An Automatic DWI/FLAIR Mismatch Assessment of Stroke Patients

DWI/FLAIR mismatch assessment for ischemic stroke patients shows promising results in determining if patients are eligible for recombinant tissue-type plasminogen activator (r-tPA) treatment. However, the mismatch criteria suffer from two major issues: binary classification of a non-binary problem and the subjectiveness of the assessor. In this article, we present a simple automatic method for segmenting stroke-related parenchymal hyperintensities on FLAIR, allowing for an automatic and continuous DWI/FLAIR mismatch assessment. We further show that our method's segmentations have comparable inter-rater agreement (DICE 0.820, SD 0.12) compared to that of two neuro-radiologists (DICE 0.856, SD 0.07), that our method appears robust to hyper-parameter choices (suggesting good generalizability), and lastly, that our methods continuous DWI/FLAIR mismatch assessment correlates to mismatch assessments made for a cohort of wake-up stroke patients at hospital submission. The proposed method shows promising results in automating the segmentation of parenchymal hyperintensity within ischemic stroke lesions and could help reduce inter-observer variability of DWI/FLAIR mismatch assessment performed in clinical environments as well as offer a continuous assessment instead of the current binary one.

Collision-constrained deformable image registration framework for discontinuity management

Topological changes like sliding motion, sources and sinks are a significant challenge in image registration. This work proposes the use of the alternating direction method of multipliers as a general framework for constraining the registration of separate objects with individual deformation fields from overlapping in image registration. This constraint is enforced by introducing a collision detection algorithm from the field of computer graphics which results in a robust divide and conquer optimization strategy using Free-Form Deformations. A series of experiments demonstrate that the proposed framework performs superior with regards to the combination of intersection prevention and image registration including synthetic examples containing complex displacement patterns. The results show compliance with the non-intersection constraints while simultaneously preventing a decrease in registration accuracy. Furthermore, the application of the proposed algorithm to the DIR-Lab data set demonstrates that the framework generalizes to real data by validating it on a lung registration problem.

A hybrid approach to full-scale reconstruction of renal arterial network

The renal vasculature, acting as a resource distribution network, plays an important role in both the physiology and pathophysiology of the kidney. However, no imaging techniques allow an assessment of the structure and function of the renal vasculature due to limited spatial and temporal resolution. To develop realistic computer simulations of renal function, and to develop new image-based diagnostic methods based on artificial intelligence, it is necessary to have a realistic full-scale model of the renal vasculature. We propose a hybrid framework to build subject-specific models of the renal vascular network by using semi-automated segmentation of large arteries and estimation of cortex area from a micro-CT scan as a starting point, and by adopting the Global Constructive Optimization algorithm for generating smaller vessels. Our results show a close agreement between the reconstructed vasculature and existing anatomical data obtained from a rat kidney with respect to morphometric and hemodynamic parameters.

Performance and Agreement When Annotating Chest X-ray Text Reports—A Preliminary Step in the Development of a Deep Learning-Based Prioritization and Detection System

A chest X-ray report is a communicative tool and can be used as data for developing artificial intelligence-based decision support systems. For both, consistent understanding and labeling is important. Our aim was to investigate how readers would comprehend and annotate 200 chest X-ray reports. Reports written between 1 January 2015 and 11 March 2022 were selected based on search words. Annotators included three board-certified radiologists, two trained radiologists (physicians), two radiographers (radiological technicians), a non-radiological physician, and a medical student. Consensus labels by two or more of the experienced radiologists were considered “gold standard”. Matthew’s correlation coefficient (MCC) was calculated to assess annotation performance, and descriptive statistics were used to assess agreement between individual annotators and labels. The intermediate radiologist had the best correlation to “gold standard” (MCC 0.77). This was followed by the novice radiologist and medical student (MCC 0.71 for both), the novice radiographer (MCC 0.65), non-radiological physician (MCC 0.64), and experienced radiographer (MCC 0.57). Our findings showed that for developing an artificial intelligence-based support system, if trained radiologists are not available, annotations from non-radiological annotators with basic and general knowledge may be more aligned with radiologists compared to annotations from sub-specialized medical staff, if their sub-specialization is outside of diagnostic radiology.

Pseudo-Label Guided Image Synthesis for Semi-Supervised COVID-19 Pneumonia Infection Segmentation

Coronavirus disease 2019 (COVID-19) has become a severe global pandemic. Accurate pneumonia infection segmentation is important for assisting doctors in diagnosing COVID-19. Deep learning-based methods can be developed for automatic segmentation, but the lack of large-scale well-annotated COVID-19 training datasets may hinder their performance. Semi-supervised segmentation is a promising solution which explores large amounts of unlabelled data, while most existing methods focus on pseudo-label refinement. In this paper, we propose a new perspective on semi-supervised learning for COVID-19 pneumonia infection segmentation, namely pseudo-label guided image synthesis. The main idea is to keep the pseudo-labels and synthesize new images to match them. The synthetic image has the same COVID-19 infected regions as indicated in the pseudo-label, and the reference style extracted from the style code pool is added to make it more realistic. We introduce two representative methods by incorporating the synthetic images into model training, including single-stage Synthesis-Assisted Cross Pseudo Supervision (SA-CPS) and multi-stage Synthesis-Assisted Self-Training (SA-ST), which can work individually as well as cooperatively. Synthesis-assisted methods expand the training data with high-quality synthetic data, thus improving the segmentation performance. Extensive experiments on two COVID-19 CT datasets for segmenting the infections demonstrate our method is superior to existing schemes for semi-supervised segmentation, and achieves the state-of-the-art performance on both datasets. Code is available at: https://github.com/FeiLyu/SASSL.

Inter- and Intra-Observer Agreement When Using a Diagnostic Labeling Scheme for Annotating Findings on Chest X-rays-An Early Step in the Development of a Deep Learning-Based Decision Support System

Consistent annotation of data is a prerequisite for the successful training and testing of artificial intelligence-based decision support systems in radiology. This can be obtained by standardizing terminology when annotating diagnostic images. The purpose of this study was to evaluate the annotation consistency among radiologists when using a novel diagnostic labeling scheme for chest X-rays. Six radiologists with experience ranging from one to sixteen years, annotated a set of 100 fully anonymized chest X-rays. The blinded radiologists annotated on two separate occasions. Statistical analyses were done using Randolph's kappa and PABAK, and the proportions of specific agreements were calculated. Fair-to-excellent agreement was found for all labels among the annotators (Randolph's Kappa, 0.40-0.99). The PABAK ranged from 0.12 to 1 for the two-reader inter-rater agreement and 0.26 to 1 for the intra-rater agreement. Descriptive and broad labels achieved the highest proportion of positive agreement in both the inter- and intra-reader analyses. Annotating findings with specific, interpretive labels were found to be difficult for less experienced radiologists. Annotating images with descriptive labels may increase agreement between radiologists with different experience levels compared to annotation with interpretive labels.

An algebra for local histograms

In this article, we consider local overlapping histograms of functions between discrete domains and codomains. We develop a simple algebra for local histograms. Based on a separation of overlapping domains into non-overlapping domains, we (1) show how these can be used to enumerate the size of the set of possible histograms given the local histogram domains, and (2) enumerate the number of functions, which share a specific choice of a set of local histograms. Finally, we present a decoding algorithm, which given a set of overlapping histograms, and calculate the set of functions, which share these histograms.

LibHip: An open-access hip joint model repository suitable for finite element method simulation

BACKGROUND AND OBJECTIVE

population-based finite element analysis of hip joints allows us to understand the effect of inter-subject variability on simulation results. Developing large subject-specific population models is challenging and requires extensive manual effort. Thus, the anatomical representations are often subjected to simplification. The discretized geometries do not guarantee conformity in shared interfaces, leading to complications in setting up simulations. Additionally, these models are not openly accessible, challenging reproducibility. Our work provides multiple subject-specific hip joint finite element models and a novel semi-automated modeling workflow.

METHODS

we reconstruct 11 healthy subject-specific models, including the sacrum, the paired pelvic bones, the paired proximal femurs, the paired hip joints, the paired sacroiliac joints, and the pubic symphysis. The bones are derived from CT scans, and the cartilages are generated from the bone geometries. We generate the whole complex's volume mesh with conforming interfaces. Our models are evaluated using both mesh quality metrics and simulation experiments.

RESULTS

the geometry of all the models are inspected by our clinical expert and show high-quality discretization with accurate geometries. The simulations produce smooth stress patterns, and the variance among the subjects highlights the effect of inter-subject variability and asymmetry in the predicted results.

CONCLUSIONS

our work is one of the largest model repositories with respect to the number of subjects and regions of interest in the hip joint area. Our detailed research data, including the clinical images, the segmentation label maps, the finite element models, and software tools, are openly accessible on GitHub and the link is provided in Moshfeghifar et al.(2022)[1]. Our aim is to empower clinical researchers to have free access to verified and reproducible models. In future work, we aim to add additional structures to our models.

Bundle geodesic convolutional neural network for diffusion-weighted imaging segmentation

Purpose

Applying machine learning techniques to magnetic resonance diffusion-weighted imaging (DWI) data is challenging due to the size of individual data samples and the lack of labeled data. It is possible, though, to learn general patterns from a very limited amount of training data if we take advantage of the geometry of the DWI data. Therefore, we present a tissue classifier based on a Riemannian deep learning framework for single-shell DWI data.

Approach

The framework consists of three layers: a lifting layer that locally represents and convolves data on tangent spaces to produce a family of functions defined on the rotation groups of the tangent spaces, i.e., a (not necessarily continuous) function on a bundle of rotational functions on the manifold; a group convolution layer that convolves this function with rotation kernels to produce a family of local functions over each of the rotation groups; a projection layer using maximization to collapse this local data to form manifold based functions.

Results

Experiments show that our method achieves the performance of the same level as state-of-the-art while using way fewer parameters in the model ( < 10 % ). Meanwhile, we conducted a model sensitivity analysis for our method. We ran experiments using a proportion (69.2%, 53.3%, and 29.4%) of the original training set and analyzed how much data the model needs for the task. Results show that this does reduce the overall classification accuracy mildly, but it also boosts the accuracy for minority classes.

Conclusions

This work extended convolutional neural networks to Riemannian manifolds, and it shows the potential in understanding structural patterns in the brain, as well as in aiding manual data annotation.

Open-Full-Jaw: An open-access dataset and pipeline for finite element models of human jaw

BACKGROUND

State-of-the-art finite element studies on human jaws are mostly limited to the geometry of a single patient. In general, developing accurate patient-specific computational models of the human jaw acquired from cone-beam computed tomography (CBCT) scans is labor-intensive and non-trivial, which involves time-consuming human-in-the-loop procedures, such as segmentation, geometry reconstruction, and re-meshing tasks. Therefore, with the current practice, researchers need to spend considerable time and effort to produce finite element models (FEMs) to get to the point where they can use the models to answer clinically-interesting questions. Besides, any manual task involved in the process makes it difficult for the researchers to reproduce identical models generated in the literature. Hence, a quantitative comparison is not attainable due to the lack of surface/volumetric meshes and FEMs.

METHODS

We share an open-access repository composed of 17 patient-specific computational models of human jaws and the utilized pipeline for generating them for reproducibility of our work. The used pipeline minimizes the required time for processing and any potential biases in the model generation process caused by human intervention. It gets the segmented geometries with irregular and dense surface meshes and provides reduced, adaptive, watertight, and conformal surface/volumetric meshes, which can directly be used in finite element (FE) analysis.

RESULTS

We have quantified the variability of our 17 models and assessed the accuracy of the developed models from three different aspects; (1) the maximum deviations from the input meshes using the Hausdorff distance as an error measurement, (2) the quality of the developed volumetric meshes, and (3) the stability of the FE models under two different scenarios of tipping and biting.

CONCLUSIONS

The obtained results indicate that the developed computational models are precise, and they consist of quality meshes suitable for various FE scenarios. We believe the provided dataset of models including a high geometrical variation obtained from 17 different models will pave the way for population studies focusing on the biomechanical behavior of human jaws.

Accuracy and consistency of intensity-based deformable image registration in 4DCT for tumor motion estimation in liver radiotherapy planning

We investigate the accuracy of intensity-based deformable image registration (DIR) for tumor localization in liver stereotactic body radiotherapy (SBRT). We included 4DCT scans to capture the breathing motion of eight patients receiving SBRT for liver metastases within a retrospective clinical study. Each patient had three fiducial markers implanted. The liver and the tumor were delineated in the mid-ventilation phase, and their positions in the other phases were estimated with deformable image registration. We tested referenced and sequential registrations strategies. The fiducial markers were the gold standard to evaluate registration accuracy. The registration errors related to measured versus estimated fiducial markers showed a mean value less than 1.6mm. The positions of some fiducial markers appeared not stable on the 4DCT throughout the respiratory phases. Markers’ center of mass tends to be a more reliable measurement. Distance errors of tumor location based on registration versus markers center of mass were less than 2mm. There were no statistically significant differences between the reference and the sequential registration, i.e., consistency and errors were comparable to resolution errors. We demonstrated that intensity-based DIR is accurate up to resolution level for locating the tumor in the liver during breathing motion.

The Added Effect of Artificial Intelligence on Physicians' Performance in Detecting Thoracic Pathologies on CT and Chest X-ray: A Systematic Review

Our systematic review investigated the additional effect of artificial intelligence-based devices on human observers when diagnosing and/or detecting thoracic pathologies using different diagnostic imaging modalities, such as chest X-ray and CT. Peer-reviewed, original research articles from EMBASE, PubMed, Cochrane library, SCOPUS, and Web of Science were retrieved. Included articles were published within the last 20 years and used a device based on artificial intelligence (AI) technology to detect or diagnose pulmonary findings. The AI-based device had to be used in an observer test where the performance of human observers with and without addition of the device was measured as sensitivity, specificity, accuracy, AUC, or time spent on image reading. A total of 38 studies were included for final assessment. The quality assessment tool for diagnostic accuracy studies (QUADAS-2) was used for bias assessment. The average sensitivity increased from 67.8% to 74.6%; specificity from 82.2% to 85.4%; accuracy from 75.4% to 81.7%; and Area Under the ROC Curve (AUC) from 0.75 to 0.80. Generally, a faster reading time was reported when radiologists were aided by AI-based devices. Our systematic review showed that performance generally improved for the physicians when assisted by AI-based devices compared to unaided interpretation.

A multi-patient analysis of the center of rotation trajectories using finite element models of the human mandible

Studying different types of tooth movements can help us to better understand the force systems used for tooth position correction in orthodontic treatments. This study considers a more realistic force system in tooth movement modeling across different patients and investigates the effect of the couple force direction on the position of the center of rotation (CRot). The finite-element (FE) models of human mandibles from three patients are used to investigate the position of the CRots for different patients’ teeth in 3D space. The CRot is considered a single point in a 3D coordinate system and is obtained by choosing the closest point on the axis of rotation to the center of resistance (CRes). A force system, consisting of a constant load and a couple (pair of forces), is applied to each tooth, and the corresponding CRot trajectories are examined across different patients. To perform a consistent inter-patient analysis, different patients’ teeth are registered to the corresponding reference teeth using an affine transformation. The selected directions and applied points of force on the reference teeth are then transformed into the registered teeth domains. The effect of the direction of the couple on the location of the CRot is also studied by rotating the couples about the three principal axes of a patient’s premolar. Our results indicate that similar patterns can be obtained for the CRot positions of different patients and teeth if the same load conditions are used. Moreover, equally rotating the direction of the couple about the three principal axes results in different patterns for the CRot positions, especially in labiolingual direction. The CRot trajectories follow similar patterns in the corresponding teeth, but any changes in the direction of the force and couple cause misalignment of the CRot trajectories, seen as rotations about the long axis of the tooth.

RootPainter3D: Interactive-machine-learning enables rapid and accurate contouring for radiotherapy

PURPOSE

Organ-at-risk contouring is still a bottleneck in radiotherapy, with many deep learning methods falling short of promised results when evaluated on clinical data. We investigate the accuracy and time-savings resulting from the use of an interactive-machine-learning method for an organ-at-risk contouring task.

METHODS

We implement an open-source interactive-machine-learning software application that facilitates corrective-annotation for deep-learning generated contours on X-ray CT images. A trained-physician contoured 933 hearts using our software by delineating the first image, starting model training, and then correcting the model predictions for all subsequent images. These corrections were added into the training data, which was used for continuously training the assisting model. From the 933 hearts, the same physician also contoured the first 10 and last 10 in Eclipse (Varian) to enable comparison in terms of accuracy and duration.

RESULTS

We find strong agreement with manual delineations, with a dice score of 0.95. The annotations created using corrective-annotation also take less time to create as more images are annotated, resulting in substantial time savings compared to manual methods. After 923 images had been delineated, hearts took 2 min and 2 s to delineate on average, which includes time to evaluate the initial model prediction and assign the needed corrections, compared to 7 min and 1 s when delineating manually.

CONCLUSIONS

Our experiment demonstrates that interactive-machine-learning with corrective-annotation provides a fast and accessible way for non computer-scientists to train deep-learning models to segment their own structures of interest as part of routine clinical workflows.

A framework for voxel-based assessment of biological effect after proton radiotherapy in pediatric brain cancer patients using multi-modal imaging

INTRODUCTION

The exact dependence of biological effect on dose and linear energy transfer (LET) in human tissue when delivering proton therapy is unknown. In this study, we propose a framework for measuring this dependency using multi-modal image-based assays with deformable registrations within imaging sessions and across time.

MATERIALS AND METHODS

3T MRI scans were prospectively collected from 6 pediatric brain cancer patients before they underwent proton therapy treatment, and every 3 months for a year after treatment. Scans included T1-weighted with contrast enhancement (T1), T2-FLAIR (T2) and fractional anisotropy (FA) images. In addition, the planning CT, dose distributions and Monte Carlo-calculated LET distributions were collected. A multi-modal deformable image registration framework was used to create a dataset of dose, LET and imaging intensities at baseline and follow-up on a voxel-by-voxel basis. We modelled the biological effect of dose and LET from proton therapy using imaging changes over time as a surrogate for biological effect. We investigated various models to show the feasibility of the framework to model imaging changes. To account for interpatient and intrapatient variations, we used a nested generalized linear mixed regression model. The models were applied to predict imaging changes over time as a function of dose and LET for each modality.

RESULTS

Using the nested models to predict imaging changes, we saw a decrease in the FA signal as a function of dose; however, the signal increased with increasing LET. Similarly, we saw an increase in T2 signal as a function of dose, but a decrease in signal with LET. We saw no changes in T1 voxel values as a function of either dose or LET.

CONCLUSIONS

The imaging changes could successfully model biological effect as a function of dose and LET using our proposed framework. Due to the low number of patients, the imaging changes observed for FA and T2 scans were not marked enough to draw any firm conclusions.

U-Sleep: resilient high-frequency sleep staging

Sleep disorders affect a large portion of the global population and are strong predictors of morbidity and all-cause mortality. Sleep staging segments a period of sleep into a sequence of phases providing the basis for most clinical decisions in sleep medicine. Manual sleep staging is difficult and time-consuming as experts must evaluate hours of polysomnography (PSG) recordings with electroencephalography (EEG) and electrooculography (EOG) data for each patient. Here, we present U-Sleep, a publicly available, ready-to-use deep-learning-based system for automated sleep staging ( sleep.ai.ku.dk ). U-Sleep is a fully convolutional neural network, which was trained and evaluated on PSG recordings from 15,660 participants of 16 clinical studies. It provides accurate segmentations across a wide range of patient cohorts and PSG protocols not considered when building the system. U-Sleep works for arbitrary combinations of typical EEG and EOG channels, and its special deep learning architecture can label sleep stages at shorter intervals than the typical 30 s periods used during training. We show that these labels can provide additional diagnostic information and lead to new ways of analyzing sleep. U-Sleep performs on par with state-of-the-art automatic sleep staging systems on multiple clinical datasets, even if the other systems were built specifically for the particular data. A comparison with consensus-scores from a previously unseen clinic shows that U-Sleep performs as accurately as the best of the human experts. U-Sleep can support the sleep staging workflow of medical experts, which decreases healthcare costs, and can provide highly accurate segmentations when human expertize is lacking.

Relationship between natriuretic peptides and left atrial mechanics and their relation to recurrence of atrial fibrillation following catheter ablation

Funding Acknowledgements

Type of funding sources: Foundation. Main funding source(s): The Danish Heart Foundation (grant no.: 09-04-R72-A2408-22545, 10-04-R78-A2929-22588, 11-04-R84-A3230-22650, and 18-R125-A8534-22083), and the Heart Centre Research Committee at Rigshospitalet, Copenhagen.

Background

The relationship between natriuretic peptides and atrial distension is not completely understood. Furthermore, how they can be used together clinically has not been fully explored.

Purpose

We sought to examine their interrelationship and how they relate to atrial fibrillation (AF) recurrence following catheter ablation.

Methods

Patients scheduled for catheter ablation as part of a randomized controlled clinical trial were included. Patients who underwent pre-operative echocardiography and had natriuretic peptide measurements performed, specifically mid-regional proANP (MR-proANP) and N-terminal proBNP (NT-proBNP), were included in this analysis. Echocardiography included assessment of atrial distension by left atrial strain.

The outcome was AF recurrence at 6 months after a 3-month blanking period. Logistic regression was performed to assess the association between log-transformed natriuretic peptides and AF. Multivariable adjustments were made for age, gender, randomization, and LVEF.

Results

Out of 99 patients 44 developed AF. No differences in natriuretic peptides nor echocardiographic measures were observed between the outcome groups.

Neither MR-proANP nor NT-proBNP were univariable predictors of AF recurrence (MR-proANP: OR = 1.06 (0.99-1.14), p = 0.09, per 10% increase; NT-proBNP: OR = 1.01 (0.98-1.05), p = 0.38, per 10% increase). These findings were unchanged after multivariable adjustments. However, atrial strain significantly modified the association between MR-proANP and AF (p for interaction = 0.009) such that MR-proANP was a significant predictor of AF in patients with high atrial strain values (OR = 1.24 (1.06-1.46), p = 0.008, per 10% increase) but not in patients with low atrial strain values. Among patients with high atrial strain values, an MR-proANP &gt; 116pmol/L was associated with a 10-fold increased risk of AF (OR = 9.78 (2.21-43.33), p = 0.003). figure.

Conclusion

Atrial natriuretic peptide predicts AF recurrence in patients with preserved atrial distension. Assessing atrial distension by echocardiography may assist the clinical interpretation of atrial natriuretic peptide concentration.

Abstract Figure.

Association between natriuretic peptides and left atrial structural and functional properties in atrial fibrillation following catheter ablation

Background

Atrial natriuretic peptides (ANP) and brain natriuretic peptides (BNP) are acutely released from the atrial myocytes upon increased mechanical distension of the atria. The relationship between imaging measures of left atrial (LA) structure and function and natriuretic peptides following catheter ablation (CA) have not been clearly delineated.

Purpose

To characterize the relationship between LA structure and function and natriuretic peptides.

Methods

We performed an echocardiographic substudy of a randomized trial of AF patients scheduled for CA. Echocardiographic measurements included: LA volume at end-systole (LAVi), at end-diastole (LAEDVi), emptying fraction (LAEF), LA reservoir strain (LAs), and global longitudinal strain (GLS). Patients were stratified by tertiles of mid-regional proANP (MR-proANP) concentrations in circulation (&lt;92 pmol/l, 92–146 pmol/l, &gt;146 pmol/l), and NT-proBNP (&lt;10pmol/l, 10–38 pmol/l, &gt;38 pmol/l). Linear regressions were performed to compare baseline echocardiographic measures to natriuretic peptide concentrations at baseline, 1 month, 3 months and 6 months of follow-up. MR-proANP and NT-proBNP were logarithm transformed in these analyses. Multivariable adjustments were made for: age, gender, AF subtype, AF burden, rhythm during echocardiogram, rhythm at study visit for blood sampling, time known with AF, beta-blocker use, and CHA-2DS2-VASc score.

Results

We included 101 patients with AF. The mean age was 58 years, 82% were men, 46% had persistent AF. Increasing tertiles of MR-proANP at baseline were associated with abnormal LA size and function (3rd vs 1st tertile: LAVi: 42mL/m2 vs 32mL/m2; LAEDVi: 31mL/m2 vs 20mL/m2; LAEF: 38% vs 26%; LAs: 27% vs 19%; GLS: −18% vs −14%) whereas both LA and left ventricular measures were associated with increasing NT-proBNP concentrations at baseline. After multivariable adjustments, only LA volumes and LAEF remained significantly associated with MR-proANP, whereas only LA volumes and GLS remained significantly associated with NT-proBNP. At follow-up, impaired LA function associated with persistently elevated concentrations, which was not the case for LAVi (figure).

Conclusion

MR-proANP reflects LA dysfunction better than NT-proBNP. Measures of LA function rather than LAVi associates with persistently elevated natriuretic peptide concentrationsw, which may indicate that functional measures are more closely associated with evidence of LA myocardial stretch than LAVi.

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): The Danish Heart Foundation

Relationship between cardiac structure and function and atrial fibrillation related hospitalizations following catheter ablation

Background

Even though catheter ablation (CA) is an effective treatment for atrial fibrillation (AF), AF-related hospitalizations and cardioversions are common following this procedure.

Purpose

To investigate whether echocardiographic measures of left atrial (LA) function could predict AF-related hospitalizations and cardioversions.

Methods

This was a substudy of a trial that randomized patients to amiodarone vs place to reduce AF recurrence following CA. Transthoracic echocardiography was performed prior to CA and included assessment of: end-systolic and end-diastolic LA volumes, emptying fraction (LAEF), atrial strain, and global longitudinal strain (GLS). Poisson regression was used to assess predictive value for AF-related hospitalizations and cardioversions. Multivariable adjustments were made for: age, gender, ejection fraction, AF burden, AF subtype, dyspnea, and class 1c antiarrhythmics.

Results

Of the 212 patients, 80 were hospitalized for AF (206 times), and 77 were cardioverted (192 times) within the 6 months follow-up period. Mean age was 60 years, 83% were men, and mean LVEF was 50%. In univariable analyses, LA volumes, LAEF and GLS were predictors of the outcomes but did not remain significant predictors after multivariable adjustments. During echocardiography 162 patients were in sinus rhythm and 50 had AF rhythm. Rhythm during the echocardiogram modified the association between GLS and outcomes (p for interaction &lt;0.05 for both endpoints), such that GLS predicted both AF-related hospitalizations and cardioversions in patients with sinus rhythm but not AF during the echocardiogram (figure).

Conclusion

Global longitudinal strain predicts AF-related hospitalizations and cardioversions after CA, but only in patients presenting in sinus rhythm during the echocardiogram. Patients presenting with impaired global longitudinal strain should be considered high-risk patients following CA who may benefit from close follow-up.

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): The Danish Heart Foundation

Restoring drifted electron microscope volumes using synaptic vesicles at sub-pixel accuracy

Imaging ultrastructures in cells using Focused Ion Beam Scanning Electron Microscope (FIB-SEM) yields section-by-section images at nano-resolution. Unfortunately, we observe that FIB-SEM often introduces sub-pixel drifts between sections, in the order of 2.5 nm. The accumulation of these drifts significantly skews distance measures and geometric structures, which standard image registration techniques fail to correct. We demonstrate that registration techniques based on mutual information and sum-of-squared-distances significantly underestimate the drift since they are agnostic to image content. For neuronal data at nano-resolution, we discovered that vesicles serve as a statistically simple geometric structure, making them well-suited for estimating the drift with sub-pixel accuracy. Here, we develop a statistical model of vesicle shapes for drift correction, demonstrate its superiority, and provide a self-contained freely available application for estimating and correcting drifted datasets with vesicles.

Collocation for Diffeomorphic Deformations in Medical Image Registration

Diffeomorphic deformation is a popular choice in medical image registration. A fundamental property of diffeomorphisms is invertibility, implying that once the relation between two points A to B is found, then the relation B to A is given per definition. Consistency is a measure of a numerical algorithm's ability to mimic this invertibility, and achieving consistency has proven to be a challenge for many state-of-the-art algorithms. We present CDD (Collocation for Diffeomorphic Deformations), a numerical solution to diffeomorphic image registration, which solves for the Stationary Velocity Field (SVF) using an implicit A-stable collocation method. CDD guarantees the preservation of the diffeomorphic properties at all discrete points and is thereby consistent to machine precision. We compared CDD's collocation method with the following standard methods: Scaling and Squaring, Forward Euler, and Runge-Kutta 4, and found that CDD is up to 9 orders of magnitude more consistent. Finally, we evaluated CDD on a number of standard bench-mark data sets and compared the results with current state-of-the-art methods: SPM-DARTEL, Diffeomorphic Demons and SyN. We found that CDD outperforms state-of-the-art methods in consistency and delivers comparable or superior registration precision.

Safety and Efficacy of Mesenchymal Stem Cells for Radiation-Induced Xerostomia: A Randomized, Placebo-Controlled Phase 1/2 Trial (MESRIX)

BACKGROUND

Salivary gland hypofunction and xerostomia are major complications to head and neck radiotherapy. This trial assessed the safety and efficacy of adipose tissue-derived mesenchymal stem cell (ASC) therapy for radiation-induced xerostomia.

PATIENT AND METHODS

This randomized, placebo-controlled phase 1/2 trial included 30 patients, randomized in a 1:1 ratio to receive ultrasound-guided transplantation of ASCs or placebo to the submandibular glands. Patients had previously received radiotherapy for a T1-2, N0-2A, human papillomavirus-positive, oropharyngeal squamous cell carcinoma. The primary outcome was the change in unstimulated whole salivary flow rate, measured before and after the intervention. All assessments were performed one month prior (baseline) and one and four months following ASC or placebo administration.

RESULTS

No adverse events were detected. Unstimulated whole salivary flow rates significantly increased in the ASC-arm at one (33%; P = .048) and four months (50%; P = .003), but not in the placebo-arm (P = .6 and P = .8), compared to baseline. The ASC-arm symptom scores significantly decreased on the xerostomia and VAS questionnaires, in the domains of thirst (-22%, P = .035) and difficulties in eating solid foods (-2%, P = .008) after four months compared to baseline. The ASC-arm showed significantly improved salivary gland functions of inorganic element secretion and absorption, at baseline and four months, compared to the placebo-arm. Core-needle biopsies showed increases in serous gland tissue and decreases in adipose and connective tissues in the ASC-arm compared to the placebo-arm (P = .04 and P = .02, respectively). MRIs showed no significant differences between groups in gland size or intensity (P < .05).

CONCLUSIONS

ASC therapy for radiation-induced hypofunction and xerostomia was safe and significantly improved salivary gland functions and patient-reported outcomes. These results should encourage further exploratory and confirmatory trials.

First-in-man mesenchymal stem cells for radiation-induced xerostomia (MESRIX): study protocol for a randomized controlled trial

Background

Salivary gland hypofunction and xerostomia are major complications following radiotherapy for head and neck cancer and may lead to debilitating oral disorders and impaired quality of life. Currently, only symptomatic treatment is available. However, mesenchymal stem cell (MSC) therapy has shown promising results in preclinical studies. Objectives are to assess safety and efficacy in a first-in-man trial on adipose-derived MSC therapy (ASC) for radiation-induced xerostomia.

Methods

This is a single-center, phase I/II, randomized, placebo-controlled, double-blinded clinical trial. A total of 30 patients are randomized in a 1:1 ratio to receive ultrasound-guided, administered ASC or placebo to the submandibular glands. The primary outcome is change in unstimulated whole salivary flow rate. The secondary outcomes are safety, efficacy, change in quality of life, qualitative and quantitative measurements of saliva, as well as submandibular gland size, vascularization, fibrosis, and secretory tissue evaluation based on contrast-induced magnetic resonance imaging (MRI) and core-needle samples. The assessments are performed at baseline (1 month prior to treatment) and 1 and 4 months following investigational intervention.

Discussion

The trial is the first attempt to evaluate the safety and efficacy of adipose-derived MSCs (ASCs) in patients with radiation-induced xerostomia. The results may provide evidence for the effectiveness of ASC in patients with salivary gland hypofunction and xerostomia and deliver valuable information for the design of subsequent trials.

Trial registration

EudraCT, Identifier: 2014-004349-29. Registered on 1 April 2015.

ClinicalTrials.gov, Identifier: NCT02513238. First received on 2 July 2015.

The trial is prospectively registered.

Electronic supplementary material

The online version of this article (doi:10.1186/s13063-017-1856-0) contains supplementary material, which is available to authorized users.

Kernel Bundle Diffeomorphic Image Registration Using Stationary Velocity Fields and Wendland Basis Functions

In this paper, we propose a multi-scale, multi-kernel shape, compactly supported kernel bundle framework for stationary velocity field-based image registration (Wendland kernel bundle stationary velocity field, wKB-SVF). We exploit the possibility of directly choosing kernels to construct a reproducing kernel Hilbert space (RKHS) instead of imposing it from a differential operator. The proposed framework allows us to minimize computational cost without sacrificing the theoretical foundations of SVF-based diffeomorphic registration. In order to recover deformations occurring at different scales, we use compactly supported Wendland kernels at multiple scales and orders to parameterize the velocity fields, and the framework allows simultaneous optimization over all scales. The performance of wKB-SVF is extensively compared to the 14 non-rigid registration algorithms presented in a recent comparison paper. On both MGH10 and CUMC12 datasets, the accuracy of wKB-SVF is improved when compared to other registration algorithms. In a disease-specific application for intra-subject registration, atrophy scores estimated using the proposed registration scheme separates the diagnostic groups of Alzheimer's and normal controls better than the state-of-the-art segmentation technique. Experimental results show that wKB-SVF is a robust, flexible registration framework that allows theoretically well-founded and computationally efficient multi-scale representation of deformations and is equally well-suited for both inter- and intra-subject image registration.

The Minimal Energetic Requirement of Sustained Awareness after Brain Injury

Differentiation of the minimally conscious state (MCS) and the unresponsive wakefulness syndrome (UWS) is a persistent clinical challenge [1]. Based on positron emission tomography (PET) studies with [(18)F]-fluorodeoxyglucose (FDG) during sleep and anesthesia, the global cerebral metabolic rate of glucose has been proposed as an indicator of consciousness [2, 3]. Likewise, FDG-PET may contribute to the clinical diagnosis of disorders of consciousness (DOCs) [4, 5]. However, current methods are non-quantitative and have important drawbacks deriving from visually guided assessment of relative changes in brain metabolism [4]. We here used FDG-PET to measure resting state brain glucose metabolism in 131 DOC patients to identify objective quantitative metabolic indicators and predictors of awareness. Quantitation of images was performed by normalizing to extracerebral tissue. We show that 42% of normal cortical activity represents the minimal energetic requirement for the presence of conscious awareness. Overall, the cerebral metabolic rate accounted for the current level, or imminent return, of awareness in 94% of the patient population, suggesting a global energetic threshold effect, associated with the reemergence of consciousness after brain injury. Our data further revealed that regional variations relative to the global resting metabolic level reflect preservation of specific cognitive or sensory modules, such as vision and language comprehension. These findings provide a simple and objective metabolic marker of consciousness, which can readily be implemented clinically. The direct correlation between brain metabolism and behavior further suggests that DOCs can fundamentally be understood as pathological neuroenergetic conditions and provide a unifying physiological basis for these syndromes.

Deformation-based atrophy computation by surface propagation and its application to Alzheimer's disease

Obtaining regional volume changes from a deformation field is more precise when using simplex counting (SC) compared with Jacobian integration (JI) due to the numerics involved in the latter. Although SC has been proposed before, numerical properties underpinning the method and a thorough evaluation of the method against JI is missing in the literature. The contributions of this paper are: (a) we propose surface propagation (SP)-a simplification to SC that significantly reduces its computational complexity; (b) we will derive the orders of approximation of SP which can also be extended to SC. In the experiments, we will begin by empirically showing that SP is indeed nearly identical to SC, and that both methods are more stable than JI in presence of moderate to large deformation noise. Since SC and SP are identical, we consider SP as a representative of both the methods for a practical evaluation against JI. In a real application on Alzheimer's disease neuroimaging initiative data, we show the following: (a) SP produces whole brain and medial temporal lobe atrophy numbers that are significantly better than JI at separating between normal controls and Alzheimer's disease patients; (b) SP produces disease group atrophy differences comparable to or better than those obtained using FreeSurfer, demonstrating the validity of the obtained clinical results. Finally, in a reproducibility study, we show that the voxel-wise application of SP yields significantly lower variance when compared to JI.

Automatic correction of dental artifacts in PET/MRI

A challenge when using current magnetic resonance (MR)-based attenuation correction in positron emission tomography/MR imaging (PET/MRI) is that the MRIs can have a signal void around the dental fillings that is segmented as artificial air-regions in the attenuation map. For artifacts connected to the background, we propose an extension to an existing active contour algorithm to delineate the outer contour using the nonattenuation corrected PET image and the original attenuation map. We propose a combination of two different methods for differentiating the artifacts within the body from the anatomical air-regions by first using a template of artifact regions, and second, representing the artifact regions with a combination of active shape models and k-nearest-neighbors. The accuracy of the combined method has been evaluated using 25 [Formula: see text]-fluorodeoxyglucose PET/MR patients. Results showed that the approach was able to correct an average of [Formula: see text] of the artifact areas.

Locally orderless registration

This paper presents a unifying approach for calculating a wide range of popular, but seemingly very different, similarity measures. Our domain is the registration of n-dimensional images sampled on a regular grid, and our approach is well suited for gradient-based optimization algorithms. Our approach is based on local intensity histograms and built upon the technique of Locally Orderless Images. Histograms by Locally Orderless Images are well posed and offer explicit control over the three inherent and unavoidable scales: the spatial resolution, intensity levels, and spatial extent of local histograms. Through Locally Orderless Images, we offer new insight into the relations between these scales. We demonstrate our unification by developing a Locally Orderless Registration algorithm for two quite different similarity measures, namely, Normalized Mutual Information and Sum of Squared Differences, and we compare these variations both theoretically and empirically. Finally, using our algorithm, we explain the empirically observed differences between two popular joint density estimation techniques used in registration: Parzen Windows and Generalized Partial Volume.

Generalized partial volume: an inferior density estimator to Parzen windows for normalized mutual information

Mutual Information (MI) and normalized mutual information (NMI) are popular choices as similarity measure for multimodal image registration. Presently, one of two approaches is often used for estimating these measures: The Parzen Window (PW) and the Generalized Partial Volume (GPV). Their theoretical relation has so far been unexplored. We present the direct connection between PW and GPV for NMI in the case of rigid and non-rigid image registration. Through step-by-step derivations of PW and GPV we clarify the difference and show that GPV is algorithmically inferior to PW from a model point of view as well as w.r.t. computational complexity. Finally, we present algorithms for both approaches for NMI which is comparable in speed to Sum of Squared Differences (SSD), and we illustrate the differences between PW and GPV on a number of registration examples.

Analysis of surfaces using constrained regression models

We present a study of the relationship between the changes in the shape of the human ear due to jaw movement and acoustical feedback (AF) in hearing aids. In particular, we analyze the deformation field of the outer ear associated with the movement of the mandible (jaw bone) to understand its effect on AF and identify local regions that play a significant role. Our data contains ear impressions of 42 hearing aid users, in two different positions: open and closed mouth, and survey data including information about experienced discomfort due to AF. We use weighted support vector machines (WSVM) to investigate the separation between the presence and lack of AF and achieve classification accuracy of 80% based on the deformation field. To robustly localize the regions of the deformation field that significantly contribute to AF we employ logistic regression penalized with elastic net (EN). By visualizing the selected variables on the mean surface, we provide clinical interpretations of the results.

Analysis of deformation of the human ear and canal caused by mandibular movement

Many hearing aid users experience physical discomfort when wearing their device. The main contributor to this problem is believed to be deformation of the ear and ear canal caused by movement of the mandible. Physical discomfort results from added pressure on soft tissue areas in the ear. Identifying features that can predict potential deformation is therefore important for identifying problematic cases in advance. A study on the physical deformation of the human ear and canal due to movement of the mandible is presented. The study is based on laser scannings of 30 pairs of ear impressions from 9 female and 21 male subjects. Two impressions have been taken from each subject, one with open mouth, and one with the mouth closed. All impressions are registered using non-rigid surface registration and a shape model is built. From each pair of impressions a deformation field is generated and propagated to the shape model, enabling the building of a deformation model in the reference frame of the shape model. A relationship between the two models is established, showing that the shape variation can explain approximately 50% of the variation in the deformation model. An hypothesis test for significance of the deformations for each deformation field reveals that all subjects have significant deformation at Tragus and in the canal. Furthermore, a relation between the magnitude of the deformation and the gender of the subject is demonstrated. The results are successfully validated by comparing the outcome to the anatomy by using a single set of high resolution histological sectionings of the region of interest.