Pasar al contenido principal

TECNICAS AVANZADAS DE APRENDIZAJE PROFUNDO PARA EL DESARROLLO DE HERRAMIENTAS DE NEUROIMAGEN

PID2023-146187OB-I00

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprograma Subprograma Estatal de Generación de Conocimiento
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2023
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT DE GIRONA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 4
Encontrada(s) 1 página(s)

Repositorio Digital de la UPF
Publicación Artículo científico (JournalArticle)

Advances in automated fetal brain MRI segmentation and biometry: insights from the FeTA 2024 challenge

  • Zalevskyi, Vladyslav
  • Marti-Juan, Gerard
  • Gonzalez Ballester, Miguel Ángel
  • Bach Cuadra, Meritxell
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., This research was funded by the Swiss National Science Foundation (182602, 215641, IZKSZ3_218590), ERA-NET Neuron MULTI-FACT project (SNSF 31NE30 203977), UKRI FLF (MR/T018119/1), the Adaptive Brain Circuits in Development and Learning Project, University Research Priority Program of the University of Zürich; by the Vontobel Foundation; by the Anna Müller Grocholski Foundation and the Prof. Max Cloetta Foundation and DFG Heisenberg funding (502024488); we acknowledge the Leenaards and Jeantet Foundations as well as CIBM Center for Biomedical Imaging, a Swiss research center of excellence founded and supported by CHUV, UNIL, EPFL, UNIGE and HUG. Diego Fajardo-Rojas would like to acknowledge funding from the EPSRC Centre for Doctoral Training in Smart Medical Imaging (EP/S022104/1). Lyuyang Tong, Bo Du and Jingwen Jiang would like to acknowledge funding from the National Key Research and Development Program of China under Grants 2023YFC2705700, the National Natural Science Foundation of China under Grants 62,306,217 and 62225113, the Postdoctoral Fellowship Program of CPSF under Grant Number GZC20231987, the China Postdoctoral Science Foundation under Grant Number 2024T170686 and 2024M752471, the Major Program (JD) of Hubei Province (2023BAA017), the Innovative Research Group Project of Hubei Province under Grants 2024AFA017. Gerard Martí-Juan is supported by the project PCI2021-122044-2A, funded by the project ERA-NET NEURON Cofund2, by MCIN/AEI/10.13039/501100011033/ and by the European Union NextGenerationEU/PRTR. R. Hamadache holds an IFUdG PhD grant from the University of Girona. R. Hamadache and X. Lladó received support by the PID2023-146187OB-I00 from the Ministerio de Ciencia e Innovación, Spain. Kiho Im, A. Gondova, and S. You received support from the National Institute of Neurological Disorders and Stroke (R01NS114087) and National Institute of Biomedical Imaging and Bioengineering (R01EB031170) of the National Institutes of Health (NIH).
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Advances in automated fetal brain MRI segmentation and biometry: insights from the FeTA 2024 challenge

  • Zalevskyi, Vladyslav
  • Marti-Juan, Gerard
  • Gonzalez Ballester, Miguel Ángel
  • Bach Cuadra, Meritxell
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., This research was funded by the Swiss National Science Foundation (182602, 215641, IZKSZ3_218590), ERA-NET Neuron MULTI-FACT project (SNSF 31NE30 203977), UKRI FLF (MR/T018119/1), the Adaptive Brain Circuits in Development and Learning Project, University Research Priority Program of the University of Zürich; by the Vontobel Foundation; by the Anna Müller Grocholski Foundation and the Prof. Max Cloetta Foundation and DFG Heisenberg funding (502024488); we acknowledge the Leenaards and Jeantet Foundations as well as CIBM Center for Biomedical Imaging, a Swiss research center of excellence founded and supported by CHUV, UNIL, EPFL, UNIGE and HUG. Diego Fajardo-Rojas would like to acknowledge funding from the EPSRC Centre for Doctoral Training in Smart Medical Imaging (EP/S022104/1). Lyuyang Tong, Bo Du and Jingwen Jiang would like to acknowledge funding from the National Key Research and Development Program of China under Grants 2023YFC2705700, the National Natural Science Foundation of China under Grants 62,306,217 and 62225113, the Postdoctoral Fellowship Program of CPSF under Grant Number GZC20231987, the China Postdoctoral Science Foundation under Grant Number 2024T170686 and 2024M752471, the Major Program (JD) of Hubei Province (2023BAA017), the Innovative Research Group Project of Hubei Province under Grants 2024AFA017. Gerard Martí-Juan is supported by the project PCI2021-122044-2A, funded by the project ERA-NET NEURON Cofund2, by MCIN/AEI/10.13039/501100011033/ and by the European Union NextGenerationEU/PRTR. R. Hamadache holds an IFUdG PhD grant from the University of Girona. R. Hamadache and X. Lladó received support by the PID2023-146187OB-I00 from the Ministerio de Ciencia e Innovación, Spain. Kiho Im, A. Gondova, and S. You received support from the National Institute of Neurological Disorders and Stroke (R01NS114087) and National Institute of Biomedical Imaging and Bioengineering (R01EB031170) of the National Institutes of Health (NIH).
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Hematoma expansion prediction in intracerebral hemorrhage patients by using synthesized CT images in an end-to-end deep learning framework

  • Yalcin, Cansu
  • Abramova, Valeriia
  • Terceño Izaga, Mikel
  • Oliver i Malagelada, Arnau
  • Silva Blas, Yolanda
  • Lladó Bardera, Xavier
Spontaneous intracerebral hemorrhage (ICH) is a type of stroke less prevalent than ischemic stroke but associated with high mortality rates. Hematoma expansion (HE) is an increase in the bleeding that affects 30%–38% of hemorrhagic stroke patients. It is observed within 24 h of onset and associated with patient worsening. Clinically it is relevant to detect the patients that will develop HE from their initial computed tomography (CT) scans which could improve patient management and treatment decisions. However, this is a significant challenge due to the predictive nature of the task and its low prevalence, which hinders the availability of large datasets with the required longitudinal information. In this work, we present an end-to-end deep learning framework capable of predicting which cases will exhibit HE using only the initial basal image. We introduce a deep learning framework based on the 2D EfficientNet B0 model to predict the occurrence of HE using initial non-contrasted CT scans and their corresponding lesion annotation as priors. We used an in-house acquired dataset of 122 ICH patients, including 35 HE cases, containing longitudinal CT scans with manual lesion annotations in both basal and follow-up (obtained within 24 h after the basal scan). Experiments were conducted using a 5-fold cross-validation strategy. We addressed the limited data problem by incorporating synthetic images into the training process. To the best of our knowledge, our approach is novel in the field of HE prediction, being the first to use image synthesis to enhance results. We studied different scenarios such as training only with the original scans, using standard image augmentation techniques, and using synthetic image generation. The best performance was achieved by adding five generated versions of each image, along with standard data augmentation, during the training process. This significantly improved (p=0.0003) the performance obtained with our baseline model using directly the original CT scans from an Accuracy of 0.56 to 0.84, F1-Score of 0.53 to 0.82, Sensitivity of 0.51 to 0.77, and Specificity of 0.60 to 0.91, respectively. The proposed approach shows promising results in predicting HE, especially with the inclusion of synthetically generated images. The obtained results highlight the significance of this research direction, which has the potential to improve the clinical management of patients with hemorrhagic stroke. The code is available at: https://github.com/NIC-VICOROB/HE-prediction-SynthCT, Cansu Yalcin té una beca FI de la Generalitat de Catalunya amb número de referència 2023 FI-1 00096. Valeriia Abramova té una beca FPI del Ministeri de Ciència, Innovació i Universidades, Espanya amb número de referència PRE2021-099121. Aquest treball ha comptat amb el suport de DPI2020-114769RB-I00 i PID2023-146187OB-I00 del Ministerio de Ciencia e Innovación, Espanya, Finançament d'accés obert gràcies a l'acord CRUE-CSIC amb Elsevier
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Topology-aware multiclass segmentation of the Circle of Willis from MRA and CTA images

  • Hamadache, Rachika E.
  • Lisazo, Clara
  • Yalcin, Cansu
  • Lal-Trehan Estrada, Uma Maria
  • Abramova, Valeriia
  • Casamitjana, Adrià
  • Oliver i Malagelada, Arnau
  • Lladó Bardera, Xavier
The Circle of Willis (CoW) is an essential network of arteries that ensures blood flow throughout the brain. From a clinical perspective, evaluating the vessels of the CoW is highly relevant as its angioarchitecture and variants are important biomarkers of neurovascular pathologies. However, achieving a topologically accurate segmentation of these vessels remains challenging due to their anatomical complexity. In this work, we propose a pipeline for the multiclass segmentation of the CoW vessels (13 possible classes), focusing on achieving both topology correctness and segmentation accuracy in magnetic resonance angiography (MRA) and computed tomography angiography (CTA) imaging techniques. We propose a deep learning framework based on the nnUNet model, together with a post-processing block that requires no additional training and that is adapted to the specific multiclass CoW segmentation task. We train and validate our framework using the publicly available TopCoW 2024 dataset (MRA and CTA) and evaluate it on the hidden test set (through an online system) and on an independent subset from the CROWN 2023 challenge dataset (MRA). The obtained results demonstrate the positive impact of our approach, achieving an average Dice (centerline Dice) scores of 0.90 (0.99) for MRA and 0.88 (0.99) for CTA on the in-domain test set, and 0.81 (0.97) on the out-of-domain test set for MRA. These high performances align with state-of-the-art methods, and rank among the top in the TopCoW 2024 challenge. The approach is publicly available for the research community at https://github.com/NIC-VICOROB/CoW-multiclass-segmentation-TopCoW24, Rachika E. Hamadache holds an IFUdG2024 grant from Universitat de Girona. Clara Lisazo holds an FI grant from the Catalan Government with reference number 2024 FI-1 00103. Cansu Yalcın holds an FI grant from the Catalan Government with reference number 2023 FI-1 00096. Uma M. Lal-Trehan Estrada holds an IFUdG2022 grant from Universitat de Girona. Valeriia Abramova holds an FPI grant from the Ministerio de Ciencia, Innovación y Universidades with reference number PRE2021-099121. Adrià Casamitjana holds a Ramón y Cajal grant from the Spanish Government with reference number RYC2024-050753-I. This work has been supported by PID2023-146187OB-I00 from the Ministerio de Ciencia, Innovación y Universidades and also by the ICREA Academia program. Open Access funding was provided thanks to the CRUE-CSIC agreement with Elsevier