arXiv subjects

Oscar Camara

Publications and source records attributed to Oscar Camara.

At least 19 recordsLinked to original sources

arXiv2026

The TopCoW Challenge -- Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

The Circle of Willis (CoW) is an important network of arteries connecting major circulations of the brain. Its vascular architecture is believed to influence the risk, severity, and outcome of serious neurovascular diseases. However, characterizing the highly variable CoW anatomy remains a manual and time-consuming expert task. The CoW is commonly imaged by two non-invasive angiographic imaging modalities, magnetic resonance angiography (MRA) and computed tomography angiography (CTA), yet few datasets with annotated CoW anatomy exist, and there have been no established benchmarks for comparing CoW segmentation algorithms. We organized the TopCoW benchmark challenge alongside the release of an annotated CoW dataset with 125 paired MRA and CTA scans from the same patients. Voxel-level annotations for 13 vessel components were created using virtual reality technology and verified by clinical experts. Participants submitted algorithms for CoW segmentation and variant classification, which we evaluated on internal and external test sets comprising 226 scans from over five centers. The benchmark includes voxel-level segmentation, CoW component detection, CoW variant classification, and two clinical application tasks. We received submissions from over 250 participants across six continents. Top-performing teams achieved over 90% Dice scores for CoW segmentation, over 80% F1 scores for detecting key vessel components, and over 70% balanced accuracy in CoW variant classification across nearly all test sets. The best algorithms also supported clinically relevant downstream tasks by accurately classifying fetal-type posterior cerebral arteries and localizing aneurysms in relation to CoW anatomy. This benchmark demonstrated the utility of CoW segmentation algorithms for some downstream clinical applications with explainability.

cs.CV↗

arXiv2026

Open-source MRI-informed computational model of human cortical folding

The human cerebral cortex, initially smooth, progressively folds during fetal brain development in utero, giving rise to cortical convolutions. Atypical cortical folding patterns can be associated with neurodevelopmental and neurological disorders. To better understand these conditions, it is crucial to first examine the factors governing healthy cortical folding. Computational modeling provides a powerful way for this purpose and has already helped understanding the influence of key biomechanical parameters on the folding pattern. However, most existing models use simplified geometries, limiting calibration and validation with fetal and neonatal brain Magnetic Resonance Imaging (MRI) and neglecting the influence of initial geometry on fold development. On the other hand, simulations on realistic brain geometries introduce additional challenges, including collision handling, fold characterization, and additional computational cost. Furthermore, model parameters are often difficult to interpret, complicating comparison, clinical translation, and calibration. Finally, computational models of cortical folding also remain rarely accessible. In this work, we introduce a novel computational model of cortical folding, developed using the open-source code FEniCS to simulate folding on a whole-brain geometry generated from fetal MRI data. We also propose a modular, interpretable, and scalable simulation framework built around this computational model and openly available to the community. It uses fetal MRI data to generate realistic input brain meshes and estimate key biomechanical parameters such as cortical growth rate. The framework also integrates a spectral metric for cortical surface analysis to optimize folding pattern predictions from an healthy fetal MRI dataset.

physics.bio-ph↗

arXiv2026

On Safer Reinforcement Learning for Sedation and Analgesia in Intensive Care

Pain management in intensive care usually involves complex trade-offs, since both inadequate and excessive treatment can compromise patient safety. Prior work on reinforcement learning for sedation and analgesia has explored how to optimize these interventions, but has not considered patient survival or partial observability. To investigate the risks of these design choices, we developed an offline deep reinforcement learning framework that suggests hourly medication doses based on recurrent state representations. Using retrospective data from 47,144 ICU stays in the MIMIC-IV database, we trained and evaluated behavior-regularized actor-critic models that prescribe continuous doses of opioids, propofol, benzodiazepines, and dexmedetomidine according to two goals: reduce pain or jointly reduce pain and 30-day post-discharge mortality. Although the two resulting policies were associated with lower pain, clinician agreement with the pain-only policy was positively correlated with mortality ($ρ$=0.119, p<0.0001), while agreement with the joint policy was negatively correlated ($ρ$=-0.316, p<0.0001). We found that such divergence arose from a different response to high levels of comorbidity. This suggests that valuing post-discharge outcomes could be critical for learning safer treatment policies, even if a short-term goal remains the primary objective.

cs.LG↗

arXiv2026

Fetpype: An Open-Source Pipeline for Reproducible Fetal Brain MRI Analysis

Fetal brain magnetic resonance imaging (MRI) is crucial for assessing neurodevelopment in utero. However, fetal MRI analysis remains technically challenging due to fetal motion, low signal-to-noise ratio, and the need for complex multi-step processing pipelines. These pipelines typically include motion correction, super-resolution reconstruction, tissue segmentation, and cortical surface extraction. While specialized tools exist for each individual processing step, integrating them into a robust, reproducible, and user-friendly end-to-end workflow remains difficult. This fragmentation limits reproducibility across studies and hinders the adoption of advanced fetal neuroimaging methods in both research and clinical contexts. Fetpype addresses this gap by providing a standardized, modular, and reproducible framework for fetal brain MRI preprocessing and analysis, enabling researchers to process raw T2-weighted acquisitions through to derived volumetric and surface-based outputs within a unified workflow. Fetpype is publicly available on GitHub at https://github.com/fetpype/fetpype.

eess.IV↗

arXiv2025

A public cardiac CT dataset featuring the left atrial appendage

Despite the success of advanced segmentation frameworks such as TotalSegmentator (TS), accurate segmentations of the left atrial appendage (LAA), coronary arteries (CAs), and pulmonary veins (PVs) remain a significant challenge in medical imaging. In this work, we present the first open-source, anatomically coherent dataset of curated, high-resolution segmentations for these structures, supplemented with whole-heart labels produced by TS on the publicly available ImageCAS dataset consisting of 1000 cardiac computed tomography angiography (CCTA) scans. One purpose of the data set is to foster novel approaches to the analysis of LAA morphology. LAA segmentations on ImageCAS were generated using a state-of-the-art segmentation framework developed specifically for high resolution LAA segmentation. We trained the network on a large private dataset with manual annotations provided by medical readers guided by a trained cardiologist and transferred the model to ImageCAS data. CA labels were improved from the original ImageCAS annotations, while PV segmentations were refined from TS outputs. In addition, we provide a list of scans from ImageCAS that contains common data flaws such as step artefacts, LAAs extending beyond the scanner's field of view, and other types of data defects.

cs.CV↗

arXiv2025

A Computational Pipeline for Advanced Analysis of 4D Flow MRI in the Left Atrium

The left atrium (LA) plays a pivotal role in modulating left ventricular filling, but our comprehension of its hemodynamics is significantly limited by the constraints of conventional ultrasound analysis. 4D flow magnetic resonance imaging (4D Flow MRI) holds promise for enhancing our understanding of atrial hemodynamics. However, the low velocities within the LA and the limited spatial resolution of 4D Flow MRI make analyzing this chamber challenging. Furthermore, the absence of dedicated computational frameworks, combined with diverse acquisition protocols and vendors, complicates gathering large cohorts for studying the prognostic value of hemodynamic parameters provided by 4D Flow MRI. In this study, we introduce the first open-source computational framework tailored for the analysis of 4D Flow MRI in the LA, enabling comprehensive qualitative and quantitative analysis of advanced hemodynamic parameters. Our framework proves robust to data from different centers of varying quality, producing high-accuracy automated segmentations (Dice $>$ 0.9 and Hausdorff 95 $<$ 3 mm), even with limited training data. Additionally, we conducted the first comprehensive assessment of energy, vorticity, and pressure parameters in the LA across a spectrum of disorders to investigate their potential as prognostic biomarkers.

cs.CV↗

arXiv2025

NUDF: Neural Unsigned Distance Fields for high resolution 3D medical image segmentation

Medical image segmentation is often considered as the task of labelling each pixel or voxel as being inside or outside a given anatomy. Processing the images at their original size and resolution often result in insuperable memory requirements, but downsampling the images leads to a loss of important details. Instead of aiming to represent a smooth and continuous surface in a binary voxel-grid, we propose to learn a Neural Unsigned Distance Field (NUDF) directly from the image. The small memory requirements of NUDF allow for high resolution processing, while the continuous nature of the distance field allows us to create high resolution 3D mesh models of shapes of any topology (i.e. open surfaces). We evaluate our method on the task of left atrial appendage (LAA) segmentation from Computed Tomography (CT) images. The LAA is a complex and highly variable shape, being thus difficult to represent with traditional segmentation methods using discrete labelmaps. With our proposed method, we are able to predict 3D mesh models that capture the details of the LAA and achieve accuracy in the order of the voxel spacing in the CT images.

eess.IV↗

arXiv2024

Unsupervised stratification of patients with myocardial infarction based on imaging and in-silico biomarkers

This study presents a novel methodology for stratifying post-myocardial infarction patients at risk of ventricular arrhythmias using patient-specific 3D cardiac models derived from late gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR) images. The method integrates imaging and computational simulation with a simplified cellular automaton model, Arrhythmic3D, enabling rapid and accurate VA risk assessment in clinical timeframes. Applied to 51 patients, the model generated thousands of personalized simulations to evaluate arrhythmia inducibility and predict VA risk. Key findings include the identification of slow conduction channels (SCCs) within scar tissue as critical to reentrant arrhythmias and the localization of high-risk zones for potential intervention. The Arrhythmic Risk Score (ARRISK), developed from simulation results, demonstrated strong concordance with clinical outcomes and outperformed traditional imaging-based risk stratification. The methodology is fully automated, requiring minimal user intervention, and offers a promising tool for improving precision medicine in cardiac care by enhancing patient-specific arrhythmia risk assessment and guiding treatment strategies.

cs.CE↗

arXiv2024

Spatio-temporal neural distance fields for conditional generative modeling of the heart

The rhythmic pumping motion of the heart stands as a cornerstone in life, as it circulates blood to the entire human body through a series of carefully timed contractions of the individual chambers. Changes in the size, shape and movement of the chambers can be important markers for cardiac disease and modeling this in relation to clinical demography or disease is therefore of interest. Existing methods for spatio-temporal modeling of the human heart require shape correspondence over time or suffer from large memory requirements, making it difficult to use for complex anatomies. We introduce a novel conditional generative model, where the shape and movement is modeled implicitly in the form of a spatio-temporal neural distance field and conditioned on clinical demography. The model is based on an auto-decoder architecture and aims to disentangle the individual variations from that related to the clinical demography. It is tested on the left atrium (including the left atrial appendage), where it outperforms current state-of-the-art methods for anatomical sequence completion and generates synthetic sequences that realistically mimics the shape and motion of the real left atrium. In practice, this means we can infer functional measurements from a static image, generate synthetic populations with specified demography or disease and investigate how non-imaging clinical data effect the shape and motion of cardiac anatomies.

cs.CV↗

arXiv2024

Signed Distance Field based Segmentation and Statistical Shape Modelling of the Left Atrial Appendage

Patients with atrial fibrillation have a 5-7 fold increased risk of having an ischemic stroke. In these cases, the most common site of thrombus localization is inside the left atrial appendage (LAA) and studies have shown a correlation between the LAA shape and the risk of ischemic stroke. These studies make use of manual measurement and qualitative assessment of shape and are therefore prone to large inter-observer discrepancies, which may explain the contradictions between the conclusions in different studies. We argue that quantitative shape descriptors are necessary to robustly characterize LAA morphology and relate to other functional parameters and stroke risk. Deep Learning methods are becoming standardly available for segmenting cardiovascular structures from high resolution images such as computed tomography (CT), but only few have been tested for LAA segmentation. Furthermore, the majority of segmentation algorithms produces non-smooth 3D models that are not ideal for further processing, such as statistical shape analysis or computational fluid modelling. In this paper we present a fully automatic pipeline for image segmentation, mesh model creation and statistical shape modelling of the LAA. The LAA anatomy is implicitly represented as a signed distance field (SDF), which is directly regressed from the CT image using Deep Learning. The SDF is further used for registering the LAA shapes to a common template and build a statistical shape model (SSM). Based on 106 automatically segmented LAAs, the built SSM reveals that the LAA shape can be quantified using approximately 5 PCA modes and allows the identification of two distinct shape clusters corresponding to the so-called chicken-wing and non-chicken-wing morphologies.

cs.CV↗

arXiv2024

Multi-Center Fetal Brain Tissue Annotation (FeTA) Challenge 2022 Results

Segmentation is a critical step in analyzing the developing human fetal brain. There have been vast improvements in automatic segmentation methods in the past several years, and the Fetal Brain Tissue Annotation (FeTA) Challenge 2021 helped to establish an excellent standard of fetal brain segmentation. However, FeTA 2021 was a single center study, and the generalizability of algorithms across different imaging centers remains unsolved, limiting real-world clinical applicability. The multi-center FeTA Challenge 2022 focuses on advancing the generalizability of fetal brain segmentation algorithms for magnetic resonance imaging (MRI). In FeTA 2022, the training dataset contained images and corresponding manually annotated multi-class labels from two imaging centers, and the testing data contained images from these two imaging centers as well as two additional unseen centers. The data from different centers varied in many aspects, including scanners used, imaging parameters, and fetal brain super-resolution algorithms applied. 16 teams participated in the challenge, and 17 algorithms were evaluated. Here, a detailed overview and analysis of the challenge results are provided, focusing on the generalizability of the submissions. Both in- and out of domain, the white matter and ventricles were segmented with the highest accuracy, while the most challenging structure remains the cerebral cortex due to anatomical complexity. The FeTA Challenge 2022 was able to successfully evaluate and advance generalizability of multi-class fetal brain tissue segmentation algorithms for MRI and it continues to benchmark new algorithms. The resulting new methods contribute to improving the analysis of brain development in utero.

eess.IV↗

arXiv2023

Evaluation of cinematic volume rendering open-source and commercial solutions for the exploration of congenital heart data

Detailed anatomical information is essential to optimize medical decisions for surgical and pre-operative planning in patients with congenital heart disease. The visualization techniques commonly used in clinical routine for the exploration of complex cardiac data are based on multi-planar reformations, maximum intensity projection, and volume rendering, which rely on basic lighting models prone to image distortion. On the other hand, cinematic rendering (CR), a three-dimensional visualization technique based on physically-based rendering methods, can create volumetric images with high fidelity. However, there are a lot of parameters involved in CR that affect the visualization results, thus being dependent on the user's experience and requiring detailed evaluation protocols to compare available solutions. In this study, we have analyzed the impact of the most relevant parameters in a CR pipeline developed in the open-source version of the MeVisLab framework for the visualization of the heart anatomy of three congenital patients and two adults from CT images. The resulting visualizations were compared to a commercial tool used in the clinics with a questionnaire filled in by clinical users, providing similar definitions of structures, depth perception, texture appearance, realism, and diagnostic ability.

cs.GR↗

arXiv2023

Unsupervised Segmentation of Fetal Brain MRI using Deep Learning Cascaded Registration

Accurate segmentation of fetal brain magnetic resonance images is crucial for analyzing fetal brain development and detecting potential neurodevelopmental abnormalities. Traditional deep learning-based automatic segmentation, although effective, requires extensive training data with ground-truth labels, typically produced by clinicians through a time-consuming annotation process. To overcome this challenge, we propose a novel unsupervised segmentation method based on multi-atlas segmentation, that accurately segments multiple tissues without relying on labeled data for training. Our method employs a cascaded deep learning network for 3D image registration, which computes small, incremental deformations to the moving image to align it precisely with the fixed image. This cascaded network can then be used to register multiple annotated images with the image to be segmented, and combine the propagated labels to form a refined segmentation. Our experiments demonstrate that the proposed cascaded architecture outperforms the state-of-the-art registration methods that were tested. Furthermore, the derived segmentation method achieves similar performance and inference time to nnU-Net while only using a small subset of annotated data for the multi-atlas segmentation task and none for training the network. Our pipeline for registration and multi-atlas segmentation is publicly available at https://github.com/ValBcn/CasReg.

cs.CV↗

arXiv2023

On The Importance of Fundamental Computational Fluid Dynamics Towards a Robust and Reliable Model of Left Atrial Flows: Is There More Than Meets the Eye?

Computational fluid dynamics (CFD) studies of left atrial flows have reached a sophisticated level, e.g., revealing plausible relationships between hemodynamics and stresses with atrial fibrillation. However, little focus has been on fundamental fluid modelling of LA flows. The purpose of this study was to investigate the spatiotemporal convergence, along with the differences between high- (HR) versus normal-resolution/accuracy (NR) solution strategies, respectively. CFD simulations on 12 patient-specific left atrial geometries, obtained from computed tomography scans, were performed by using a second-order accurate and space/time centered solver. The convergence studies showed an average variability of around 30\% and 55\% for time averaged wall shear stress (WSS), oscillatory shear index (OSI), relative residence time (RRT), and endothelial cell activation potential (ECAP), even between intermediate spatial and temporal resolutions, in the left atrium (LA) and left atrial appendage (LAA), respectively. The comparison between HR and NR simulations showed good correlation in the LA for WSS, RRT, and ECAP (R^2 > 0.9), but not for OSI (R^2 = 0.63). However, there were poor correlations in the LAA especially for OSI, RRT, and ECAP (R^2 = 0.55, 0.63, and 0.61, respectively), except for WSS (R^2 = 0.81). The errors are comparable to differences previously reported with disease correlations. To robustly predict atrial hemodynamics and stresses, numerical resolutions of 10M elements and 10k time-steps per cycle seem necessary (i.e., one order of magnitude higher than normally used in both space and time). In conclusion, attention to fundamental numerical aspects is essential towards establishing a plausible, robust, and reliable model of LA flows.

physics.flu-dyn↗

arXiv2022

Attri-VAE: attribute-based interpretable representations of medical images with variational autoencoders

Deep learning (DL) methods where interpretability is intrinsically considered as part of the model are required to better understand the relationship of clinical and imaging-based attributes with DL outcomes, thus facilitating their use in the reasoning behind medical decisions. Latent space representations built with variational autoencoders (VAE) do not ensure individual control of data attributes. Attribute-based methods enforcing attribute disentanglement have been proposed in the literature for classical computer vision tasks in benchmark data. In this paper, we propose a VAE approach, the Attri-VAE, that includes an attribute regularization term to associate clinical and medical imaging attributes with different regularized dimensions in the generated latent space, enabling a better-disentangled interpretation of the attributes. Furthermore, the generated attention maps explained the attribute encoding in the regularized latent space dimensions. Using the Attri-VAE approach we analyzed healthy and myocardial infarction patients with clinical, cardiac morphology, and radiomics attributes. The proposed model provided an excellent trade-off between reconstruction fidelity, disentanglement, and interpretability, outperforming state-of-the-art VAE approaches according to several quantitative metrics. The resulting latent space allowed the generation of realistic synthetic data in the trajectory between two distinct input samples or along a specific attribute dimension to better interpret changes between different cardiac conditions.

eess.IV↗

arXiv2022

Geometric Deep Learning for the Assessment of Thrombosis Risk in the Left Atrial Appendage

The assessment of left atrial appendage (LAA) thrombogenesis has experienced major advances with the adoption of patient-specific computational fluid dynamics (CFD) simulations. Nonetheless, due to the vast computational resources and long execution times required by fluid dynamics solvers, there is an ever-growing body of work aiming to develop surrogate models of fluid flow simulations based on neural networks. The present study builds on this foundation by developing a deep learning (DL) framework capable of predicting the endothelial cell activation potential (ECAP), linked to the risk of thrombosis, solely from the patient-specific LAA geometry. To this end, we leveraged recent advancements in Geometric DL, which seamlessly extend the unparalleled potential of convolutional neural networks (CNN), to non-Euclidean data such as meshes. The model was trained with a dataset combining 202 synthetic and 54 real LAA, predicting the ECAP distributions instantaneously, with an average mean absolute error of 0.563. Moreover, the resulting framework manages to predict the anatomical features related to higher ECAP values even when trained exclusively on synthetic cases.

cs.LG↗

arXiv2022

In-silico analysis of the influence of pulmonary vein configuration on left atrial haemodynamics and thrombus formation in a large cohort

Atrial fibrillation (AF) is considered the most common human arrhythmia. Around 99\% of thrombi in non-valvular AF are formed in the left atrial appendage (LAA). Studies suggest that abnormal LAA haemodynamics and the subsequently stagnated flow are the factors triggering clot formation. However, the relation between LAA morphology, the blood pattern and the triggering is not fully understood. Moreover, the impact of structures such as the pulmonary veins (PVs) on LA haemodynamics has not been thoroughly studied due to the difficulties of acquiring appropriate data. On the other hand, in-silico studies and flow simulations allow a thorough analysis of haemodynamics, analysing the 4D nature of blood flow patterns under different boundary conditions. However, the reduced number of cases reported on the literature of these studies has been a limitation. The main goal of this work was to study the influence of PVs on left atrium (LA) and LAA haemodynamics. Computational fluid dynamics simulations were run on 52 patients, the largest cohort so far in the literature, where different parameters were individually studied: pulmonary veins orientation and configuration; LAA and LA volumes and its ratio; and flow velocities. Our computational analysis showed how the right pulmonary vein height and angulation have a great influence on LA haemodynamics. Additionally, we found that LAA with great bending with its tip pointing towards the mitral valve could contribute to favour flow stagnation.

physics.med-ph↗

arXiv2022

Generalizing electrocardiogram delineation -- Training convolutional neural networks with synthetic data augmentation

Obtaining per-beat information is a key task in the analysis of cardiac electrocardiograms (ECG), as many downstream diagnosis tasks are dependent on ECG-based measurements. Those measurements, however, are costly to produce, especially in recordings that change throughout long periods of time. However, existing annotated databases for ECG delineation are small, being insufficient in size and in the array of pathological conditions they represent. This article delves has two main contributions. First, a pseudo-synthetic data generation algorithm was developed, based in probabilistically composing ECG traces given "pools" of fundamental segments, as cropped from the original databases, and a set of rules for their arrangement into coherent synthetic traces. The generation of conditions is controlled by imposing expert knowledge on the generated trace, which increases the input variability for training the model. Second, two novel segmentation-based loss functions have been developed, which attempt at enforcing the prediction of an exact number of independent structures and at producing closer segmentation boundaries by focusing on a reduced number of samples. The best performing model obtained an $F_1$-score of 99.38\% and a delineation error of $2.19 \pm 17.73$ ms and $4.45 \pm 18.32$ ms for all wave's fiducials (onsets and offsets, respectively), as averaged across the P, QRS and T waves for three distinct freely available databases. The excellent results were obtained despite the heterogeneous characteristics of the tested databases, in terms of lead configurations (Holter, 12-lead), sampling frequencies ($250$, $500$ and $2,000$ Hz) and represented pathophysiologies (e.g., different types of arrhythmias, sinus rhythm with structural heart disease), hinting at its generalization capabilities, while outperforming current state-of-the-art delineation approaches.

cs.LG↗