Publications

dc.title: A-eye: Automated 3D MRI segmentation and morphometric feature extraction for eye and orbit atlas construction dc.contributor.author: Barranco, Jaime; Luyken, Adrian Konstantin; Jia, Yiwei; Kebiri, Hamza; Stachs, Philipp; Gordaliza, Pedro M; Esteban, Oscar; Aleman, Yasser; Sznitman, Raphael; Streckenbach, Felix; Stachs, Oliver; Langner, Sönke; Franceschiello, Benedetta; Cuadra, Meritxell Bach dc.description.abstract: In this study we introduce automated 3D segmentation of the healthy human adult eye and orbit from Magnetic Resonance Images, to improve ophthalmic diagnostics and treatments. Past efforts have primarily focused on small sample sizes and varied imaging modalities. Here, we leverage a large-scale dataset of T1-weighted MRI of 1245 subjects and the deep learning-based nnU-Net for MR-Eye segmentation tasks. The results showcase robust and accurate 3D segmentation of lens, globe, optic nerve, rectus muscles, and orbital fat. We also present the automated estimation of key ophthalmic morphometry biomarkers such as axial length and volumetry, while benchmarking correlations between body mass index and eye structure volumes. Quality control protocols are introduced through the pipeline to ensure the reliability of the segmented large-scale data, further enhancing the applicability of our algorithm in clinical research. As a major outcome we provide the first large-scale unbiased eye atlases (female, male, and combined) towards standardization of spatial normalization tools for MR-Eye. Copyright: © 2026 Barranco et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

dc.title: High-resolution whole-brain magnetic resonance spectroscopic imaging in youth at risk for psychosis dc.contributor.author: Céléreau, Edgar; Lucchetti, Federico; Alemán-Gómez, Yasser; Dwir, Daniella; Cleusix, Martine; Ledoux, Jean-Baptiste; Jenni, Raoul; Conchon, Caroline; Bach Cuadra, Meritxell; Schilliger, Zoé; Solida, Alessandra; Armando, Marco; Plessen, Kerstin Jessica; Hagmann, Patric; Conus, Philippe; Klauser, Antoine; Klauser, Paul dc.description.abstract: Advances in three-dimensional magnetic resonance spectroscopic imaging (3D-MRSI) allow for the high-resolution mapping of multiple neurometabolites throughout the entire brain in vivo and within clinically compatible time frames. Leveraging this capability, we created a voxel-based pipeline that corrects and spatially normalizes whole-brain maps of total N-acetylaspartate (tNAA), myo-inositol (Ins), choline compounds (Cho), glutamate + glutamine (Glx) and creatine + phosphocreatine (tCr). We examined 2 different 3D-MRSI dataset: first, a clinical sample of adolescents and young adults at risk for psychosis (n = 21) meeting DSM-5 criteria for Attenuated Psychosis Syndrome (APS) or Schizotypal Personality Disorder (SCZT), and age-/sex-matched healthy controls (n = 13); and second, a non-clinical sample of adolescents (n = 61) scanned on a different site. The objective of the study was threefold: first, to assess the reproducibility of 3D-MRSI measures across datasets and scanning sites; second, to validate the feasibility of whole-brain, voxel-based analyses on 3D-MRSI data; and third, to test the sensitivity of this approach. Metabolite distributions showed reproducible regional variation in standard space between the two independent samples and scanning sites (r ranging from 0.82 to 0.99). Relative to controls, at-risk participants exhibited higher tNAA levels in frontal grey matter; the SCZT subgroup additionally displayed widespread cortical and subcortical elevations of Ins levels compared with both APS and controls. Voxel-based analyses of structural (i.e., gray and white matter volumes or densities) and diffusion (i.e., generalized fractional anisotropy) parameters yielded no significant differences between patients and controls. These preliminary findings suggest that high-resolution 3D-MRSI may be sensitive enough to detect subtle neurometabolic alterations at the group level in the early stages of psychotic disorders when structural or diffusion measures show no difference. High-resolution whole-brain metabolic mapping may have the potential to help with early identification of young people at risk for psychosis or other mental disorders. © 2026 The Authors. Published under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.

dc.title: Fetpype: An Open-Source Pipeline for Reproducible Fetal Brain MRI Analysis dc.contributor.author: Sanchez, Thomas; Martí-Juan, Gerard; Meunier, David; Ballester, Miguel Angel Gonzalez; Camara, Oscar; Eixarch, Elisenda; Piella, Gemma; Cuadra, Meritxell Bach; Auzias, Guillaume dc.description.abstract: 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

dc.title: A comparative study of deep learning for cortical lesion MRI segmentation with explainability analysis in multiple sclerosis dc.contributor.author: Molchanova, Nataliia; Cagol, Alessandro; Ocampo-Pineda, Mario; Lu, Po-Jui; Weigel, Matthias; Chen, Xinjie; Beck, Erin S; Tsagkas, Charidimos; Reich, Daniel S; Bulcke, Colin Vanden; Stölting, Anna; Borrelli, Serena; Maggi, Pietro; Lugo, Sebastian Baez; Lemay, Delphine Ribes; Depeursinge, Adrien; Granziera, Cristina; Müller, Henning; Gordaliza, Pedro M; Bach Cuadra, Meritxell dc.description.abstract: Cortical lesions (CLs) have emerged as valuable biomarkers in multiple sclerosis (MS), offering high diagnostic specificity and prognostic relevance. However, their routine clinical integration remains limited due to subtle magnetic resonance imaging (MRI) appearance, challenges in expert annotation, and a lack of standardized automated methods. We present a multi-centric comparative study of CL detection and segmentation in MRI. A total of 656 MRI scans, including clinical trial and research data from four institutions, were acquired at 3T and 7T using MP2RAGE and MPRAGE sequences with expert-consensus annotations. We rely on the self-configuring nnU-Net framework, designed for medical imaging segmentation, and propose adaptations tailored to the improved CL detection. We evaluated model generalization through out-of-distribution testing, demonstrating promising lesion detection capabilities with an F1-score of 0.64 and 0.5 in and out of the domain, respectively. We also analyze internal model features and model errors for a better understanding of AI decision-making. Our study examines how data variability, lesion ambiguity, and protocol differences impact model performance, offering future recommendations to address these barriers to clinical adoption. Furthermore, we designed and implemented a medical expert questionnaire for better assessment of clinical value of the model predictions. To reinforce the reproducibility, the implementation and models will be publicly accessible and ready to use at GitHub and Zenodo. Copyright © 2026 The Authors. Published by Elsevier Inc. All rights reserved.

dc.title: Standardized methodology for assessing the presence, variants and area of the interthalamic adhesion using anatomical MRI (SNAP-IA): multicentric validation on 565 healthy individuals and multiple neurological disorders dc.contributor.author: Vidal, Julie P; Forno, Gonzalo; Hornberger, Michael; Cuadra, Meritxell Bach; Danet, Lola; Kumar, Vinod J; Péran, Patrice; Tourdias, Thomas; Barbeau, Emmanuel J dc.description.abstract: The interthalamic adhesion (IA) connects both thalami. Emerging research suggests it may support thalamo-cortical connectivity and could be involved in neurodevelopmental and neuropsychiatric conditions. However, inconsistent MRI evaluation hinders progress on this subject. We developed SNAP-IA, a standardized anatomical imaging protocol for consistent IA identification and quantification. This work leveraged the expertise from seven research teams (Toulouse, Santiago, Southampton, Lausanne, Tübingen, and Bordeaux). SNAP-IA includes three steps: (1) determination of IA presence/absence on T1-weighted MRI; (2) classification of IA variants (simple, broad, double, bilobar, and filiform); (3) segmentation-based area assessment. It was tested on 500 controls (20–69 yo) and patients (stroke, schizophrenia, bipolar disorder, and ADHD) with 0.6–1 mm isotropic T1-weighted MRI (3T to 9.4T). SNAP-IA application achieved high inter-dataset agreement (mean Dice ≈ 0.92), with an average identification time of 35 s. The IA was absent in 22.8% of controls. Simple and broad variants constituted 95% of identified IA while some variants (double, filiform) were observed less frequently. At 3T, females had a higher presence rate (84.4%) than males (69.8%) and a larger IA area. ANCOVA indicated that both age and gender were highly predictive of IA area, decreasing by 0.25 mm²/year. At 9.4T, absence rates were significantly higher (34.6%) than at 3T (18.1%, p = 0.002). Mean IA area did not differ significantly between 3T and 9.4T. Patients with neurodevelopmental or neuropsychiatric disorders had two times less IA presence, with significantly smaller IA. SNAP-IA provides a reliable, reproducible framework for anatomical IA assessment across populations and MRI sequences, enabling future research into its structural and functional roles and supporting automated, large-scale AI studies. The online version contains supplementary material available at 10.1007/s00429-026-03097-6.

dc.title: Advances in automated fetal brain MRI segmentation and biometry: Insights from the FeTA 2024 challenge dc.contributor.author: Zalevskyi, Vladyslav; Sanchez, Thomas; Kaandorp, Misha; Roulet, Margaux; Fajardo-Rojas, Diego; Li, Liu; Hutter, Jana; Li, Hongwei Bran; Barkovich, Matthew J; Ji, Hui; Wilhelmi, Luca; Dändliker, Aline; Steger, Céline; Koob, Mériam; Gomez, Yvan; Jakovčić, Anton; Klaić, Melita; Adžić, Ana; Marković, Pavel; Grabarić, Gracia; Rados, Milan; Aviles Verdera, Jordina; Kasprian, Gregor; Dovjak, Gregor; Gaubert-Rachmühl, Raphael; Aschwanden, Maurice; Zeng, Qi; Karimi, Davood; Peruzzo, Denis; Ciceri, Tommaso; Longari, Giorgio; Hamadache, Rachika E; Bouzid, Amina; Lladó, Xavier; Chiarella, Simone; Martí-Juan, Gerard; González Ballester, Miguel Ángel; Castellaro, Marco; Pinamonti, Marco; Visani, Valentina; Cremese, Robin; Sam, Keïn; Gaudfernau, Fleur; Ahir, Param; Parikh, Mehul; Zenk, Maximilian; Baumgartner, Michael; Maier-Hein, Klaus; Tianhong, Li; Hong, Yang; Longfei, Zhao; Preloznik, Domen; Špiclin, Žiga; Won Choi, Jae; Li, Muyang; Fu, Jia; Wang, Guotai; Jiang, Jingwen; Tong, Lyuyang; Du, Bo; Gondova, Andrea; You, Sungmin; Im, Kiho; Qayyum, Abdul; Mazher, Moona; Niederer, Steven A; Jakab, Andras; Licandro, Roxane; Payette, Kelly; Bach Cuadra, Meritxell dc.description.abstract: 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. Copyright © 2026 The Author(s). Published by Elsevier B.V. All rights reserved.

dc.title: Trustworthy AI in medical image analysis: A unified perspective built on robustness and layers of trust dc.contributor.author: Zuluaga, Maria A.; Išgum, Ivana; Bach Cuadra, Meritxell

dc.title: Towards contrast- and pathology-agnostic clinical fetal brain MRI segmentation using SynthSeg dc.contributor.author: Shang, Ziyao; Kaandorp, Misha; Payette, Kelly; Fernandez Garcia, Marina; Licandro, Roxane; Langs, Georg; Aviles Verdera, Jordina; Hutter, Jana; Menze, Bjoern; Kasprian, Gregor; Bach Cuadra, Meritxell; Jakab, Andras dc.description.abstract: Magnetic resonance imaging (MRI) has played a crucial role in fetal neurodevelopmental research. Structural annotations of MR images are an important step for quantitative analysis of the developing human brain, with Deep Learning providing an automated alternative for this otherwise tedious manual process. However, segmentation performances of Convolutional Neural Networks often suffer from domain shift, where the network fails when applied to subjects that deviate from the distribution with which it is trained on. In this work, we aim to train networks capable of automatically segmenting fetal brain MRIs with a wide range of domain shifts pertaining to differences in subject physiology and acquisition environments, in particular shape-based differences commonly observed in pathological cases. We introduce a novel data-driven train-time sampling strategy that seeks to fully exploit the diversity of a given training dataset to enhance the domain generalizability of the trained networks. We adapted our sampler, together with other existing data augmentation techniques, to the SynthSeg framework, a generator that utilizes domain randomization to generate diverse training data. We ran thorough experimentations and ablation studies on a wide range of training/testing data to test the validity of the approaches. Our networks achieved notable improvements in the segmentation quality on testing subjects with intense anatomical abnormalities (p < 1e-4), though at the cost of a slighter decrease in performance in cases with fewer abnormalities. Our work also lays the foundation for future works on creating and adapting data-driven sampling strategies for other training pipelines. Copyright © 2026 The Author(s). Published by Elsevier Inc. All rights reserved.

dc.title: Instance-level quantitative saliency in multiple sclerosis lesion segmentation dc.contributor.author: Spagnolo, Federico; Molchanova, Nataliia; Cuadra, Meritxell Bach; Ocampo-Pineda, Mario; Melie-Garcia, Lester; Granziera, Cristina; Andrearczyk, Vincent; Depeursinge, Adrien dc.description.abstract: In recent years, explainable methods for artificial intelligence (XAI) have tried to reveal and describe models' decision mechanisms in the case of classification and even for segmentation. However, XAI methods for semantic segmentation and in particular for single specific instances (e.g. one given lesion among others of the same class in medical imaging) have yet to be developed to understand what drove the detection and contouring of the latter, which is crucial for all multi-lesional diseases. We proposed instance-level explanation maps for semantic segmentation extending both SmoothGrad and Grad-CAM++ methods and yielding quantitative instance saliency for the former. The instance-level methods were applied to the segmentation of white matter lesions (WML), a magnetic resonance imaging (MRI) biomarker in multiple sclerosis (MS). 687 patients diagnosed with MS for a total of 4023 FLAIR and MPRAGE MRI scans were collected at the University Hospital of Basel, Switzerland. WM lesion masks were annotated by four expert clinicians on baseline and follow-up imaging. Three deep learning networks-a 3D U-Net, nnU-Net, and Swin UNETR-were trained and tested on these data (test normalized Dice score, respectively of 0.71, 0.78, 0.80; true positive rate of 79%, 78%, and 85%; false discovery rate of 37%, 38%, and 36%; false negative rate of 20%, 22%, and 14%), then saliency maps were computed. Consistent with clinical practice, the proposed instance saliency maps revealed that the model relied more on FLAIR than MPRAGE to segment WMLs, with positive saliency values inside a lesion and negative in its neighborhood. FLAIR hyperintensity combined with healthy WM around the lesion border was required for their detection. Beyond the aforementioned sanity checks, we observed that peak values of the generated saliency maps presented distributions that differ significantly between TP, FN, FP and TN predictions, suggesting that the quantitative nature of the proposed saliency could be used to identify errors. In conclusion, we introduced two XAI methods to generate quantitative instance-level explanations in semantic segmentation. The proposed XAI maps can be applied to any architecture and could serve as a basis to (i) improve model performance (e.g. reducing FPs), (ii) optimize their internal architecture (e.g. patch size), and (iii) justify the model's decisions to the end users, which are contextualized to a specific lesion instance of interest. © 2026. The Author(s).

dc.title: Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis dc.contributor.author: Wynen, Maxence; Vanden Bulcke, Colin; Borrelli, Serena; Gordaliza, Pedro M; Stölting, Anna; Guisset, François; Cordier, Clément; Martire, Maria Sofia; Tamanti, Agnese; Macq, Benoit; Sati, Pascal; Filippi, Massimo; Calabrese, Massimiliano; Absinta, Martina; Reich, Daniel S; Bach Cuadra, Meritxell; Maggi, Pietro dc.description.abstract: Multiple sclerosis diagnostic criteria lack optimal specificity, leading to potential misdiagnosis. Advanced magnetic resonance imaging (MRI) biomarkers like the central vein sign, cortical lesions and paramagnetic rim lesions are highly specific to multiple sclerosis and could potentially improve diagnostic accuracy. In this study, we applied machine learning techniques to a retrospective, multicentric dataset of 322 multiple sclerosis/multiple sclerosis-mimic (204/118) and 84 prodromal multiple sclerosis/non-multiple sclerosis (43/41) adult patients, incorporating the central vein sign, cortical lesions and paramagnetic rim lesions. We compared (5 × 2 cross-validation combined F-test) the diagnostic performance of 71 machine learning models, each corresponding to a distinct combination of full-count or simplified biomarker inputs, against the baseline dissemination in space McDonald criteria. The aim was to evaluate the multiple sclerosis diagnostic power of combining these biomarkers in an MRI-only diagnostic framework. 51 of the 71 models significantly outperformed the dissemination in space criterion (P < 0.05), with balanced accuracy improvements up to 13.0% (confidence interval: [+10.5; +17.0]). The best overall model (random forest, using full-count assessments) achieved 95.7% (confidence interval: [93.2; 99.7]) balanced accuracy; the best simplified model (logistic regression, using only simplified assessments) reached 94.7% with no significant difference with the former (P = 0.29). Notably, 12/51 high-performing models used only simplified assessments. To further investigate the models' generalizability, external validation on two out-of-distribution test sets using bootstrapping (1000 resamples) confirmed these results and highlighted a more robust generalization for the best model using solely simplified biomarkers. On the first external test set (n = 37, Verona), the simplified model achieved 97.2% balanced accuracy, while the full-count model reached 93.3% (versus 83.3% for baseline). On the second test set (n = 84, prodromal cases), the simplified model achieved 92.6% (versus 60.1% for baseline) showing competitive performance against the full-count model (93.9%). Both models improved all key performance metrics-balanced accuracy, sensitivity, specificity, precision and F1 score-over the baseline on both test sets (all P < 0.0001). Within a non-invasive MRI-only diagnostic framework, these results show that the incorporation of advanced imaging biomarkers into the multiple sclerosis-MRI diagnostic criteria significantly enhances the diagnostic accuracy-a statement holding true even when using simplified central vein sign, cortical lesions and paramagnetic rim lesions assessments. The study also provides a publicly available online diagnostic tool, facilitating further interaction, validation and clinical support (https://www.msdiagnostictool.org). © The Author(s) 2026. Published by Oxford University Press on behalf of the Guarantors of Brain.