The Medical Segmentation Decathlon

By M. Antonelli, Annika Reinke, S. Bakas, Keyvan Farahani, AnnetteKopp-Schneider, Bennett A. Landman, Geert J. S. Litjens, Bjoern H Menze, Olaf Ronneberger, Ronald M.Summers and 48 more. Nature Communications, 2021

1,513 citations recorded by the source corpus.

International challenges have become the de facto standard for comparative assessment of image analysis algorithms. Although segmentation is the most widely investigated medical image processing task, the various challenges have been organized to focus only on specific clinical tasks. We organized the Medical Segmentation Decathlon (MSD)—a biomedical image analysis challenge, in which algorithms compete in a multitude of both tasks and modalities to investigate the hypothesis that a method capable of performing well on multiple tasks will generalize well to a previously unseen task and potentially outperform a custom-designed solution. MSD results confirmed this hypothesis, moreover, MSD winner continued generalizing well to a wide range of other clinical problems for the next two years. Three main conclusions can be drawn from this study: (1) state-of-the-art image segmentation algorithms generalize well when retrained on unseen tasks; (2) consistent algorithmic performance across multiple tasks is a strong surrogate of algorithmic generalizability; (3) the training of accurate AI segmentation models is now commoditized to scientists that are not versed in AI model training. International challenges have become the de facto standard for comparative assessment of image analysis algorithms. Here, the authors present the results of a biomedical image segmentation challenge, showing that a method capable of performing well on multiple tasks will generalize well to a previously unseen task.

Subfields: Computer Vision and Pattern Recognition, Health Informatics, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/979a9f247700d00ff2c3f0612d5eb001379f93c8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Health Informatics

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Medical Image Segmentation Review: The Success of U-Net (2022, background): “where γ is set in the interval [1, 3] to enable the loss function to concentrate more on incorrectly classified predictions that are less accurate.”
Segment Anything Model for Medical Image Segmentation: Current Applications and Future Directions (2024, background): “Accurate segmentation can provide reliable volumetric and shape information of target structures, thereby assisting in many further clinical applications like disease diagnosis, quantitative analysis, and surgical planning [3, 4, 5].”
Segment anything in medical images (2023, neutral): “However, a signiﬁcant limitation of many current medical image segmentation models is their task-speciﬁc nature.”
MedMNIST v2 - A large-scale lightweight benchmark for 2D and 3D biomedical image classification (2021, background): “As reported in the winning solutions 6,7 , hyperparameter tuning, pre/post-processing, model ensemble strategies and training/test-time augmentation could be more important than the machine learning part ( e.g. , model architectures, learning scheme).”
nnFormer: Volumetric Medical Image Segmentation via a 3D Transformer (2021, positive): “For thoroughly comparing nnFormer to previous convnet-and transformer-based architecture, we conduct experiments on three datasets/tasks: the brain tumor segmentation task in Medical Segmentation Decathlon (MSD) [36], Synapse multi-organ segmentation [37] and Automatic Cardiac Diagnosis Challenge (A”
PMC-VQA: Visual Instruction Tuning for Medical Visual Question Answering (2023, neutral): “Modality Source Images QA pairs VQA-RAD [18] Radiology MedPix ® database 0.3k 3.5k PathVQA [12] Pathology PEIR Digital Library [14] 5k 32.8k SLAKE [23] Radiology MSD [3], ChestX-ray8 [36], CHAOS [15] 0.7k 14k VQA-Med-2021 [5] Radiology MedPix ® database 5k 5k PMC-VQA Mixture * PubMed Central ® 149k ”
Metrics Reloaded: Recommendations for image analysis validation (2022, background): “(2) Liver segmentation in Computed Tomography (CT) images [2, 80].”
The Liver Tumor Segmentation Benchmark (LiTS) (2019, methodology): “In order to evaluate the state-of-the-art methods for automated liver and liver tumor segmentation, we organized the Liver Tumor Segmentation Challenge (LiTS) in three events: 1) in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI) 2017, 2) with MICCAI 2017 and 3) as a d”
Swin UNETR: Swin Transformers for Semantic Segmentation of Brain Tumors in MRI Images (2022, methodology): “UNETR has shown promising results for brain tumor segmentation using the MSD dataset [1].”
AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation (2022, methodology): “Abdomen Organ Dataset #Organs #Scans #Slices # Anns per Scan Modality Region Year MSD-Liver† [1] 1 201 29,402 3410K CT France 2018 MSD-Spleen† [1] 1 61 1,563 40K CT United States 2018 MSD-Prostate† [1] 1 48 712 15K MRI Netherlands 2018 MSD-Pancreas† [1] 1 420 13,141 144K CT United States 2018 Kits† ”
MONAI: An open-source framework for deep learning in healthcare (2022, background): “Another example, the Medical Segmentation Decathlon [30], represents a set of 10 challenge datasets, widely used to demonstrate network performance across a range of segmentation tasks.”
Self-Supervised Pre-Training of Swin Transformers for 3D Medical Image Analysis (2021, background): “MSD: Medical Segmentation Decathlon (MSD) dataset [1] comprises of 10 segmentation tasks from different organs and image modalities.”
CLIP-Driven Universal Model for Organ Segmentation and Tumor Detection (2023, background): “MSD CT Tasks [1] 9 947 Spl, Liver and Tumor∗, Lung Tumor∗, Colon Tumor∗, Pan and Tumor∗, Hepatic Vessel and Tumor∗ 12.”