuracy, AUC/ROC, sensitivity, and specificity of 94.73%, 98.6%, 95.1%, and 94.3%, respectively. CONCLUSION Deep autoencoder networks are a powerful tool for modelling complex miRNA-phenotype associations in cancer. The proposed DCC improves classification accuracy by learning from the biological context of both samples and miRNAs, using anatomical and genomic annotation. Analyzing the deep structure of DCCs with backpropagation can also facilitate biological discovery, by performing gene ontology searches on the most highly significant features.
Collapse
Affiliation(s)
- Blake Pyman
- School of Computing, Queen's University, Kingston, Ontario K7L 3N6, Canada http://www.queensu.ca/,
| | | | | | | | | | | |
Collapse
|
43
|
Azizi S, Bayat S, Yan P, Tahmasebi A, Kwak JT, Xu S, Turkbey B, Choyke P, Pinto P, Wood B, Mousavi P, Abolmaesumi P. Deep Recurrent Neural Networks for Prostate Cancer Detection: Analysis of Temporal Enhanced Ultrasound. IEEE Trans Med Imaging 2018; 37:2695-2703. [PMID: 29994471 PMCID: PMC7983161 DOI: 10.1109/tmi.2018.2849959] [Citation(s) in RCA: 21] [Impact Index Per Article: 3.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 05/29/2023]
Abstract
Temporal enhanced ultrasound (TeUS), comprising the analysis of variations in backscattered signals from a tissue over a sequence of ultrasound frames, has been previously proposed as a new paradigm for tissue characterization. In this paper, we propose to use deep recurrent neural networks (RNN) to explicitly model the temporal information in TeUS. By investigating several RNN models, we demonstrate that long short-term memory (LSTM) networks achieve the highest accuracy in separating cancer from benign tissue in the prostate. We also present algorithms for in-depth analysis of LSTM networks. Our in vivo study includes data from 255 prostate biopsy cores of 157 patients. We achieve area under the curve, sensitivity, specificity, and accuracy of 0.96, 0.76, 0.98, and 0.93, respectively. Our result suggests that temporal modeling of TeUS using RNN can significantly improve cancer detection accuracy over previously presented works.
Collapse
|
44
|
Barnes J, Li S, Goyal A, Abolmaesumi P, Mousavi P, Loock HP. Broadband Vibration Detection in Tissue Phantoms Using a Fiber Fabry–Perot Cavity. IEEE Trans Biomed Eng 2018; 65:921-927. [DOI: 10.1109/tbme.2017.2731663] [Citation(s) in RCA: 12] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/07/2022]
|
45
|
Azizi S, Van Woudenberg N, Sojoudi S, Li M, Xu S, Abu Anas EM, Yan P, Tahmasebi A, Kwak JT, Turkbey B, Choyke P, Pinto P, Wood B, Mousavi P, Abolmaesumi P. Toward a real-time system for temporal enhanced ultrasound-guided prostate biopsy. Int J Comput Assist Radiol Surg 2018; 13:1201-1209. [PMID: 29589258 DOI: 10.1007/s11548-018-1749-z] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/30/2018] [Accepted: 03/21/2018] [Indexed: 01/17/2023]
Abstract
PURPOSE We have previously proposed temporal enhanced ultrasound (TeUS) as a new paradigm for tissue characterization. TeUS is based on analyzing a sequence of ultrasound data with deep learning and has been demonstrated to be successful for detection of cancer in ultrasound-guided prostate biopsy. Our aim is to enable the dissemination of this technology to the community for large-scale clinical validation. METHODS In this paper, we present a unified software framework demonstrating near-real-time analysis of ultrasound data stream using a deep learning solution. The system integrates ultrasound imaging hardware, visualization and a deep learning back-end to build an accessible, flexible and robust platform. A client-server approach is used in order to run computationally expensive algorithms in parallel. We demonstrate the efficacy of the framework using two applications as case studies. First, we show that prostate cancer detection using near-real-time analysis of RF and B-mode TeUS data and deep learning is feasible. Second, we present real-time segmentation of ultrasound prostate data using an integrated deep learning solution. RESULTS The system is evaluated for cancer detection accuracy on ultrasound data obtained from a large clinical study with 255 biopsy cores from 157 subjects. It is further assessed with an independent dataset with 21 biopsy targets from six subjects. In the first study, we achieve area under the curve, sensitivity, specificity and accuracy of 0.94, 0.77, 0.94 and 0.92, respectively, for the detection of prostate cancer. In the second study, we achieve an AUC of 0.85. CONCLUSION Our results suggest that TeUS-guided biopsy can be potentially effective for the detection of prostate cancer.
Collapse
Affiliation(s)
| | | | - Samira Sojoudi
- The University of British Columbia, Vancouver, BC, Canada
| | - Ming Li
- National Institutes of Health, Bethesda, MD, USA
| | - Sheng Xu
- National Institutes of Health, Bethesda, MD, USA
| | | | - Pingkun Yan
- Rensselaer Polytechnic Institute, Troy, NY, USA
| | | | | | | | - Peter Choyke
- National Institutes of Health, Bethesda, MD, USA
| | - Peter Pinto
- National Institutes of Health, Bethesda, MD, USA
| | | | | | | |
Collapse
|
46
|
Azizi S, Rajaram A, Bayat S, Mohamed T, Walus K, Abolmaesumi P, Mousavi P, Anas EMA. 3D tissue mimicking biophantoms for ultrasound imaging: bioprinting and image analysis. Medical Imaging 2018: Image-Guided Procedures, Robotic Interventions, and Modeling 2018. [DOI: 10.1117/12.2293930] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 09/01/2023]
|
47
|
Bayat S, Azizi S, Daoud MI, Nir G, Imani F, Gerardo CD, Yan P, Tahmasebi A, Vignon F, Sojoudi S, Wilson S, Iczkowski KA, Lucia MS, Goldenberg L, Salcudean SE, Abolmaesumi P, Mousavi P. Investigation of Physical Phenomena Underlying Temporal-Enhanced Ultrasound as a New Diagnostic Imaging Technique: Theory and Simulations. IEEE Trans Ultrason Ferroelectr Freq Control 2018; 65:400-410. [PMID: 29505407 DOI: 10.1109/tuffc.2017.2785230] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/08/2023]
Abstract
Temporal-enhanced ultrasound (TeUS) is a novel noninvasive imaging paradigm that captures information from a temporal sequence of backscattered US radio frequency data obtained from a fixed tissue location. This technology has been shown to be effective for classification of various in vivo and ex vivo tissue types including prostate cancer from benign tissue. Our previous studies have indicated two primary phenomena that influence TeUS: 1) changes in tissue temperature due to acoustic absorption and 2) micro vibrations of tissue due to physiological vibration. In this paper, first, a theoretical formulation for TeUS is presented. Next, a series of simulations are carried out to investigate micro vibration as a source of tissue characterizing information in TeUS. The simulations include finite element modeling of micro vibration in synthetic phantoms, followed by US image generation during TeUS imaging. The simulations are performed on two media, a sparse array of scatterers and a medium with pathology mimicking scatterers that match nuclei distribution extracted from a prostate digital pathology data set. Statistical analysis of the simulated TeUS data shows its ability to accurately classify tissue types. Our experiments suggest that TeUS can capture the microstructural differences, including scatterer density, in tissues as they react to micro vibrations.
Collapse
|
48
|
Nahlawi L, Goncalves C, Imani F, Gaed M, Gomez JA, Moussa M, Gibson E, Fenster A, Ward A, Abolmaesumi P, Shatkay H, Mousavi P. Stochastic Modeling of Temporal Enhanced Ultrasound: Impact of Temporal Properties on Prostate Cancer Characterization. IEEE Trans Biomed Eng 2017; 65:1798-1809. [PMID: 29989922 DOI: 10.1109/tbme.2017.2778007] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [What about the content of this article? (0)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/09/2022]
Abstract
OBJECTIVES Temporal enhanced ultrasound (TeUS) is a new ultrasound-based imaging technique that provides tissue-specific information. Recent studies have shown the potential of TeUS for improving tissue characterization in prostate cancer diagnosis. We study the temporal properties of TeUS-temporal order and length-and present a new framework to assess their impact on tissue information. METHODS We utilize a probabilistic modeling approach using hidden Markov models (HMMs) to capture the temporal signatures of malignant and benign tissues from TeUS signals of nine patients. We model signals of benign and malignant tissues (284 and 286 signals, respectively) in their original temporal order as well as under order permutations. We then compare the resulting models using the Kullback-Liebler divergence and assess their performance differences in characterization. Moreover, we train HMMs using TeUS signals of different durations and compare their model performance when differentiating tissue types. RESULTS Our findings demonstrate that models of order-preserved signals perform statistically significantly better (85% accuracy) in tissue characterization compared to models of order-altered signals (62% accuracy). The performance degrades as more changes in signal order are introduced. Additionally, models trained on shorter sequences perform as accurately as models of longer sequences. CONCLUSION The work presented here strongly indicates that temporal order has substantial impact on TeUS performance; thus, it plays a significant role in conveying tissue-specific information. Furthermore, shorter TeUS signals can relay sufficient information to accurately distinguish between tissue types. SIGNIFICANCE Understanding the impact of TeUS properties facilitates the process of its adopting in diagnostic procedures and provides insights on improving its acquisition.
Collapse
|
49
|
Azizi S, Bayat S, Yan P, Tahmasebi A, Nir G, Kwak JT, Xu S, Wilson S, Iczkowski KA, Lucia MS, Goldenberg L, Salcudean SE, Pinto PA, Wood B, Abolmaesumi P, Mousavi P. Detection and grading of prostate cancer using temporal enhanced ultrasound: combining deep neural networks and tissue mimicking simulations. Int J Comput Assist Radiol Surg 2017; 12:1293-1305. [PMID: 28634789 PMCID: PMC7900902 DOI: 10.1007/s11548-017-1627-0] [Citation(s) in RCA: 25] [Impact Index Per Article: 3.6] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/10/2017] [Accepted: 05/01/2017] [Indexed: 10/19/2022]
Abstract
PURPOSE : Temporal Enhanced Ultrasound (TeUS) has been proposed as a new paradigm for tissue characterization based on a sequence of ultrasound radio frequency (RF) data. We previously used TeUS to successfully address the problem of prostate cancer detection in the fusion biopsies. METHODS : In this paper, we use TeUS to address the problem of grading prostate cancer in a clinical study of 197 biopsy cores from 132 patients. Our method involves capturing high-level latent features of TeUS with a deep learning approach followed by distribution learning to cluster aggressive cancer in a biopsy core. In this hypothesis-generating study, we utilize deep learning based feature visualization as a means to obtain insight into the physical phenomenon governing the interaction of temporal ultrasound with tissue. RESULTS : Based on the evidence derived from our feature visualization, and the structure of tissue from digital pathology, we build a simulation framework for studying the physical phenomenon underlying TeUS-based tissue characterization. CONCLUSION : Results from simulation and feature visualization corroborated with the hypothesis that micro-vibrations of tissue microstructure, captured by low-frequency spectral features of TeUS, can be used for detection of prostate cancer.
Collapse
Affiliation(s)
| | - Sharareh Bayat
- The University of British Columbia, Vancouver, BC, Canada
| | - Pingkun Yan
- Philips Research North America, Cambridge, MA, USA
| | | | - Guy Nir
- The University of British Columbia, Vancouver, BC, Canada
| | - Jin Tae Kwak
- Sejong University, Gwangjin-Gu, Seoul, South Korea
| | - Sheng Xu
- National Institutes of Health, Bethesda, MD, USA
| | | | | | | | | | | | | | | | | | | |
Collapse
|
50
|
Azizi S, Mousavi P, Yan P, Tahmasebi A, Kwak JT, Xu S, Turkbey B, Choyke P, Pinto P, Wood B, Abolmaesumi P. Transfer learning from RF to B-mode temporal enhanced ultrasound features for prostate cancer detection. Int J Comput Assist Radiol Surg 2017; 12:1111-1121. [PMID: 28349507 DOI: 10.1007/s11548-017-1573-x] [Citation(s) in RCA: 20] [Impact Index Per Article: 2.9] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/29/2017] [Accepted: 03/18/2017] [Indexed: 02/06/2023]
Abstract
PURPOSE We present a method for prostate cancer (PCa) detection using temporal enhanced ultrasound (TeUS) data obtained either from radiofrequency (RF) ultrasound signals or B-mode images. METHODS For the first time, we demonstrate that by applying domain adaptation and transfer learning methods, a tissue classification model trained on TeUS RF data (source domain) can be deployed for classification using TeUS B-mode data alone (target domain), where both data are obtained on the same ultrasound scanner. This is a critical step for clinical translation of tissue classification techniques that primarily rely on accessing RF data, since this imaging modality is not readily available on all commercial scanners in clinics. Proof of concept is provided for in vivo characterization of PCa using TeUS B-mode data, where different nonlinear processing filters in the pipeline of the RF to B-mode conversion result in a distribution shift between the two domains. RESULTS Our in vivo study includes data obtained in MRI-guided targeted procedure for prostate biopsy. We achieve comparable area under the curve using TeUS RF and B-mode data for medium to large cancer tumor sizes in biopsy cores (>4 mm). CONCLUSION Our result suggests that the proposed adaptation technique is successful in reducing the divergence between TeUS RF and B-mode data.
Collapse
Affiliation(s)
| | | | - Pingkun Yan
- Philips Research North America, Cambridge, MA, USA
| | | | | | - Sheng Xu
- National Institutes of Health, Bethesda, MD, USA
| | | | - Peter Choyke
- National Institutes of Health, Bethesda, MD, USA
| | - Peter Pinto
- National Institutes of Health, Bethesda, MD, USA
| | | | | |
Collapse
|