Accurate fetal brain tissue segmentation and biometric measurement are essential for monitoring neurodevelopment and detecting abnormalities in utero. The Fetal Tissue Annotation (FeTA) Challenges have established robust multi-center benchmarks for evaluating state-of-the-art segmentation methods. This paper presents the results of the 2024 challenge edition, which introduced three key innovations. First, we introduced a topology-aware metric based on the Euler characteristic difference (ED) to overcome the performance plateau observed with traditional metrics like Dice or Hausdorff distance (HD), as the performance of the best models in segmentation surpassed the inter-rater variability. While the best teams reached similar scores in Dice (0.81-0.82) and HD95 (2.1-2.3 mm), ED provided greater discriminative power: the winning method achieved an ED of 20.9, representing roughly a 50% improvement over the second- and third-ranked teams despite comparable Dice scores. Second, we introduced a new 0.55T low-field MRI test set, which, when paired with high-quality super-resolution reconstruction, achieved the highest segmentation performance across all test cohorts (Dice=0.86, HD95=1.69, ED=6.26). This provides the first quantitative evidence that low-cost, low-field MRI can match or surpass high-field systems in automated fetal brain segmentation. Third, the new biometry estimation task exposed a clear performance gap: although the best model reached a mean average percentage error (MAPE) of 7.72%, most submissions failed to outperform a simple gestational-age-based linear regression model (MAPE=9.56%), and all remained above inter-rater variability with a MAPE of 5.38%. Finally, by analyzing the top-performing models from FeTA 2024 alongside those from previous challenge editions, we identify ensembles of 3D nnU-Net trained on both real and synthetic data with both image- and anatomy-level augmentations as the most effective approaches for fetal brain segmentation. Our quantitative analysis reveals that acquisition site, super-resolution strategy, and image quality are the primary sources of domain shift, informing recommendations to enhance the robustness and generalizability of automated fetal brain analysis methods.

A major contributor to permanent vision loss is glaucoma. Early diagnosis is crucial for preventing vision loss due to glaucoma, making glaucoma screening essential. A more affordable method of glaucoma screening can be achieved by applying artificial intelligence to evaluate color fundus photographs (CFPs). We present the Justified Referral in AI Glaucoma Screening (JustRAIGS) challenge to further develop these AI algorithms for glaucoma screening and to assess their efficacy. To support this challenge, we have generated a distinctive big dataset containing more than 110,000 meticulously labeled CFPs obtained from approximately 60,000 patients and 500 distinct screening centers in the USA. Our objective is to assess the practicality of creating advanced and dependable AI systems that can take a CFP as input and produce the probability of referable glaucoma, as well as outputs for glaucoma justification by integrating both binary and multi-label classification tasks. This paper presents the evaluation of solutions provided by nine teams, recognizing the team with the highest level of performance. The highest achieved score of sensitivity at a specificity level of 95% was 85%, and the highest achieved score of Hamming losses average was 0.13. Additionally, we test the top three participants' algorithms on an external dataset to validate the performance and generalization of these models. The outcomes of this research can offer valuable insights into the development of intelligent systems for detecting glaucoma. Ultimately, findings can aid in the early detection and treatment of glaucoma patients, hence decreasing preventable vision impairment and blindness caused by glaucoma.

The widespread adoption of pre-trained models in artificial intelligence poses several evaluation challenges, primarily due to limited access to comprehensive information including training data, test data, loss function, and hyperparameter values. Traditional evaluation metrics, like accuracy, precision, and recall, demand the availability of train and test data, making it difficult to determine the expected performance or quality of the deep learning models. This paper addresses this challenge by introducing a comprehensive empirical analysis of Convolutional Neural Networks (CNNs) using Dynamic Mode Decomposition (DMD) theory. Feature maps from the convolutional layers are used to model the trained CNN as a linear dynamical system. By constructing snapshots for DMD analysis, we decompose the system’s dynamics into individual modes and eigenvalues, providing a linear approximation of the intricate CNN behavior. We also introduce DMD-based metrics, including entropy of DMD-eigenvalues (EDMD), DMD eigen gap/spectral gap, and zero amplitude DMD modes, to quantify the effectiveness and quality of the learned CNN (well-trained vs poorly-trained) without the need to access the test data. Results indicate that the entropy of DMD-eigenvalues (EDMD) correlates well with the classical evaluation metrics and also distinguishes between well-trained and poorly-trained models. Additionally, we explored the potential of interpreting CNNs through stable DMD modes and their corresponding eigenvalues. Our approach was thoroughly evaluated on both shallow and deep models across various datasets for classification and segmentation tasks, demonstrating its effectiveness in quantifying CNN learning evolution and task performance. We also conducted a comprehensive analysis to examine the correlation between the proposed metrics and classical evaluation metrics on pre-trained models, specifically using ImageNet for classification (20 models) and Pascal VOC for segmentation (3 models) tasks. The code is available at: https://github.com/sikha2552/DMD_XAI . • Propose a DMD formulation for CNNs as a linear dynamical system using feature maps extracted from the filters of the convolutional blocks to construct snapshots for DMD analysis. This approach aims to analyze and interpret the behavior of CNNs and their learning evolution by decomposing the system’s dynamics into individual modes using DMD. These modes are then employed to evaluate and understand the network’s performance over epochs. • Introduce DMD-based metrics to quantify the quality of the learned CNNs, which includes: Entropy of DMD-eigenvalues (EDMD), DMD Eigen Gap/ Spectral Gap, Zero Amplitude DMD Modes, and analyze their correlation with the classical task-specific evaluation metrics (test accuracy, IoU, Dice). • Performs an extensive evaluation of the proposed quality metrics across various shallow and deep classification and segmentation models trained end-to-end on diverse datasets. Additionally, conduct a comprehensive analysis to explore the correlation between the proposed metrics and classical evaluation metrics on pretrained models, specifically using ImageNet for classification and the Pascal VOC dataset for segmentation tasks. • Present a novel approach for interpreting CNNs using the stable DMD modes and their associated DMD-eigenvalues. Compare the interpretation maps generated by the proposed DMD-based method with state-of-the-art CNN interpretability methods, including Grad-CAM, Grad-CAM++, and ScoreCAM.

Monitoring the healing progress of diabetic foot ulcers is a challenging process. Accurate segmentation of foot ulcers can help podiatrists to quantitatively measure the size of wound regions to assist prediction of healing status. The main challenge in this field is the lack of publicly available manual delineation, which can be time consuming and laborious. Recently, methods based on deep learning have shown excellent results in automatic segmentation of medical images, however, they require large-scale datasets for training, and there is limited consensus on which methods perform the best. The 2022 Diabetic Foot Ulcers segmentation challenge was held in conjunction with the 2022 International Conference on Medical Image Computing and Computer Assisted Intervention, which sought to address these issues and stimulate progress in this research domain. A training set of 2000 images exhibiting diabetic foot ulcers was released with corresponding segmentation ground truth masks. Of the 72 (approved) requests from 47 countries, 26 teams used this data to develop fully automated systems to predict the true segmentation masks on a test set of 2000 images, with the corresponding ground truth segmentation masks kept private. Predictions from participating teams were scored and ranked according to their average Dice similarity coefficient of the ground truth masks and prediction masks. The winning team achieved a Dice of 0.7287 for diabetic foot ulcer segmentation. This challenge has now entered a live leaderboard stage where it serves as a challenging benchmark for diabetic foot ulcer segmentation.

The early detection of glaucoma is essential in preventing visual impairment. Artificial intelligence (AI) can be used to analyze color fundus photographs (CFPs) in a cost-effective manner, making glaucoma screening more accessible. While AI models for glaucoma screening from CFPs have shown promising results in laboratory settings, their performance decreases significantly in real-world scenarios due to the presence of out-of-distribution and low-quality images. To address this issue, we propose the Artificial Intelligence for Robust Glaucoma Screening (AIROGS) challenge. This challenge includes a large dataset of around 113,000 images from about 60,000 patients and 500 different screening centers, and encourages the development of algorithms that are robust to ungradable and unexpected input data. We evaluated solutions from 14 teams in this paper and found that the best teams performed similarly to a set of 20 expert ophthalmologists and optometrists. The highest-scoring team achieved an area under the receiver operating characteristic curve of 0.99 (95% CI: 0.98-0.99) for detecting ungradable images on-the-fly. Additionally, many of the algorithms showed robust performance when tested on three other publicly available datasets. These results demonstrate the feasibility of robust AI-enabled glaucoma screening.

MOTIVATION: Low back pain is responsible for more global disability than any other condition. Its incidence is closely related to intervertebral disc (IVD) failure, which is likely caused by an accumulation of microtrauma within the IVD. Crucial factors in microtrauma development are not entirely known yet, probably because their exploration in vivo or in vitro remains tremendously challenging. In silico modelling is, therefore, definitively appealing, and shall include approaches to integrate influences of multiple cell stimuli at the microscale. Accordingly, this study introduces a hybrid Agent-based (AB) model in IVD research and exploits network modelling solutions in systems biology to mimic the cellular behaviour of Nucleus Pulposus cells exposed to a 3D multifactorial biochemical environment, based on mathematical integrations of existing experimental knowledge. Cellular activity reflected by mRNA expression of Aggrecan, Collagen type I, Collagen type II, MMP-3 and ADAMTS were calculated for inflamed and non-inflamed cells. mRNA expression over long periods of time is additionally determined including cell viability estimations. Model predictions were eventually validated with independent experimental data. RESULTS: As it combines experimental data to simulate cell behaviour exposed to a multifactorial environment, the present methodology was able to reproduce cell death within 3 days under glucose deprivation and a 50% decrease in cell viability after 7 days in an acidic environment. Cellular mRNA expression under non-inflamed conditions simulated a quantifiable catabolic shift under an adverse cell environment, and model predictions of mRNA expression of inflamed cells provide new explanation possibilities for unexpected results achieved in experimental research. AVAILABILITYAND IMPLEMENTATION: The AB model as well as used mathematical functions were built with open source software. Final functions implemented in the AB model and complete AB model parameters are provided as Supplementary Material. Experimental input and validation data were provided through referenced, published papers. The code corresponding to the model can be shared upon request and shall be reused after proper training. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

3D echocardiography is an increasingly popular tool for assessing cardiac remodelling in the right ventricle (RV). It allows quantification of the cardiac chambers without any geometric assumptions, which is the main weakness of 2D echocardiography. However, regional quantification of geometry and function is limited by the lower spatial and temporal resolution and the scarcity of identifiable anatomical landmarks, especially within the ventricular cavity. We developed a technique for regionally assessing the volume of 3 relevant RV volumetric regions: apical, inlet and outflow. The proposed parcellation method is based on the geodesic distances to anatomical landmarks that are easily identifiable in the images: the apex and the tricuspid and pulmonary valves, each associated to a region. Based on these distances, we define a partition in the endocardium at end-diastole (ED). This partition is then interpolated to the blood cavity using the Laplace equation, which allows to compute regional volumes. For obtaining an end-systole (ES) partition, the endocardial partition is transported from ED to ES using a commercial image-based tracking software, and then the interpolation process is repeated. We assessed the intra- and inter-observer reproducibility using a 10-subjects dataset containing repeated quantifications of the same images, obtaining intra- and inter- observer errors (7-12% and 10-23% respectively). Finally, we propose a novel synthetic mesh generation algorithm that deforms a template mesh imposing a user-defined strain to a template mesh. We used this method to create a new dataset for involving distinct types of remodelling that were used to assess the sensitivity of the parcellation method to identify volume changes affecting different parts. We show that the parcellation method is adequate for capturing local circumferential and global circumferential and longitudinal RV remodelling, which are the most clinically relevant cases.

Fetoscopic laser photocoagulation is the most effective treatment for Twin-to-Twin Transfusion Syndrome, a condition affecting twin pregnancies in which there is a deregulation of blood circulation through the placenta, that can be fatal to both babies. For the purposes of surgical planning, we design the first automatic approach to detect and segment the intrauterine cavity from axial, sagittal and coronal MRI stacks. Our methodology relies on the ability of capsule networks to successfully capture the part-whole interdependency of objects in the scene, particularly for unique class instances (i.e., intrauterine cavity). The presented deep Q-CapsNet reinforcement learning framework is built upon a context-adaptive detection policy to generate a bounding box of the womb. A capsule architecture is subsequently designed to segment (or refine) the whole intrauterine cavity. This network is coupled with a strided nnU-Net feature extractor, which encodes discriminative feature maps to construct strong primary capsules. The method is robustly evaluated with and without the localization stage using 13 performance measures, and directly compared with 15 state-of-the-art deep neural networks trained on 71 singleton and monochorionic twin pregnancies. An average Dice score above 0.91 is achieved for all ablations, revealing the potential of our approach to be used in clinical practice.

Twin-to-twin transfusion syndrome (TTTS) is characterized by an unbalanced blood transfer through placental abnormal vascular connections. Prenatal ultrasound (US) is the imaging technique to monitor monochorionic pregnancies and diagnose TTTS. Fetoscopic laser photocoagulation is an elective treatment to coagulate placental communications between both twins. To locate the anomalous connections ahead of surgery, preoperative planning is crucial. In this context, we propose a novel multi-task stacked generative adversarial framework to jointly learn synthetic fetal US generation, multi-class segmentation of the placenta, its inner acoustic shadows and peripheral vasculature, and placenta shadowing removal. Specifically, the designed architecture is able to learn anatomical relationships and global US image characteristics. In addition, we also extract for the first time the umbilical cord insertion on the placenta surface from 3D HD-flow US images. The database consisted of 70 US volumes including singleton, mono- and dichorionic twins at 17-37 gestational weeks. Our experiments show that 71.8% of the synthesized US slices were categorized as realistic by clinicians, and that the multi-class segmentation achieved Dice scores of 0.82 ± 0.13, 0.71 ± 0.09, and 0.72 ± 0.09, for placenta, acoustic shadows, and vasculature, respectively. Moreover, fetal surgeons classified 70.2% of our completed placenta shadows as satisfactory texture reconstructions. The umbilical cord was successfully detected on 85.45% of the volumes. The framework developed could be implemented in a TTTS fetal surgery planning software to improve the intrauterine scene understanding and facilitate the location of the optimum fetoscope entry point.

Delineation of the left ventricular cavity, myocardium, and right ventricle from cardiac magnetic resonance images (multi-slice 2-D cine MRI) is a common clinical task to establish diagnosis. The automation of the corresponding tasks has thus been the subject of intense research over the past decades. In this paper, we introduce the "Automatic Cardiac Diagnosis Challenge" dataset (ACDC), the largest publicly available and fully annotated dataset for the purpose of cardiac MRI (CMR) assessment. The dataset contains data from 150 multi-equipments CMRI recordings with reference measurements and classification from two medical experts. The overarching objective of this paper is to measure how far state-of-the-art deep learning methods can go at assessing CMRI, i.e., segmenting the myocardium and the two ventricles as well as classifying pathologies. In the wake of the 2017 MICCAI-ACDC challenge, we report results from deep learning methods provided by nine research groups for the segmentation task and four groups for the classification task. Results show that the best methods faithfully reproduce the expert analysis, leading to a mean value of 0.97 correlation score for the automatic extraction of clinical indices and an accuracy of 0.96 for automatic diagnosis. These results clearly open the door to highly accurate and fully automatic analysis of cardiac CMRI. We also identify scenarios for which deep learning methods are still failing. Both the dataset and detailed results are publicly available online, while the platform will remain open for new submissions.

Dual Energy X-ray Absorptiometry (DXA) is the standard exam for osteoporosis diagnosis and fracture risk evaluation at the spine. However, numerous patients with bone fragility are not diagnosed as such. In fact, standard analysis of DXA images does not differentiate between trabecular and cortical bone; neither specifically assess of the bone density in the vertebral body, which is where most of the osteoporotic fractures occur. Quantitative computed tomography (QCT) is an alternative technique that overcomes limitations of DXA-based diagnosis. However, due to the high cost and radiation dose, QCT is not used for osteoporosis management. We propose a method that provides a 3-D subject-specific shape and density estimation of the lumbar spine from a single anteroposterior (AP) DXA image. A 3-D statistical shape and density model is built, using a training set of QCT scans, and registered onto the AP DXA image so that its projection matches it. Cortical and trabecular bone compartments are segmented using a model-based algorithm. Clinical measurements are performed at different bone compartments. Accuracy was evaluated by comparing DXA-derived to QCT-derived 3-D measurements for a validation set of 180 subjects. The shape accuracy was 1.51 mm at the total vertebra and 0.66 mm at the vertebral body. Correlation coefficients between DXA and QCT-derived measurements ranged from 0.81 to 0.97. The method proposed offers an insightful 3-D analysis of the lumbar spine, which could potentially improve osteoporosis and fracture risk assessment in patients who had an AP DXA scan of the lumbar spine without any additional examination.

Active Shape Models are a powerful and well known method to perform face alignment. In some applications it is common to have shape information available beforehand, such as previously detected landmarks. Introducing this prior knowledge to the statistical model may result of great advantage but it is challenging to maintain this priors unchanged once the statistical model constraints are applied. We propose a new weighted-regularized projection into the parameter space which allows us to obtain shapes that at the same time fulfill the imposed shape constraints and are plausible according to the statistical model. The performed experiments show how using this projection better performance than competing state of the art methods is achieved.

MOTIVATION: Biological mechanisms contributing to atherogenesis are multiple and complex. The early stage of atherosclerosis (AS) is characterized by the accumulation of low-density lipoprotein (LDL) droplets, leading to the creation of foam cells (FC). To address the difficulty to explore the dynamics of interactions that controls this process, this study aimed to develop a model of agents and infer on the most influential cell- and molecule-related parameters. RESULTS: FC started to accumulate after six to eight months of simulated hypercholesterolemia. A sensitivity analysis revealed the strong influence of LDL oxidation rate on the risk of FC creation, which was exploited to model the antioxidant effect of statins. Combined with an empirical simulation of the drug ability to decrease the level of LDL, the virtual statins treatment led to reductions of oxidized LDL levels similar to reductions measured in vivo. AVAILABILITY AND IMPLEMENTATION: An Open source software was used to develop the agent-based model of early AS. Two different concentrations of LDL agents were imposed in the intima layer to simulate healthy and hypercholesterolemia groups of 'virtual patients'. The interactions programmed between molecules and cells were based on experiments and models reported in the literature. A factorial sensitivity analysis explored the respective effects of the less documented model parameters as (i) agent migration speed, (ii) LDL oxidation rate and (iii) concentration of autoantibody agents. Finally, the response of the model to known perturbations was assessed by introducing statins agents, able to reduce the oxidation rate of LDL agents and the LDL boundary concentrations. CONTACT: [email protected] information: Supplementary data are available at Bioinformatics online.

Studies have demonstrated the feasibility of late Gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging for guiding the management of patients with sequelae to myocardial infarction, such as ventricular tachycardia and heart failure. Clinical implementation of these developments necessitates a reproducible and reliable segmentation of the infarcted regions. It is challenging to compare new algorithms for infarct segmentation in the left ventricle (LV) with existing algorithms. Benchmarking datasets with evaluation strategies are much needed to facilitate comparison. This manuscript presents a benchmarking evaluation framework for future algorithms that segment infarct from LGE CMR of the LV. The image database consists of 30 LGE CMR images of both humans and pigs that were acquired from two separate imaging centres. A consensus ground truth was obtained for all data using maximum likelihood estimation. Six widely-used fixed-thresholding methods and five recently developed algorithms are tested on the benchmarking framework. Results demonstrate that the algorithms have better overlap with the consensus ground truth than most of the n-SD fixed-thresholding methods, with the exception of the Full-Width-at-Half-Maximum (FWHM) fixed-thresholding method. Some of the pitfalls of fixed thresholding methods are demonstrated in this work. The benchmarking evaluation framework, which is a contribution of this work, can be used to test and benchmark future algorithms that detect and quantify infarct in LGE CMR images of the LV. The datasets, ground truth and evaluation code have been made publicly available through the website: https://www.cardiacatlas.org/web/guest/challenges.

Better understanding of the anatomical variability of the human cochlear is important for the design and function of Cochlear Implants. Proper non-rigid alignment of high-resolution cochlear μ CT data is a challenge for the typical cubic B-spline registration model. In this paper we study one way of incorporating skeleton-based similarity as an anatomical registration prior. We extract a centerline skeleton of the cochlear spiral, and generate corresponding parametric pseudo-landmarks between samples. These correspondences are included in the cost function of a typical cubic B-spline registration model to provide a more global guidance of the alignment. The resulting registrations are evaluated using different metrics for accuracy and model behavior, and compared to the results of a registration without the prior.

Correspondence establishment is a key step in statistical shape model building. There are several automated methods for solving this problem in 3D, but they usually can only handle objects with simple topology, like that of a sphere or a disc. We propose an extension to correspondence establishment over a population based on the optimization of the minimal description length function, allowing considering objects with arbitrary topology. Instead of using a fixed structure of kernel placement on a sphere for the systematic manipulation of point landmark positions, we rely on an adaptive, hierarchical organization of surface patches. This hierarchy can be built on surfaces of arbitrary topology and the resulting patches are used as a basis for a consistent, multi-scale modification of the surfaces' parameterization, based on point distribution models. The feasibility of the approach is demonstrated on synthetic models with different topologies.

Automatic identification and extraction of bone contours from x-ray images is an essential first step task for further medical image analysis. In this paper we propose a 3D statistical model based framework for the proximal femur contour extraction from calibrated x-ray images. The automatic initialization is solved by an Estimation of Bayesian Network Algorithm to fit a multiple component geometrical model to the x-ray data. The contour extraction is accomplished by a non-rigid 2D/3D registration between a 3D statistical model and the x-ray images, in which bone contours are extracted by a graphical model based Bayesian inference. Preliminary experiments on clinical data sets verified its validity.

We present a method for the analysis of basal ganglia (including the thalamus) for accurate detection of human spongiform encephalopathy in multisequence magnetic resonance imaging (MRI) of the brain. One common feature of most forms of prion protein diseases is the appearance of hyperintensities in the deep grey matter area of the brain in T2-weighted magnetic resonance (MR) images. We employ T1, T2, and Flair-T2 MR sequences for the detection of intensity deviations in the internal nuclei. First, the MR data are registered to a probabilistic atlas and normalized in intensity. Then smoothing is applied with edge enhancement. The segmentation of hyperintensities is performed using a model of the human visual system. For more accurate results, a priori anatomical data from a segmented atlas are employed to refine the registration and remove false positives. The results are robust over the patient data and in accordance with the clinical ground truth. Our method further allows the quantification of intensity distributions in basal ganglia. The caudate nuclei are highlighted as main areas of diagnosis of sporadic Creutzfeldt-Jakob Disease (sCJD), in agreement with the histological data. The algorithm permitted the classification of the intensities of abnormal signals in sCJD patient FLAIR images with a higher hypersignal in caudate nuclei (10/10) and putamen (6/10) than in thalami. Defining normalized MRI measures of the intensity relations between the internal grey nuclei of patients, we robustly differentiate sCJD and variant CJD (vCJD) patients, in an attempt to create an automatic classification tool of human spongiform encephalopathies.