Incidence, risk factors, and clinical implications of postoperative blood in or near the resection cavity after glioma surgery

Introduction

Postoperative hematomas that require reoperation are a serious, but uncommon complication to glioma surgery. However, smaller blood volumes are frequently observed, but their clinical significance is less known.

Research question

What are the incidence rates, risk factors, and patient-reported outcomes of all measurable blood in or near the resection cavity on postoperative MRI in diffuse glioma patients?

Material and methods

We manually segmented intradural and extradural blood from early postoperative MRI of 292 diffuse glioma resections. Potential associations between blood volume and tumor characteristics, demographics, and perioperative factors were explored using non-parametric methods. The assessed outcomes were generic and disease-specific patient-reported HRQoL.

Results

Out of the 292 MRI scans included, 184 (63%) had intradural blood, and 212 (73%) had extradural blood in or near the resection cavity. The median blood volumes were 0.4 mL and 3.0 mL, respectively. Intradural blood volume was associated with tumor volume, intraoperative blood loss, and EOR. Extradural blood volume was associated with age and tumor volume. Greater intradural blood volume was associated with less headache and cognitive improvement, but not after adjustments for tumor volume.

Discussion and conclusions

Postoperative blood on early postoperative MRI is common. Intradural blood volumes tend to be larger in patients with larger tumors, more intraoperative blood loss, or undergoing subtotal resections. Extradural blood volumes tend to be larger in younger patients with larger tumors. Postoperative blood in or near the resection cavity that does not require reoperation does not seem to affect HRQoL in diffuse glioma patients.

Growth dynamics of untreated meningiomas

Background

Knowledge about meningioma growth characteristics is needed for developing biologically rational follow-up routines. In this study of untreated meningiomas followed with repeated magnetic resonance imaging (MRI) scans, we studied growth dynamics and explored potential factors associated with tumor growth.

Methods

In a single-center cohort study, we included 235 adult patients with radiologically suspected intracranial meningioma and at least 3 MRI scans during follow-up. Tumors were segmented using an automatic algorithm from contrast-enhanced T1 series, and, if needed, manually corrected. Potential meningioma growth curves were statistically compared: linear, exponential, linear radial, or Gompertzian. Factors associated with growth were explored.

Results

In 235 patients, 1394 MRI scans were carried out in the median 5-year observational period. Of the models tested, a Gompertzian growth curve best described growth dynamics of meningiomas on group level. 59% of the tumors grew, 27% remained stable, and 14% shrunk. Only 13 patients (5%) underwent surgery during the observational period and were excluded after surgery. Tumor size at the time of diagnosis, multifocality, and length of follow-up were associated with tumor growth, whereas age, sex, presence of peritumoral edema, and hyperintense T2-signal were not significant factors.

Conclusions

Untreated meningiomas follow a Gompertzian growth curve, indicating that increasing and potentially doubling subsequent follow-up intervals between MRIs seems biologically reasonable, instead of fixed time intervals. Tumor size at diagnosis is the strongest predictor of future growth, indicating a potential for longer follow-up intervals for smaller tumors. Although most untreated meningiomas grow, few require surgery.

Segmentation of glioblastomas in early post-operative multi-modal MRI with deep neural networks

Extent of resection after surgery is one of the main prognostic factors for patients diagnosed with glioblastoma. To achieve this, accurate segmentation and classification of residual tumor from post-operative MR images is essential. The current standard method for estimating it is subject to high inter- and intra-rater variability, and an automated method for segmentation of residual tumor in early post-operative MRI could lead to a more accurate estimation of extent of resection. In this study, two state-of-the-art neural network architectures for pre-operative segmentation were trained for the task. The models were extensively validated on a multicenter dataset with nearly 1000 patients, from 12 hospitals in Europe and the United States. The best performance achieved was a 61% Dice score, and the best classification performance was about 80% balanced accuracy, with a demonstrated ability to generalize across hospitals. In addition, the segmentation performance of the best models was on par with human expert raters. The predicted segmentations can be used to accurately classify the patients into those with residual tumor, and those with gross total resection.

Raidionics: an open software for pre- and postoperative central nervous system tumor segmentation and standardized reporting

For patients suffering from central nervous system tumors, prognosis estimation, treatment decisions, and postoperative assessments are made from the analysis of a set of magnetic resonance (MR) scans. Currently, the lack of open tools for standardized and automatic tumor segmentation and generation of clinical reports, incorporating relevant tumor characteristics, leads to potential risks from inherent decisions' subjectivity. To tackle this problem, the proposed Raidionics open-source software has been developed, offering both a user-friendly graphical user interface and stable processing backend. The software includes preoperative segmentation models for each of the most common tumor types (i.e., glioblastomas, lower grade gliomas, meningiomas, and metastases), together with one early postoperative glioblastoma segmentation model. Preoperative segmentation performances were quite homogeneous across the four different brain tumor types, with an average Dice around 85% and patient-wise recall and precision around 95%. Postoperatively, performances were lower with an average Dice of 41%. Overall, the generation of a standardized clinical report, including the tumor segmentation and features computation, requires about ten minutes on a regular laptop. The proposed Raidionics software is the first open solution enabling an easy use of state-of-the-art segmentation models for all major tumor types, including preoperative and postsurgical standardized reports.

Treatment of Liver Metastases With Focused Ultrasound and Microbubbles in Patients With Colorectal Cancer Receiving Chemotherapy

OBJECTIVE

Pre-clinical trials have obtained promising results that focused ultrasound (FUS) combined with microbubbles (MBs) increases tumor uptake and the therapeutic effect of drugs. The aims of the study described here were to investigate whether FUS and MBs could improve the effect of chemotherapy in patients with liver metastases from colorectal cancer and to investigate the safety and feasibility of using FUS + MBs.

METHODS

We included 17 patients with liver metastases from colorectal cancer, selected two lesions in each patient's liver and randomized the lesions for, respectively, treatment with FUS + MBs or control. After chemotherapy (FOLFIRI or FOLFOXIRI), the lesions were treated with FUS (frequency = 1.67 MHz, mechanical index = 0.5, pulse repetition frequency = 0.33 Hz, 33 oscillations, duty cycle = 0.2%-0.4% and MBs (SonoVue) for 35 min). Nine boluses of MBs were injected intravenously at 3.5 min intervals. Patients were scheduled for four cycles of treatment. Changes in the size of metastases were determined from computed tomography images.

RESULTS

Treatment with FUS + MBs is safe at the settings used. There was considerable variation in treatment response between lesions and mixed response between lesions receiving only chemotherapy. There is a tendency toward larger-volume reduction in lesions treated with FUS + MBs compared with control lesions, but a mixed response to chemotherapy and lesion heterogeneity make it difficult to interpret the results.

CONCLUSION

The combination of FUS and MBs is a safe, feasible and available strategy for improving the effect of chemotherapy in cancer patients. Therapeutic effect was not demonstrated in this trial. Multicenter trials with standardized protocols should be performed.

Learning deep abdominal CT registration through adaptive loss weighting and synthetic data generation

PURPOSE

This study aims to explore training strategies to improve convolutional neural network-based image-to-image deformable registration for abdominal imaging.

METHODS

Different training strategies, loss functions, and transfer learning schemes were considered. Furthermore, an augmentation layer which generates artificial training image pairs on-the-fly was proposed, in addition to a loss layer that enables dynamic loss weighting.

RESULTS

Guiding registration using segmentations in the training step proved beneficial for deep-learning-based image registration. Finetuning the pretrained model from the brain MRI dataset to the abdominal CT dataset further improved performance on the latter application, removing the need for a large dataset to yield satisfactory performance. Dynamic loss weighting also marginally improved performance, all without impacting inference runtime.

CONCLUSION

Using simple concepts, we improved the performance of a commonly used deep image registration architecture, VoxelMorph. In future work, our framework, DDMR, should be validated on different datasets to further assess its value.

H2G-Net: A multi-resolution refinement approach for segmentation of breast cancer region in gigapixel histopathological images

Over the past decades, histopathological cancer diagnostics has become more complex, and the increasing number of biopsies is a challenge for most pathology laboratories. Thus, development of automatic methods for evaluation of histopathological cancer sections would be of value. In this study, we used 624 whole slide images (WSIs) of breast cancer from a Norwegian cohort. We propose a cascaded convolutional neural network design, called H2G-Net, for segmentation of breast cancer region from gigapixel histopathological images. The design involves a detection stage using a patch-wise method, and a refinement stage using a convolutional autoencoder. To validate the design, we conducted an ablation study to assess the impact of selected components in the pipeline on tumor segmentation. Guiding segmentation, using hierarchical sampling and deep heatmap refinement, proved to be beneficial when segmenting the histopathological images. We found a significant improvement when using a refinement network for post-processing the generated tumor segmentation heatmaps. The overall best design achieved a Dice similarity coefficient of 0.933±0.069 on an independent test set of 90 WSIs. The design outperformed single-resolution approaches, such as cluster-guided, patch-wise high-resolution classification using MobileNetV2 (0.872±0.092) and a low-resolution U-Net (0.874±0.128). In addition, the design performed consistently on WSIs across all histological grades and segmentation on a representative × 400 WSI took ~ 58 s, using only the central processing unit. The findings demonstrate the potential of utilizing a refinement network to improve patch-wise predictions. The solution is efficient and does not require overlapping patch inference or ensembling. Furthermore, we showed that deep neural networks can be trained using a random sampling scheme that balances on multiple different labels simultaneously, without the need of storing patches on disk. Future work should involve more efficient patch generation and sampling, as well as improved clustering.

P11.37.B When to resect or biopsy for patients with supratentorial glioblastoma: a multivariable prediction model

Background

The prospects of a patient with suspected glioblastoma may rely heavily on the indication for surgical resection versus biopsy only. Biopsy percentages vary considerably across hospitals and guidelines for treatment of glioblastoma lack criteria for surgical decision-making. To identify patient and tumor characteristics associated with the decision to resect or biopsy a glioblastoma and to develop and validate a prediction model for decision support.

Material and Methods

Clinical data and pre-operative MRI scans were collected for adults who underwent first-time surgery for supratentorial glioblastoma from a registry-based cohort study of 12 hospitals from the Netherlands, Germany, France, Italy, and the United States between 1st of January 2007 and 31st of December 2011. The main outcome was the type of surgical procedure: surgical resection or biopsy only. Predictors were patient- and tumor-related characteristics. Radiological factors were extracted from MRI using an automated tumor segmentation method. A prediction model was constructed using multivariable logistic regression analysis. The model was cross-validated and externally validated with a leave-one-hospital-out approach.

Results

Out of 1053 patients treated for glioblastoma, 28% underwent biopsy only. Biopsy rates varied from 15-40% across hospitals. The prediction model showed excellent discrimination with an average area under the curve of 0.86. Of the patient-related characteristics, younger age was associated more with resection and Karnofsky Performance Score of 60 or less with biopsy. Of the tumor-related characteristics, a location in the right hemisphere, unifocality, no tumor midline crossing, and no involvement of the cortical spinal tract, were associated with resection, as well as a high expected resectability index, a location in the right occipital lobe, and a higher percentage of tumor in Schaefer’s dorsal or ventral attention, limbic, and default networks. External validation proved acceptable to outstanding discrimination with areas under the curve ranging between 0.79 and 0.92 for hospitals.

Conclusion

A prediction model is presented and validated to support the decision to resect or to biopsy a patient with a suspected supratentorial glioblastoma. In this prediction model, tumor-related characteristics were more informative than patient-related factors. This may support surgical decision-making for individual patients, or facilitate comparisons of patient cohorts between surgeons or institutions.

Preoperative Brain Tumor Imaging: Models and Software for Segmentation and Standardized Reporting

For patients suffering from brain tumor, prognosis estimation and treatment decisions are made by a multidisciplinary team based on a set of preoperative MR scans. Currently, the lack of standardized and automatic methods for tumor detection and generation of clinical reports, incorporating a wide range of tumor characteristics, represents a major hurdle. In this study, we investigate the most occurring brain tumor types: glioblastomas, lower grade gliomas, meningiomas, and metastases, through four cohorts of up to 4,000 patients. Tumor segmentation models were trained using the AGU-Net architecture with different preprocessing steps and protocols. Segmentation performances were assessed in-depth using a wide-range of voxel and patient-wise metrics covering volume, distance, and probabilistic aspects. Finally, two software solutions have been developed, enabling an easy use of the trained models and standardized generation of clinical reports: Raidionics and Raidionics-Slicer. Segmentation performances were quite homogeneous across the four different brain tumor types, with an average true positive Dice ranging between 80 and 90%, patient-wise recall between 88 and 98%, and patient-wise precision around 95%. In conjunction to Dice, the identified most relevant other metrics were the relative absolute volume difference, the variation of information, and the Hausdorff, Mahalanobis, and object average symmetric surface distances. With our Raidionics software, running on a desktop computer with CPU support, tumor segmentation can be performed in 16–54 s depending on the dimensions of the MRI volume. For the generation of a standardized clinical report, including the tumor segmentation and features computation, 5–15 min are necessary. All trained models have been made open-access together with the source code for both software solutions and validation metrics computation. In the future, a method to convert results from a set of metrics into a final single score would be highly desirable for easier ranking across trained models. In addition, an automatic classification of the brain tumor type would be necessary to replace manual user input. Finally, the inclusion of post-operative segmentation in both software solutions will be key for generating complete post-operative standardized clinical reports.

Pretreatment patient-reported cognitive function in patients with diffuse glioma

Purpose

Cognitive function is frequently assessed with objective neuropsychological tests, but patient-reported cognitive function is less explored. We aimed to investigate the preoperative prevalence of patient-reported cognitive impairment in patients with diffuse glioma compared to a matched reference group and explore associated factors.

Methods

We included 237 patients with diffuse glioma and 474 age- and gender-matched controls from the general population. Patient-reported cognitive function was measured using the cognitive function subscale in the European Organisation for Research and Treatment of Cancer QLQ-C30 questionnaire. The transformed scale score (0–100) was dichotomized, with a score of ≤ 75 indicating clinically important patient-reported cognitive impairment. Factors associated with preoperative patient-reported cognitive impairment were explored in a multivariable regression analysis.

Results

Cognitive impairment was reported by 49.8% of the diffuse glioma patients and by 23.4% in the age- and gender-matched reference group (p < 0.001). Patients with diffuse glioma had 3.2 times higher odds (95% CI 2.29, 4.58, p < 0.001) for patient-reported cognitive impairment compared to the matched reference group. In the multivariable analysis, large tumor volume, left tumor lateralization, and low Karnofsky Performance Status score were found to be independent predictors for preoperative patient-reported cognitive impairment.

Conclusions

Our findings demonstrate that patient-reported cognitive impairment is a common symptom in patients with diffuse glioma pretreatment, especially in patients with large tumor volumes, left tumor lateralization, and low functional levels. Patient-reported cognitive function may provide important information about patients’ subjective cognitive health and disease status and may serve as a complement to or as a screening variable for subsequent objective testing.

Brain infarctions after glioma surgery: prevalence, radiological characteristics and risk factors

Background

Prevalence, radiological characteristics, and risk factors for peritumoral infarctions after glioma surgery are not much studied. In this study, we assessed shape, volume, and prevalence of peritumoral infarctions and investigated possible associated factors.

Methods

In a prospective single-center cohort study, we included all adult patients operated for diffuse gliomas from January 2007 to December 2018. Postoperative infarctions were segmented using early postoperative MRI images, and volume, shape, and location of postoperative infarctions were assessed. Heatmaps of the distribution of tumors and infarctions were created.

Results

MRIs from 238 (44%) of 539 operations showed restricted diffusion in relation to the operation cavity, interpreted as postoperative infarctions. Of these, 86 (36%) were rim-shaped, 103 (43%) were sector-shaped, 40 (17%) were a combination of rim- and sector-shaped, and six (3%) were remote infarctions. Median infarction volume was 1.7 cm³ (IQR 0.7–4.3, range 0.1–67.1). Infarctions were more common if the tumor was in the temporal lobe, and the map shows more infarctions in the periventricular watershed areas. Sector-shaped infarctions were more often seen in patients with known cerebrovascular disease (47.6% vs. 25.5%, p = 0.024). There was a positive correlation between infarction volume and tumor volume (r = 0.267, p < 0.001) and infarction volume and perioperative bleeding (r = 0.176, p = 0.014). Moreover, there was a significant positive association between age and larger infarction volumes (r = 0.193, p = 0.003). Infarction rates and infarction volumes varied across individual surgeons, p = 0.037 (range 32–72%) and p = 0.026.

Conclusions

In the present study, peritumoral infarctions occurred in 44% after diffuse glioma operations. Infarctions were more common in patients operated for tumors in the temporal lobe but were not more common following recurrent surgeries. Sector-shaped infarctions were more common in patients with known cerebrovascular disease. Increasing age, larger tumors, and more perioperative bleeding were factors associated with infarction volumes. The risk of infarctions and infarction volumes may also be surgeon-dependent.

Are there predilection sites for intracranial meningioma? A population-based atlas

Meningioma is the most common benign intracranial tumor and is believed to arise from arachnoid cap cells of arachnoid granulations. We sought to develop a population-based atlas from pre-treatment MRIs to explore the distribution of intracranial meningiomas and to explore risk factors for development of intracranial meningiomas in different locations. All adults (≥ 18 years old) diagnosed with intracranial meningiomas and referred to the department of neurosurgery from a defined catchment region between 2006 and 2015 were eligible for inclusion. Pre-treatment T1 contrast-enhanced MRI-weighted brain scans were used for semi-automated tumor segmentation to develop the meningioma atlas. Patient variables used in the statistical analyses included age, gender, tumor locations, WHO grade and tumor volume. A total of 602 patients with intracranial meningiomas were identified for the development of the brain tumor atlas from a wide and defined catchment region. The spatial distribution of meningioma within the brain is not uniform, and there were more tumors in the frontal region, especially parasagittally, along the anterior part of the falx, and on the skull base of the frontal and middle cranial fossa. More than 2/3 meningioma patients were females (p < 0.001) who also were more likely to have multiple meningiomas (p < 0.01), while men more often have supratentorial meningiomas (p < 0.01). Tumor location was not associated with age or WHO grade. The distribution of meningioma exhibits an anterior to posterior gradient in the brain. Distribution of meningiomas in the general population is not dependent on histopathological WHO grade, but may be gender-related.

Meningioma Segmentation in T1-Weighted MRI Leveraging Global Context and Attention Mechanisms

Purpose: Meningiomas are the most common type of primary brain tumor, accounting for ~30% of all brain tumors. A substantial number of these tumors are never surgically removed but rather monitored over time. Automatic and precise meningioma segmentation is, therefore, beneficial to enable reliable growth estimation and patient-specific treatment planning. Methods: In this study, we propose the inclusion of attention mechanisms on top of a U-Net architecture used as backbone: (i) Attention-gated U-Net (AGUNet) and (ii) Dual Attention U-Net (DAUNet), using a three-dimensional (3D) magnetic resonance imaging (MRI) volume as input. Attention has the potential to leverage the global context and identify features' relationships across the entire volume. To limit spatial resolution degradation and loss of detail inherent to encoder-decoder architectures, we studied the impact of multi-scale input and deep supervision components. The proposed architectures are trainable end-to-end and each concept can be seamlessly disabled for ablation studies. Results: The validation studies were performed using a five-fold cross-validation over 600 T1-weighted MRI volumes from St. Olavs Hospital, Trondheim University Hospital, Norway. Models were evaluated based on segmentation, detection, and speed performances, and results are reported patient-wise after averaging across all folds. For the best-performing architecture, an average Dice score of 81.6% was reached for an F1-score of 95.6%. With an almost perfect precision of 98%, meningiomas smaller than 3 ml were occasionally missed hence reaching an overall recall of 93%. Conclusion: Leveraging global context from a 3D MRI volume provided the best performances, even if the native volume resolution could not be processed directly due to current GPU memory limitations. Overall, near-perfect detection was achieved for meningiomas larger than 3 ml, which is relevant for clinical use. In the future, the use of multi-scale designs and refinement networks should be further investigated. A larger number of cases with meningiomas below 3 ml might also be needed to improve the performance for the smallest tumors.

Glioblastoma Surgery Imaging-Reporting and Data System: Validation and Performance of the Automated Segmentation Task

For patients with presumed glioblastoma, essential tumor characteristics are determined from preoperative MR images to optimize the treatment strategy. This procedure is time-consuming and subjective, if performed by crude eyeballing or manually. The standardized GSI-RADS aims to provide neurosurgeons with automatic tumor segmentations to extract tumor features rapidly and objectively. In this study, we improved automatic tumor segmentation and compared the agreement with manual raters, describe the technical details of the different components of GSI-RADS, and determined their speed. Two recent neural network architectures were considered for the segmentation task: nnU-Net and AGU-Net. Two preprocessing schemes were introduced to investigate the tradeoff between performance and processing speed. A summarized description of the tumor feature extraction and standardized reporting process is included. The trained architectures for automatic segmentation and the code for computing the standardized report are distributed as open-source and as open-access software. Validation studies were performed on a dataset of 1594 gadolinium-enhanced T1-weighted MRI volumes from 13 hospitals and 293 T1-weighted MRI volumes from the BraTS challenge. The glioblastoma tumor core segmentation reached a Dice score slightly below 90%, a patientwise F1-score close to 99%, and a 95th percentile Hausdorff distance slightly below 4.0 mm on average with either architecture and the heavy preprocessing scheme. A patient MRI volume can be segmented in less than one minute, and a standardized report can be generated in up to five minutes. The proposed GSI-RADS software showed robust performance on a large collection of MRI volumes from various hospitals and generated results within a reasonable runtime.

Glioblastoma Surgery Imaging-Reporting and Data System: Standardized Reporting of Tumor Volume, Location, and Resectability Based on Automated Segmentations

Treatment decisions for patients with presumed glioblastoma are based on tumor characteristics available from a preoperative MR scan. Tumor characteristics, including volume, location, and resectability, are often estimated or manually delineated. This process is time consuming and subjective. Hence, comparison across cohorts, trials, or registries are subject to assessment bias. In this study, we propose a standardized Glioblastoma Surgery Imaging Reporting and Data System (GSI-RADS) based on an automated method of tumor segmentation that provides standard reports on tumor features that are potentially relevant for glioblastoma surgery. As clinical validation, we determine the agreement in extracted tumor features between the automated method and the current standard of manual segmentations from routine clinical MR scans before treatment. In an observational consecutive cohort of 1596 adult patients with a first time surgery of a glioblastoma from 13 institutions, we segmented gadolinium-enhanced tumor parts both by a human rater and by an automated algorithm. Tumor features were extracted from segmentations of both methods and compared to assess differences, concordance, and equivalence. The laterality, contralateral infiltration, and the laterality indices were in excellent agreement. The native and normalized tumor volumes had excellent agreement, consistency, and equivalence. Multifocality, but not the number of foci, had good agreement and equivalence. The location profiles of cortical and subcortical structures were in excellent agreement. The expected residual tumor volumes and resectability indices had excellent agreement, consistency, and equivalence. Tumor probability maps were in good agreement. In conclusion, automated segmentations are in excellent agreement with manual segmentations and practically equivalent regarding tumor features that are potentially relevant for neurosurgical purposes. Standard GSI-RADS reports can be generated by open access software.

Fast meningioma segmentation in T1-weighted magnetic resonance imaging volumes using a lightweight 3D deep learning architecture

Purpose: Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large number of both surgically treated meningiomas and untreated meningiomas followed at the outpatient clinic. Approach: We studied two different three-dimensional (3D) neural network architectures: (i) a simple encoder-decoder similar to a 3D U-Net, and (ii) a lightweight multi-scale architecture [Pulmonary Lobe Segmentation Network (PLS-Net)]. In addition, we studied the impact of different training schemes. For the validation studies, we used 698 T1-weighted MR volumes from St. Olav University Hospital, Trondheim, Norway. The models were evaluated in terms of detection accuracy, segmentation accuracy, and training/inference speed. Results: While both architectures reached a similar Dice score of 70% on average, the PLS-Net was more accurate with an F 1 -score of up to 88%. The highest accuracy was achieved for the largest meningiomas. Speed-wise, the PLS-Net architecture tended to converge in about 50 h while 130 h were necessary for U-Net. Inference with PLS-Net takes less than a second on GPU and about 15 s on CPU. Conclusions: Overall, with the use of mixed precision training, it was possible to train competitive segmentation models in a relatively short amount of time using the lightweight PLS-Net architecture. In the future, the focus should be brought toward the segmentation of small meningiomas ( < 2    ml ) to improve clinical relevance for automatic and early diagnosis and speed of growth estimates.

Spatial distribution of malignant transformation in patients with low-grade glioma

Background

Malignant transformation represents the natural evolution of diffuse low-grade gliomas (LGG). This is a catastrophic event, causing neurocognitive symptoms, intensified treatment and premature death. However, little is known concerning the spatial distribution of malignant transformation in patients with LGG.

Materials and methods

Patients histopathological diagnosed with LGG and subsequent radiological malignant transformation were identified from two different institutions. We evaluated the spatial distribution of malignant transformation with (1) visual inspection and (2) segmentations of longitudinal tumor volumes. In (1) a radiological transformation site < 2 cm from the tumor on preceding MRI was defined local transformation. In (2) overlap with pretreatment volume after importation into a common space was defined as local transformation. With a centroid model we explored if there were particular patterns of transformations within relevant subgroups.

Results

We included 43 patients in the clinical evaluation, and 36 patients had MRIs scans available for longitudinal segmentations. Prior to malignant transformation, residual radiological tumor volumes were > 10 ml in 93% of patients. The transformation site was considered local in 91% of patients by clinical assessment. Patients treated with radiotherapy prior to transformation had somewhat lower rate of local transformations (83%). Based upon the segmentations, the transformation was local in 92%. We did not observe any particular pattern of transformations in examined molecular subgroups.

Conclusion

Malignant transformation occurs locally and within the T2w hyperintensities in most patients. Although LGG is an infiltrating disease, this data conceptually strengthens the role of loco-regional treatments in patients with LGG.

Vision-based and marker-less surgical tool detection and tracking: a review of the literature

In recent years, tremendous progress has been made in surgical practice for example with Minimally Invasive Surgery (MIS). To overcome challenges coming from deported eye-to-hand manipulation, robotic and computer-assisted systems have been developed. Having real-time knowledge of the pose of surgical tools with respect to the surgical camera and underlying anatomy is a key ingredient for such systems. In this paper, we present a review of the literature dealing with vision-based and marker-less surgical tool detection. This paper includes three primary contributions: (1) identification and analysis of data-sets used for developing and testing detection algorithms, (2) in-depth comparison of surgical tool detection methods from the feature extraction process to the model learning strategy and highlight existing shortcomings, and (3) analysis of validation techniques employed to obtain detection performance results and establish comparison between surgical tool detectors. The papers included in the review were selected through PubMed and Google Scholar searches using the keywords: "surgical tool detection", "surgical tool tracking", "surgical instrument detection" and "surgical instrument tracking" limiting results to the year range 2000 2015. Our study shows that despite significant progress over the years, the lack of established surgical tool data-sets, and reference format for performance assessment and method ranking is preventing faster improvement.

Automatic data-driven real-time segmentation and recognition of surgical workflow

PURPOSE

With the intention of extending the perception and action of surgical staff inside the operating room, the medical community has expressed a growing interest towards context-aware systems. Requiring an accurate identification of the surgical workflow, such systems make use of data from a diverse set of available sensors. In this paper, we propose a fully data-driven and real-time method for segmentation and recognition of surgical phases using a combination of video data and instrument usage signals, exploiting no prior knowledge. We also introduce new validation metrics for assessment of workflow detection.

METHODS

The segmentation and recognition are based on a four-stage process. Firstly, during the learning time, a Surgical Process Model is automatically constructed from data annotations to guide the following process. Secondly, data samples are described using a combination of low-level visual cues and instrument information. Then, in the third stage, these descriptions are employed to train a set of AdaBoost classifiers capable of distinguishing one surgical phase from others. Finally, AdaBoost responses are used as input to a Hidden semi-Markov Model in order to obtain a final decision.

RESULTS

On the MICCAI EndoVis challenge laparoscopic dataset we achieved a precision and a recall of 91 % in classification of 7 phases.

CONCLUSION

Compared to the analysis based on one data type only, a combination of visual features and instrument signals allows better segmentation, reduction of the detection delay and discovery of the correct phase order.

Detecting Surgical Tools by Modelling Local Appearance and Global Shape

Detecting tools in surgical videos is an important ingredient for context-aware computer-assisted surgical systems. To this end, we present a new surgical tool detection dataset and a method for joint tool detection and pose estimation in 2d images. Our two-stage pipeline is data-driven and relaxes strong assumptions made by previous works regarding the geometry, number, and position of tools in the image. The first stage classifies each pixel based on local appearance only, while the second stage evaluates a tool-specific shape template to enforce global shape. Both local appearance and global shape are learned from training data. Our method is validated on a new surgical tool dataset of 2 476 images from neurosurgical microscopes, which is made freely available. It improves over existing datasets in size, diversity and detail of annotation. We show that our method significantly improves over competitive baselines from the computer vision field. We achieve 15% detection miss-rate at 10(-1) false positives per image (for the suction tube) over our surgical tool dataset. Results indicate that performing semantic labelling as an intermediate task is key for high quality detection.

Unsupervised Trajectory Segmentation for Surgical Gesture Recognition in Robotic Training

Dexterity and procedural knowledge are two critical skills that surgeons need to master to perform accurate and safe surgical interventions. However, current training systems do not allow us to provide an in-depth analysis of surgical gestures to precisely assess these skills. Our objective is to develop a method for the automatic and quantitative assessment of surgical gestures. To reach this goal, we propose a new unsupervised algorithm that can automatically segment kinematic data from robotic training sessions. Without relying on any prior information or model, this algorithm detects critical points in the kinematic data that define relevant spatio-temporal segments. Based on the association of these segments, we obtain an accurate recognition of the gestures involved in the surgical training task. We, then, perform an advanced analysis and assess our algorithm using datasets recorded during real expert training sessions. After comparing our approach with the manual annotations of the surgical gestures, we observe 97.4% accuracy for the learning purpose and an average matching score of 81.9% for the fully automated gesture recognition process. Our results show that trainees workflow can be followed and surgical gestures may be automatically evaluated according to an expert database. This approach tends toward improving training efficiency by minimizing the learning curve.

Automatic knowledge-based recognition of low-level tasks in ophthalmological procedures

PURPOSE

Surgical process models (SPMs) have recently been created for situation-aware computer-assisted systems in the operating room. One important challenge in this area is the automatic acquisition of SPMs. The purpose of this study is to present a new method for the automatic detection of low-level surgical tasks, that is, the sequence of activities in a surgical procedure, from microscope video images only. The level of granularity that we addressed in this work is symbolized by activities formalized by triplets <action, surgical tool, anatomical structure> .

METHODS

Using the results of our latest work on the recognition of surgical phases in cataract surgeries, and based on the hypothesis that most activities occur in one or two phases only, we created a light-weight ontology, formalized as a hierarchical decomposition into phases and activities. Information concerning the surgical tools, the areas where tools are used and three other visual cues were detected through an image-based approach and combined with the information of the current surgical phase within a knowledge-based recognition system. Knowing the surgical phases before the activity, recognition allows supervised classification to be adapted to the phase. Multiclass Support Vector Machines were chosen as a classification algorithm.

RESULTS

Using a dataset of 20 cataract surgeries, and identifying 25 possible pairs of activities, a frame-by-frame recognition rate of 64.5 % was achieved with the proposed system.

CONCLUSIONS

The addition of human knowledge to traditional bottom-up approaches based on image analysis appears to be promising for low-level task detection. The results of this work could be used for the automatic indexation of post-operative videos.

Surgical tools recognition and pupil segmentation for cataract surgical process modeling

In image-guided surgery, a new generation of Computer-Assisted-Surgical (CAS) systems based on information from the Operating Room (OR) has recently been developed to improve situation awareness in the OR. Our main project is to develop an application-dependant framework able to extract high-level tasks (surgical phases) using microscope videos data only. In this paper, we present two methods: one method to segment the pupil and one to extract and recognize surgical tools. We show how both methods improve the accuracy of the framework for analysis of cataract surgery videos, to detect eight surgical phases.

A framework for the recognition of high-level surgical tasks from video images for cataract surgeries

The need for a better integration of the new generation of computer-assisted-surgical systems has been recently emphasized. One necessity to achieve this objective is to retrieve data from the operating room (OR) with different sensors, then to derive models from these data. Recently, the use of videos from cameras in the OR has demonstrated its efficiency. In this paper, we propose a framework to assist in the development of systems for the automatic recognition of high-level surgical tasks using microscope videos analysis. We validated its use on cataract procedures. The idea is to combine state-of-the-art computer vision techniques with time series analysis. The first step of the framework consisted in the definition of several visual cues for extracting semantic information, therefore, characterizing each frame of the video. Five different pieces of image-based classifiers were, therefore, implemented. A step of pupil segmentation was also applied for dedicated visual cue detection. Time series classification algorithms were then applied to model time-varying data. Dynamic time warping and hidden Markov models were tested. This association combined the advantages of all methods for better understanding of the problem. The framework was finally validated through various studies. Six binary visual cues were chosen along with 12 phases to detect, obtaining accuracies of 94%.