Deep Learning Techniques for Automatic MRI Cardiac Multi-structures Segmentation and Diagnosis: Is the Problem Solved?

Olivier Bernard 1 Alain Lalande 2 Clement Zotti 3 Frederic Cervenansky 4 Xin Yang 5 Pheng-Ann Heng 5 Irem Cetin 6 Karim Lekadir 6 Oscar Camara 6 Miguel Angel Gonzalez Ballester 7 Gerard Sanroma 6 Sandy Napel 8 Steffen Petersen 9 Georgios Tziritas 10 Elias Grinias 10 Mahendra Khened 11 Varghese Alex Kollerathu 11 Ganapathy Krishnamurthi 11 Marc-Michel Rohé 12, 13 Xavier Pennec 12, 13 Maxime Sermesant 12, 13 Fabian Isensee 14 Paul Jager 14 Klaus Maier-Hein 14 Peter Full 15 Ivo Wolf 16 Sandy Engelhardt 16 Chrisitan Baumgartner 17 Lisa Koch 18 Jelmer Wolterink 19 Ivana Isgum 19 Yeonggul Jang 20 Yoonmi Hong 20 Jay Patravali 21 Shubham Jain 21 Olivier Humbert 22 Pierre-Marc Jodoin 3
1 Images et Modèles
CREATIS - Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé
4 Service Informatique et développements
CREATIS - Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé
12 ASCLEPIOS - Analysis and Simulation of Biomedical Images
CRISAM - Inria Sophia Antipolis - Méditerranée
22 TIRO - UMR E4320 - Transporteurs en Imagerie et Radiothérapie en Oncologie
CEA - Commissariat à l'énergie atomique et aux énergies alternatives : DRF/BIAM, CNRS - Centre National de la Recherche Scientifique, TIRO-MATOs - UMR E4320 : UMR E4320
Abstract : 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 classificationtask. 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.
Document type :
Journal articles
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01803621
Contributor : Ivan Porcherot <>
Submitted on : Wednesday, May 30, 2018 - 3:52:58 PM
Last modification on : Tuesday, April 2, 2019 - 2:39:58 AM

Links full text

Identifiers

Citation

Olivier Bernard, Alain Lalande, Clement Zotti, Frederic Cervenansky, Xin Yang, et al.. Deep Learning Techniques for Automatic MRI Cardiac Multi-structures Segmentation and Diagnosis: Is the Problem Solved?. IEEE Transactions on Medical Imaging, Institute of Electrical and Electronics Engineers, 2018, 37 (11), pp.2514-2525. ⟨10.1109/TMI.2018.2837502⟩. ⟨hal-01803621⟩

Share

Metrics

Record views

448