File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: The Liver Tumor Segmentation Benchmark (LiTS)

TitleThe Liver Tumor Segmentation Benchmark (LiTS)
Authors
KeywordsBenchmark
CT
Deep learning
Liver
Liver tumor
Segmentation
Issue Date1-Feb-2023
PublisherElsevier
Citation
Medical Image Analysis, 2023, v. 84 How to Cite?
Abstract

In this work, we report the set-up and results of the Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI) 2017 and the International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017 and 2018. The image dataset is diverse and contains primary and secondary tumors with varied sizes and appearances with various lesion-to-background levels (hyper-/hypo-dense), created in collaboration with seven hospitals and research institutions. Seventy-five submitted liver and liver tumor segmentation algorithms were trained on a set of 131 computed tomography (CT) volumes and were tested on 70 unseen test images acquired from different patients. We found that not a single algorithm performed best for both liver and liver tumors in the three events. The best liver segmentation algorithm achieved a Dice score of 0.963, whereas, for tumor segmentation, the best algorithms achieved Dices scores of 0.674 (ISBI 2017), 0.702 (MICCAI 2017), and 0.739 (MICCAI 2018). Retrospectively, we performed additional analysis on liver tumor detection and revealed that not all top-performing segmentation algorithms worked well for tumor detection. The best liver tumor detection method achieved a lesion-wise recall of 0.458 (ISBI 2017), 0.515 (MICCAI 2017), and 0.554 (MICCAI 2018), indicating the need for further research. LiTS remains an active benchmark and resource for research, e.g., contributing the liver-related segmentation tasks in http://medicaldecathlon.com/. In addition, both data and online evaluation are accessible via https://competitions.codalab.org/competitions/17094.


Persistent Identifierhttp://hdl.handle.net/10722/333924
ISSN
2023 Impact Factor: 10.7
2023 SCImago Journal Rankings: 4.112
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorBilic, P-
dc.contributor.authorChrist, P-
dc.contributor.authorLi, HB-
dc.contributor.authorVorontsov, E-
dc.contributor.authorBen-Cohen, A-
dc.contributor.authorKaissis, G-
dc.contributor.authorSzeskin, A-
dc.contributor.authorJacobs, C-
dc.contributor.authorMamani, GEH-
dc.contributor.authorChartrand, G-
dc.contributor.authorLohofer, F-
dc.contributor.authorHolch, JW-
dc.contributor.authorSommer, W-
dc.contributor.authorHofmann, F-
dc.contributor.authorHostettler, A-
dc.contributor.authorLev-Cohain, N-
dc.contributor.authorDrozdzal, M-
dc.contributor.authorAmitai, MM-
dc.contributor.authorVivanti, R-
dc.contributor.authorSosna, J-
dc.contributor.authorEzhov, I-
dc.contributor.authorSekuboyina, A-
dc.contributor.authorNavarro, F-
dc.contributor.authorKofler, F-
dc.contributor.authorPaetzold, JC-
dc.contributor.authorShit, S-
dc.contributor.authorHu, XB-
dc.contributor.authorLipkova, J-
dc.contributor.authorRempfler, M-
dc.contributor.authorPiraud, M-
dc.contributor.authorKirschke, J-
dc.contributor.authorWiestler, B-
dc.contributor.authorZhang, ZH-
dc.contributor.authorHulsemeyer, C-
dc.contributor.authorBeetz, M-
dc.contributor.authorEttlinger, F-
dc.contributor.authorAntonelli, M-
dc.contributor.authorBae, W-
dc.contributor.authorBellver, M-
dc.contributor.authorBi, L-
dc.contributor.authorChen, H-
dc.contributor.authorChlebus, G-
dc.contributor.authorDam, EB-
dc.contributor.authorDou, Q-
dc.contributor.authorFu, CW-
dc.contributor.authorGeorgescu, B-
dc.contributor.authorGiro-I-Nieto, X-
dc.contributor.authorGruen, F-
dc.contributor.authorHan, X-
dc.contributor.authorHeng, PA-
dc.contributor.authorHesser, J-
dc.contributor.authorMoltz, JH-
dc.contributor.authorIgel, C-
dc.contributor.authorIsensee, F-
dc.contributor.authorJager, P-
dc.contributor.authorJia, FC-
dc.contributor.authorKaluva, KC-
dc.contributor.authorKhened, M-
dc.contributor.authorKim, I-
dc.contributor.authorKim, JH-
dc.contributor.authorKim, S-
dc.contributor.authorKohl, S-
dc.contributor.authorKonopczynski, T-
dc.contributor.authorKori, A-
dc.contributor.authorKrishnamurthi, G-
dc.contributor.authorLi, F-
dc.contributor.authorLi, HC-
dc.contributor.authorLi, JB-
dc.contributor.authorLi, XM-
dc.contributor.authorLowengrub, J-
dc.contributor.authorMa, J-
dc.contributor.authorMaier-Hein, K-
dc.contributor.authorManinis, KK-
dc.contributor.authorMeine, H-
dc.contributor.authorMerhof, D-
dc.contributor.authorPai, A-
dc.contributor.authorPerslev, M-
dc.contributor.authorPetersen, J-
dc.contributor.authorPont-Tuset, J-
dc.contributor.authorQi, J-
dc.contributor.authorQi, XJ-
dc.contributor.authorRippel, O-
dc.contributor.authorRoth, K-
dc.contributor.authorSarasua, I-
dc.contributor.authorSchenk, A-
dc.contributor.authorShen, ZM-
dc.contributor.authorTorres, J-
dc.contributor.authorWachinger, C-
dc.contributor.authorWang, CL-
dc.contributor.authorWeninger, L-
dc.contributor.authorWu, JR-
dc.contributor.authorXu, DG-
dc.contributor.authorYang, XP-
dc.contributor.authorYu, SCH-
dc.contributor.authorYuan, YD-
dc.contributor.authorYue, M-
dc.contributor.authorZhang, LP-
dc.contributor.authorCardoso, J-
dc.contributor.authorBakas, S-
dc.contributor.authorBraren, R-
dc.contributor.authorHeinemann, V-
dc.contributor.authorPal, C-
dc.contributor.authorTang, A-
dc.contributor.authorKadoury, S-
dc.contributor.authorSoler, L-
dc.contributor.authorvan Ginneken, B-
dc.contributor.authorGreenspan, H-
dc.contributor.authorJoskowicz, L-
dc.contributor.authorMenze, B-
dc.date.accessioned2023-10-10T03:14:26Z-
dc.date.available2023-10-10T03:14:26Z-
dc.date.issued2023-02-01-
dc.identifier.citationMedical Image Analysis, 2023, v. 84-
dc.identifier.issn1361-8415-
dc.identifier.urihttp://hdl.handle.net/10722/333924-
dc.description.abstract<p>In this work, we report the set-up and results of the Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with the IEEE International Symposium on <a href="https://www.sciencedirect.com/topics/medicine-and-dentistry/biomedical-imaging" title="Learn more about Biomedical Imaging from ScienceDirect's AI-generated Topic Pages">Biomedical Imaging</a> (ISBI) 2017 and the International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017 and 2018. The image dataset is diverse and contains primary and secondary tumors with varied sizes and appearances with various lesion-to-background levels (hyper-/hypo-dense), created in collaboration with seven <a href="https://www.sciencedirect.com/topics/medicine-and-dentistry/hospital" title="Learn more about hospitals from ScienceDirect's AI-generated Topic Pages">hospitals</a> and research institutions. Seventy-five submitted liver and liver tumor segmentation algorithms were trained on a set of 131 computed tomography (CT) volumes and were tested on 70 unseen test images acquired from different <a href="https://www.sciencedirect.com/topics/medicine-and-dentistry/patient" title="Learn more about patients from ScienceDirect's AI-generated Topic Pages">patients</a>. We found that not a single algorithm performed best for both liver and liver tumors in the three events. The best <a href="https://www.sciencedirect.com/topics/computer-science/liver-segmentation" title="Learn more about liver segmentation from ScienceDirect's AI-generated Topic Pages">liver segmentation</a> algorithm achieved a Dice score of 0.963, whereas, for tumor segmentation, the best algorithms achieved Dices scores of 0.674 (ISBI 2017), 0.702 (MICCAI 2017), and 0.739 (MICCAI 2018). Retrospectively, we performed additional analysis on liver <a href="https://www.sciencedirect.com/topics/computer-science/tumor-detection" title="Learn more about tumor detection from ScienceDirect's AI-generated Topic Pages">tumor detection</a> and revealed that not all top-performing segmentation algorithms worked well for tumor detection. The best liver tumor detection method achieved a lesion-wise recall of 0.458 (ISBI 2017), 0.515 (MICCAI 2017), and 0.554 (MICCAI 2018), indicating the need for further research. LiTS remains an active benchmark and resource for research, e.g., contributing the liver-related <a href="https://www.sciencedirect.com/topics/computer-science/segmentation-task" title="Learn more about segmentation tasks from ScienceDirect's AI-generated Topic Pages">segmentation tasks</a> in <a href="http://medicaldecathlon.com/">http://medicaldecathlon.com/</a>. In addition, both data and online evaluation are accessible via <a href="https://competitions.codalab.org/competitions/17094">https://competitions.codalab.org/competitions/17094</a>.</p>-
dc.languageeng-
dc.publisherElsevier-
dc.relation.ispartofMedical Image Analysis-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectBenchmark-
dc.subjectCT-
dc.subjectDeep learning-
dc.subjectLiver-
dc.subjectLiver tumor-
dc.subjectSegmentation-
dc.titleThe Liver Tumor Segmentation Benchmark (LiTS)-
dc.typeArticle-
dc.identifier.doi10.1016/j.media.2022.102680-
dc.identifier.scopuseid_2-s2.0-85143743210-
dc.identifier.volume84-
dc.identifier.isiWOS:000928244200004-
dc.identifier.issnl1361-8415-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats