{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T20:39:13Z","timestamp":1792787953547,"version":"4.5.0"},"reference-count":35,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T00:00:00Z","timestamp":1789516800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T00:00:00Z","timestamp":1790208000000},"content-version":"vor","delay-in-days":8,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Julius-Maximilians-Universit\u00e4t W\u00fcrzburg","id":[{"id":"https:\/\/ror.org\/00fbnyb24","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Med Imaging"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Ejection fraction (EF) is a central measure of cardiac function, but echocardiographic EF assessment remains reader-dependent and sensitive to acquisition quality. Deep learning can automate EF estimation, yet performance measured on a single development dataset may not transfer to data acquired under different imaging and annotation conventions. We investigated whether anatomically constrained weakly supervised learning improves cross-dataset EF estimation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We propose CAFEx, a contrastive-augmented feature extraction pipeline that combines left ventricular segmentation, echocardiography-specific augmentation, mask-derived anatomical features, and temporal EF regression. The model was trained on EchoNet-Dynamic using video-level EF labels with limited dense annotations for the anatomical component. External evaluation was performed on CAMUS after image harmonization and recalculation of an apical-four-chamber monoplane EF endpoint. Performance was compared with reproduced segmentation-based, direct video-regression, graph-based, transformer-based, and feature-extraction baselines using Dice score, mean absolute error, and coefficient of determination.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      In the harmonized train-on-EchoNet\/test-on-CAMUS benchmark, anatomically constrained models degraded less than direct video-regression baselines. CAFEx achieved 90.73% Dice on CAMUS and improved CAMUS EF prediction over the strongest reproduced feature-extraction baseline, increasing\n                      <jats:inline-formula>\n                        <jats:tex-math>$$R^2$$<\/jats:tex-math>\n                      <\/jats:inline-formula>\n                      from 0.38 to 0.54 and reducing mean absolute EF error from 7.89 to 6.70 percentage points. Poor-quality videos and extreme EF values remained challenging.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Weakly supervised EF estimation can benefit from an anatomical bottleneck learned with limited dense annotation. The findings support controlled external validation under explicit image and target harmonization, while further prospective validation, calibration, and quality-control mechanisms are needed before clinical deployment.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s12880-026-02781-7","type":"journal-article","created":{"date-parts":[[2026,9,16]],"date-time":"2026-09-16T18:14:25Z","timestamp":1789582465000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Weakly supervised anatomical feature learning for cross-dataset ejection fraction estimation from echocardiography videos"],"prefix":"10.1186","volume":"26","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1593-3300","authenticated-orcid":false,"given":"Adrian","family":"Krenzer","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Viktoria","family":"Wieser","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tobias","family":"Friedetzki","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,9,16]]},"reference":[{"issue":"19","key":"2781_CR1","doi-asserted-by":"publisher","first-page":"2360","DOI":"10.1016\/j.jacc.2018.08.2162","volume":"72","author":"TH Marwick","year":"2018","unstructured":"Marwick TH. Ejection fraction pros and cons. J Am Coll Cardiol. 2018;72(19):2360\u201379. https:\/\/doi.org\/10.1016\/j.jacc.2018.08.2162.","journal-title":"J Am Coll Cardiol"},{"issue":"5","key":"2781_CR2","doi-asserted-by":"publisher","first-page":"1047","DOI":"10.1007\/s10554-021-02482-y","volume":"38","author":"S Kim","year":"2022","unstructured":"Kim S, Park H-B, Jeon J, Arsanjani R, Heo R, Lee S-E, et al. Fully automated quantification of cardiac chamber and function assessment in 2-d echocardiography: clinical feasibility of deep learning-based algorithms. Int J Cardiovasc Imag. 2022;38(5):1047\u201359. https:\/\/doi.org\/10.1007\/s10554-021-02482-y.","journal-title":"Int J Cardiovasc Imag"},{"issue":"1","key":"2781_CR3","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.jjcc.2013.02.017","volume":"62","author":"H Iwano","year":"2013","unstructured":"Iwano H, Little WC. Heart failure: what does ejection fraction have to do with it? J Cardiol. 2013;62(1):1\u20133. https:\/\/doi.org\/10.1016\/j.jjcc.2013.02.017.","journal-title":"J Cardiol"},{"issue":"4","key":"2781_CR4","doi-asserted-by":"publisher","first-page":"736","DOI":"10.1016\/S0735-1097(03)00789-7","volume":"42","author":"JP Curtis","year":"2003","unstructured":"Curtis JP, Sokol SI, Wang Y, Rathore SS, Ko DT, Jadbabaie F, et al. The association of left ventricular ejection fraction, mortality, and cause of death in stable outpatients with heart failure. J Am Coll Cardiol. 2003;42(4):736\u201342. https:\/\/doi.org\/10.1016\/S0735-1097(03)00789-7.","journal-title":"J Am Coll Cardiol"},{"issue":"3","key":"2781_CR5","doi-asserted-by":"publisher","first-page":"233","DOI":"10.1093\/ehjci\/jev014","volume":"16","author":"RM Lang","year":"2015","unstructured":"Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American society of echocardiography and the European association of cardiovascular imaging. Eur Heart J - Cardiovasc Imag. 2015;16(3):233\u201371. https:\/\/doi.org\/10.1093\/ehjci\/jev014.","journal-title":"Eur Heart J - Cardiovasc Imag"},{"key":"2781_CR6","doi-asserted-by":"publisher","unstructured":"Smistad E, Ostvik A, Salte IM, Leclerc S, Bernard O, Lovstakken L. Fully automatic real-time ejection fraction and mapse measurements in 2d echocardiography using deep neural networks. 2018 IEEE International Ultrasonics Symposium (IUS). 2018, pp. 1\u20134. doi: https:\/\/doi.org\/10.1109\/ULTSYM.2018.8579886.","DOI":"10.1109\/ULTSYM.2018.8579886"},{"issue":"7802","key":"2781_CR7","doi-asserted-by":"publisher","first-page":"252","DOI":"10.1038\/s41586-020-2145-8","volume":"580","author":"D Ouyang","year":"2020","unstructured":"Ouyang D, He B, Ghorbani A, Yuan N, Ebinger J, Langlotz C, et al. Video-based ai for beat-to-beat assessment of cardiac function. Nature. 2020;580(7802):252\u201356. https:\/\/doi.org\/10.1038\/s41586-020-2145-8.","journal-title":"Nature"},{"key":"2781_CR8","first-page":"234","volume-title":"Medical image computing and computer-assisted intervention \u2013 MICCAI 2015","author":"O Ronneberger","year":"2015","unstructured":"Ronneberger O, Fischer P, Brox T. U-net: convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells WM, Frangi AF, editors. Medical image computing and computer-assisted intervention \u2013 MICCAI 2015. Cham: Springer; 2015. p. 234\u201341."},{"key":"2781_CR9","doi-asserted-by":"crossref","unstructured":"Isensee F, Maier-Hein KH. An attempt at beating the 3D U-Net. 2019.","DOI":"10.24926\/548719.001"},{"key":"2781_CR10","doi-asserted-by":"crossref","unstructured":"Chen L-C, Papandreou G, Schroff F, Adam H. Rethinking atrous convolution for semantic image segmentation. 2017.","DOI":"10.1007\/978-3-030-01234-2_49"},{"key":"2781_CR11","doi-asserted-by":"publisher","first-page":"370","DOI":"10.1007\/978-3-031-16440-8_36","volume-title":"Medical image computing and computer assisted intervention \u2013 MICCAI 2022","author":"R Muhtaseb","year":"2022","unstructured":"Muhtaseb R, Yaqub M. Echocotr: estimation of the left ventricular ejection fraction from spatiotemporal echocardiography. In: Wang L, Dou Q, Fletcher PT, Speidel S, Li S, editors. Medical image computing and computer assisted intervention \u2013 MICCAI 2022. Cham: Springer; 2022. p. 370\u201379."},{"issue":"14","key":"2781_CR12","doi-asserted-by":"publisher","first-page":"2365","DOI":"10.3390\/diagnostics13142365","volume":"13","author":"M Liao","year":"2023","unstructured":"Liao M, Lian Y, Yao Y, Chen L, Gao F, Xu L, et al. Left ventricle segmentation in echocardiography with transformer. Diagn (Basel). 2023;13(14):2365. https:\/\/doi.org\/10.3390\/diagnostics13142365.","journal-title":"Diagn (Basel)"},{"key":"2781_CR13","unstructured":"Li K, Wang Y, Gao P, Song G, Liu Y, Li H, et al. UniFormer: unified transformer for efficient spatiotemporal representation learning. 2022."},{"key":"2781_CR14","doi-asserted-by":"publisher","first-page":"360","DOI":"10.1007\/978-3-031-16440-8_35","volume-title":"Medical image computing and computer assisted intervention \u2013 MICCAI 2022","author":"M Mokhtari","year":"2022","unstructured":"Mokhtari M, Tsang T, Abolmaesumi P, Liao R. Echognn: explainable ejection fraction estimation with graph neural networks. In: Wang L, Dou Q, Fletcher PT, Speidel S, Li S, editors. Medical image computing and computer assisted intervention \u2013 MICCAI 2022. Cham: Springer; 2022. p. 360\u201369."},{"issue":"13","key":"2781_CR15","doi-asserted-by":"publisher","first-page":"2155","DOI":"10.3390\/diagnostics13132155","volume":"13","author":"S Batool","year":"2023","unstructured":"Batool S, Taj IA, Ghafoor M. Ejection fraction estimation from echocardiograms using optimal left ventricle feature extraction based on clinical methods. Diagnostics. 2023;13(13):2155. https:\/\/doi.org\/10.3390\/diagnostics13132155.","journal-title":"Diagnostics"},{"key":"2781_CR16","doi-asserted-by":"publisher","first-page":"268","DOI":"10.1007\/978-3-030-58607-2_16","volume-title":"Computer vision \u2013 ECCV 2020","author":"W Van Gansbeke","year":"2020","unstructured":"Van Gansbeke W, Vandenhende S, Georgoulis S, Proesmans M, Van Gool L. Scan: learning to classify images without labels. In: Vedaldi A, Bischof H, Brox T, Frahm J-M, editors. Computer vision \u2013 ECCV 2020. Cham: Springer; 2020. p. 268\u201385."},{"issue":"2","key":"2781_CR17","doi-asserted-by":"publisher","first-page":"373","DOI":"10.1007\/s10994-019-05855-6","volume":"109","author":"JE Engelen","year":"2020","unstructured":"Engelen JE, Hoos HH. A survey on semi-supervised learning. Mach Learn. 2020;109(2):373\u2013440. https:\/\/doi.org\/10.1007\/s10994-019-05855-6.","journal-title":"Mach Learn"},{"key":"2781_CR18","doi-asserted-by":"publisher","first-page":"680","DOI":"10.1007\/978-3-031-12053-4_50","volume-title":"Medical image understanding and analysis","author":"M Saeed","year":"2022","unstructured":"Saeed M, Muhtaseb R, Yaqub M. Contrastive pretraining for echocardiography segmentation with limited data. In: Yang G, Aviles-Rivero A, Roberts M, Sch\u00f6nlieb C-B, editors. Medical image understanding and analysis. Springer; 2022. p. 680\u201391."},{"key":"2781_CR19","unstructured":"Maani F, Ukaye A, Saadi N, Saeed N, Yaqub M. UniLVSeg: unified left ventricular segmentation with sparsely annotated echocardiogram videos through self-supervised temporal masking and weakly supervised training. 2023."},{"key":"2781_CR20","unstructured":"Stanford AIMI. EchoNet-dynamic dataset. 2024. https:\/\/stanfordaimi.azurewebsites.net\/datasets\/834e1cd1-92f7-4268-9daa-d359198b310a."},{"issue":"9","key":"2781_CR21","doi-asserted-by":"publisher","first-page":"2198","DOI":"10.1109\/TMI.2019.2900516","volume":"38","author":"S Leclerc","year":"2019","unstructured":"Leclerc S, Smistad E, Pedrosa J, Stvik A, Cervenansky F, Espinosa F, et al. Deep learning for segmentation using an open large-scale dataset in 2d echocardiography. IEEE Trans Med Imag. 2019;38(9):2198\u2013210. https:\/\/doi.org\/10.1109\/TMI.2019.2900516.","journal-title":"IEEE Trans Med Imag"},{"key":"2781_CR22","unstructured":"CREATIS: Camus Dataset. 2024. https:\/\/www.creatis.insa-lyon.fr\/Challenge\/camus\/databases.html."},{"key":"2781_CR23","doi-asserted-by":"publisher","unstructured":"Zyuzin V, Mukhtarov A, Neustroev D, Chumarnaya T. Segmentation of 2d echocardiography images using residual blocks in u-net architectures. 2020 Ural Symposium on Biomedical Engineering, Radioelectronics and Information Technology (USBEREIT). 2020, pp. 499\u2013502. doi: https:\/\/doi.org\/10.1109\/USBEREIT48449.2020.9117678.","DOI":"10.1109\/USBEREIT48449.2020.9117678"},{"key":"2781_CR24","doi-asserted-by":"publisher","first-page":"58","DOI":"10.1016\/j.ejmp.2019.10.001","volume":"67","author":"S Moradi","year":"2019","unstructured":"Moradi S, Oghli MG, Alizadehasl A, Shiri I, Oveisi N, Oveisi M, et al. Mfp-unet: a novel deep learning based approach for left ventricle segmentation in echocardiography. Phys Med. 2019;67:58\u201369. https:\/\/doi.org\/10.1016\/j.ejmp.2019.10.001.","journal-title":"Phys Med"},{"key":"2781_CR25","doi-asserted-by":"publisher","first-page":"159755","DOI":"10.1109\/ACCESS.2021.3122256","volume":"9","author":"A Amer","year":"2021","unstructured":"Amer A, Ye X, Janan F. Resdunet: a deep learning-based left ventricle segmentation method for echocardiography. IEEE Access. 2021;9:159755\u201363. https:\/\/doi.org\/10.1109\/ACCESS.2021.3122256.","journal-title":"IEEE Access"},{"key":"2781_CR26","unstructured":"Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, et al. Attention is all you need. In: Guyon I, Luxburg UV, Bengio S, Wallach H, Fergus R, Vishwanathan S, Garnett R, editors. Advances in neural information processing systems. Vol. 30. 2017."},{"issue":"1","key":"2781_CR27","doi-asserted-by":"publisher","first-page":"2104","DOI":"10.1038\/s41598-024-52480-y","volume":"14","author":"J Alven","year":"2024","unstructured":"Alven J, Hagberg E, Hagerman D, Petersen R, Hjelmgren O. A deep multi-stream model for robust prediction of left ventricular ejection fraction in 2d echocardiography. Sci Rep. 2024;14(1):2104. https:\/\/doi.org\/10.1038\/s41598-024-52480-y.","journal-title":"Sci Rep"},{"issue":"1","key":"2781_CR28","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1093\/nsr\/nwx106","volume":"5","author":"Z-H Zhou","year":"2018","unstructured":"Zhou Z-H. A brief introduction to weakly supervised learning. Natl Sci Rev. 2018;5(1):44\u201353. https:\/\/doi.org\/10.1093\/nsr\/nwx106.","journal-title":"Natl Sci Rev"},{"issue":"4","key":"2781_CR29","doi-asserted-by":"publisher","first-page":"6802","DOI":"10.1109\/TNNLS.2024.3399164","volume":"36","author":"X Zhang","year":"2025","unstructured":"Zhang X, Xiao Z, Wu X, Chen Y, Zhao J, Hu Y, et al. Pyramid pixel context adaption network for medical image classification with supervised contrastive learning. IEEE Trans Neural Netw Learn Syst. 2025;36(4):6802\u201315. https:\/\/doi.org\/10.1109\/TNNLS.2024.3399164.","journal-title":"IEEE Trans Neural Netw Learn Syst"},{"key":"2781_CR30","doi-asserted-by":"publisher","unstructured":"Li H, Li H, Zhao W, Fu H, Su X, Hu Y, et al. Frequency-mixed single-source domain generalization for medical image segmentation. Med Image Comput Comput Assisted Intervention \u2013 MICCAI 2023. 2023;127\u201336. https:\/\/doi.org\/10.1007\/978-3-031-43987-2_13.","DOI":"10.1007\/978-3-031-43987-2_13"},{"key":"2781_CR31","unstructured":"Ouyang D. EchoNet-dynamic GitHub repository. 2024. https:\/\/github.com\/echonet\/dynamic."},{"issue":"14","key":"2781_CR32","doi-asserted-by":"publisher","first-page":"2365","DOI":"10.3390\/diagnostics13142365","volume":"13","author":"M Liao","year":"2023","unstructured":"Liao M, Lian Y, Yao Y, Chen L, Gao F, Xu L, et al. Left ventricle segmentation in echocardiography with transformer. Diagnostics. 2023;13(14):2365. https:\/\/doi.org\/10.3390\/diagnostics13142365.","journal-title":"Diagnostics"},{"key":"2781_CR33","doi-asserted-by":"publisher","first-page":"104557","DOI":"10.1016\/j.bspc.2022.104557","volume":"82","author":"C Sfakianakis","year":"2023","unstructured":"Sfakianakis C, Simantiris G, Tziritas G. Gudu: geometrically-constrained ultrasound data augmentation in u-net for echocardiography semantic segmentation. Biomed Signal Process Control. 2023;82:104557. https:\/\/doi.org\/10.1016\/j.bspc.2022.104557.","journal-title":"Biomed Signal Process Control"},{"issue":"14","key":"2781_CR34","doi-asserted-by":"publisher","first-page":"5481","DOI":"10.5194\/gmd-15-5481-2022","volume":"15","author":"TO Hodson","year":"2022","unstructured":"Hodson TO. Root-mean-square error (rmse) or mean absolute error (mae): when to use them or not. Geosci Model Dev. 2022;15(14):5481\u201387. https:\/\/doi.org\/10.5194\/gmd-15-5481-2022.","journal-title":"Geosci Model Dev"},{"key":"2781_CR35","unstructured":"Scikit-Learn Developers: Scikit-Learn R2-score definition. 2024. https:\/\/scikit-learn.org\/stable\/modules\/model_evaluation.html#r2-score."}],"container-title":["BMC Medical Imaging"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12880-026-02781-7","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12880-026-02781-7.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12880-026-02781-7.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T11:03:13Z","timestamp":1790247793000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1186\/s12880-026-02781-7"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,16]]},"references-count":35,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["2781"],"URL":"https:\/\/doi.org\/10.1186\/s12880-026-02781-7","relation":{},"ISSN":["1471-2342"],"issn-type":[{"value":"1471-2342","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,16]]},"assertion":[{"value":"3 May 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"7 September 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 September 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"This study is a secondary analysis of publicly available, de-identified research datasets. No new human participants, human tissue, or identifiable participant data were collected. Ethics approval and consent to participate were not required for this analysis.","order":1,"name":"Ethics","label":"Ethics approval and consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable. The manuscript does not contain identifiable individual person data, images, or videos.","order":2,"name":"Ethics","label":"Consent for publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":3,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"470"}}