{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,12]],"date-time":"2026-06-12T21:05:52Z","timestamp":1781298352881,"version":"3.54.1"},"reference-count":38,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2021,2,15]],"date-time":"2021-02-15T00:00:00Z","timestamp":1613347200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,2,15]],"date-time":"2021-02-15T00:00:00Z","timestamp":1613347200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100002347","name":"Bundesministerium f\u00fcr Bildung und Forschung","doi-asserted-by":"publisher","award":["01EO1504"],"award-info":[{"award-number":["01EO1504"]}],"id":[{"id":"10.13039\/501100002347","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Med Imaging"],"published-print":{"date-parts":[[2021,12]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Image segmentation is a common task in medical imaging e.g., for volumetry analysis in cardiac MRI. Artificial neural networks are used to automate this task with performance similar to manual operators. However, this performance is only achieved in the narrow tasks networks are trained on. Performance drops dramatically when data characteristics differ from the training set properties. Moreover, neural networks are commonly considered black boxes, because it is hard to understand how they make decisions and why they fail. Therefore, it is also hard to predict whether they will generalize and work well with new data. Here we present a generic method for segmentation model interpretation. Sensitivity analysis is an approach where model input is modified in a controlled manner and the effect of these modifications on the model output is evaluated. This method yields insights into the sensitivity of the model to these alterations and therefore to the importance of certain features on segmentation performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>We present an open-source Python library (misas), that facilitates the use of sensitivity analysis with arbitrary data and models. We show that this method is a suitable approach to answer practical questions regarding use and functionality of segmentation models. We demonstrate this in two case studies on cardiac magnetic resonance imaging. The first case study explores the suitability of a published network for use on a public dataset the network has not been trained on. The second case study demonstrates how sensitivity analysis can be used to evaluate the robustness of a newly trained model.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Sensitivity analysis is a useful tool for deep learning developers as well as users such as clinicians. It extends their toolbox, enabling and improving interpretability of segmentation models. Enhancing our understanding of neural networks through sensitivity analysis also assists in decision making. Although demonstrated only on cardiac magnetic resonance images this approach and software are much more broadly applicable.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s12880-021-00551-1","type":"journal-article","created":{"date-parts":[[2021,2,17]],"date-time":"2021-02-17T09:23:12Z","timestamp":1613553792000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":38,"title":["Sensitivity analysis for interpretation of machine learning based segmentation models in cardiac MRI"],"prefix":"10.1186","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6620-807X","authenticated-orcid":false,"given":"Markus J.","family":"Ankenbrand","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liliia","family":"Shainberg","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Michael","family":"Hock","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"David","family":"Lohr","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Laura M.","family":"Schreiber","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2021,2,15]]},"reference":[{"key":"551_CR1","doi-asserted-by":"publisher","first-page":"18","DOI":"10.1016\/j.media.2016.05.004","volume":"35","author":"M Havaei","year":"2017","unstructured":"Havaei M, Davy A, Warde-Farley D, Biard A, Courville A, Bengio Y, et al. Brain tumor segmentation with deep neural networks. Med Image Anal. 2017;35:18\u201331.","journal-title":"Med Image Anal"},{"issue":"5","key":"551_CR2","doi-asserted-by":"publisher","first-page":"e190103","DOI":"10.1148\/ryai.2020190103","volume":"2","author":"RS Eijgelaar","year":"2020","unstructured":"Eijgelaar RS, Visser M, M\u00fcller DMJ, Barkhof F, Vrenken H, Herk M, et al. Robust deep learning-based segmentation of glioblastoma on routine clinical MRI scans using sparsified training. Radiol Artif Intell. 2020;2(5):e190103.","journal-title":"Radiol Artif Intell"},{"issue":"9","key":"551_CR3","doi-asserted-by":"publisher","first-page":"1342","DOI":"10.1038\/s41591-018-0107-6","volume":"24","author":"J De Fauw","year":"2018","unstructured":"De Fauw J, Ledsam JR, Romera-Paredes B, Nikolov S, Tomasev N, Blackwell S, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat Med. 2018;24(9):1342\u201350.","journal-title":"Nat Med"},{"key":"551_CR4","unstructured":"Jakhar K, Bajaj R, Gupta RJA. Pneumothorax segmentation: deep learning image segmentation to predict pneumothorax; 2019. arXiv:1912.07329."},{"key":"551_CR5","doi-asserted-by":"publisher","first-page":"25","DOI":"10.3389\/fcvm.2020.00025","volume":"7","author":"C Chen","year":"2020","unstructured":"Chen C, Qin C, Qiu H, Tarroni G, Duan J, Bai W, et al. Deep learning for cardiac image segmentation: a review. Front Cardiovasc Med. 2020;7:25.","journal-title":"Front Cardiovasc Med"},{"issue":"1","key":"551_CR6","doi-asserted-by":"publisher","first-page":"61","DOI":"10.1186\/s12968-019-0575-y","volume":"21","author":"T Leiner","year":"2019","unstructured":"Leiner T, Rueckert D, Suinesiaputra A, Bae\u00dfler B, Nezafat R, I\u0161gum I, et al. Machine learning in cardiovascular magnetic resonance: basic concepts and applications. J Cardiovasc Magn Reson. 2019;21(1):61.","journal-title":"J Cardiovasc Magn Reson"},{"issue":"8, Part 1","key":"551_CR7","doi-asserted-by":"publisher","first-page":"1549","DOI":"10.1016\/j.jcmg.2019.06.009","volume":"12","author":"G Litjens","year":"2019","unstructured":"Litjens G, Ciompi F, Wolterink JM, de Vos BD, Leiner T, Teuwen J, et al. State-of-the-art deep learning in cardiovascular image analysis. JACC Cardiovasc Imaging. 2019;12(8, Part 1):1549\u201365.","journal-title":"JACC Cardiovasc Imaging"},{"key":"551_CR8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3389\/fcvm.2019.00133","volume":"6","author":"SE Petersen","year":"2019","unstructured":"Petersen SE, Abdulkareem M, Leiner T. Artificial intelligence will transform cardiac imaging\u2014opportunities and challenges. Front Cardiovasc Med. 2019;6:1.","journal-title":"Front Cardiovasc Med"},{"issue":"2","key":"551_CR9","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1007\/s10334-018-0718-4","volume":"32","author":"S Moccia","year":"2019","unstructured":"Moccia S, Banali R, Martini C, Muscogiuri G, Pontone G, Pepi M, et al. Development and testing of a deep learning-based strategy for scar segmentation on CMR-LGE images. Magn Reson Mater Phys Biol Med. 2019;32(2):187\u201395.","journal-title":"Magn Reson Mater Phys Biol Med"},{"issue":"3","key":"551_CR10","doi-asserted-by":"publisher","first-page":"e190043","DOI":"10.1148\/ryai.2020190043","volume":"2","author":"M Reyes","year":"2020","unstructured":"Reyes M, Meier R, Pereira S, Silva CA, Dahlweid F-M, Tengg-Kobligk H, et al. On the interpretability of artificial intelligence in radiology: challenges and opportunities. Radiol Artif Intell. 2020;2(3):e190043.","journal-title":"Radiol Artif Intell"},{"issue":"2","key":"551_CR11","first-page":"345","volume":"21","author":"A Subbaswamy","year":"2020","unstructured":"Subbaswamy A, Saria S. From development to deployment: dataset shift, causality, and shift-stable models in health AI. Biostatistics. 2020;21(2):345\u201352.","journal-title":"Biostatistics"},{"key":"551_CR12","doi-asserted-by":"crossref","unstructured":"Oakden-Rayner L, Dunnmon J, Carneiro G, R\u00e9 C. Hidden stratification causes clinically meaningful failures in machine learning for medical imaging. arXiv:1909.12475 [cs, stat] (2019).","DOI":"10.1145\/3368555.3384468"},{"issue":"12","key":"551_CR13","doi-asserted-by":"publisher","first-page":"1212","DOI":"10.1001\/jama.2020.12067","volume":"324","author":"A Kaushal","year":"2020","unstructured":"Kaushal A, Altman R, Langlotz C. Geographic distribution of US cohorts used to train deep learning algorithms. JAMA. 2020;324(12):1212\u20133.","journal-title":"JAMA"},{"key":"551_CR14","doi-asserted-by":"crossref","unstructured":"Geirhos R, Jacobsen J-H, Michaelis C, Zemel R, Brendel W, Bethge M, et al. Shortcut learning in deep neural networks. 2020. arXiv:2004.07780.","DOI":"10.1038\/s42256-020-00257-z"},{"key":"551_CR15","doi-asserted-by":"publisher","first-page":"101636","DOI":"10.1016\/j.media.2020.101636","volume":"61","author":"FM Guo","year":"2020","unstructured":"Guo FM, Ng M, Goubran M, Petersen SE, Piechnik SK, Neubauer S, et al. Improving cardiac MRI convolutional neural network segmentation on small training datasets and dataset shift: a continuous kernel cut approach. Med Image Anal. 2020;61:101636.","journal-title":"Med Image Anal"},{"key":"551_CR16","doi-asserted-by":"crossref","unstructured":"Hann E, Biasiolli L, Zhang Q, Popescu IA, Werys K, Lukaschuk E, et al. Quality control-driven image segmentation towards reliable automatic image analysis in large-scale cardiovascular magnetic resonance aortic cine imaging. Medical Image Computing and Computer Assisted Intervention\u2014MICCAI 2019. Lecture Notes in Computer Science2019. p. 750\u20138.","DOI":"10.1007\/978-3-030-32245-8_83"},{"key":"551_CR17","doi-asserted-by":"crossref","unstructured":"Wang S, Tarroni G, Qin C, Mo Y, Dai C, Chen C, et al. Deep generative model-based quality control for cardiac MRI segmentation. Medical Image Computing and Computer Assisted Intervention\u2014MICCAI 2020. Lecture Notes in Computer Science2020. p. 88\u201397.","DOI":"10.1007\/978-3-030-59719-1_9"},{"key":"551_CR18","unstructured":"Vilone G, Longo L. Explainable artificial intelligence: a systematic review; 2020. arXiv:2006.00093."},{"key":"551_CR19","unstructured":"Kokhlikyan N, Miglani V, Mart\u00edn M, Wang E, Alsallakh B, Reynolds J, et al. Captum: a unified and generic model interpretability library for PyTorch; 2020. arXiv:2009.07896."},{"key":"551_CR20","unstructured":"Sundararajan M, Taly A, Yan Q, editors. Axiomatic attribution for deep networks. ICML; 2017."},{"key":"551_CR21","doi-asserted-by":"crossref","unstructured":"Selvaraju RR, Das A, Vedantam R, Cogswell M, Parikh D, Batra DJIJoCV. Grad-CAM: visual explanations from deep networks via gradient-based localization. 2019;128:336\u201359.","DOI":"10.1007\/s11263-019-01228-7"},{"key":"551_CR22","unstructured":"Dhamdhere K, Sundararajan M, Yan Q. How important is a neuron? 2019. arXiv:1805.12233."},{"key":"551_CR23","doi-asserted-by":"crossref","unstructured":"Widrow B, Hoff ME. Adaptive switching circuits. 1960 IRE WESCON Conv Rec. 1960(4):96\u2013104.","DOI":"10.21236\/AD0241531"},{"key":"551_CR24","first-page":"4943","volume":"33","author":"H Shu","year":"2019","unstructured":"Shu H, Zhu H. Sensitivity analysis of deep neural networks. Proc AAAI Conf Artif Intell. 2019;33:4943\u201350.","journal-title":"Proc AAAI Conf Artif Intell"},{"issue":"7","key":"551_CR25","doi-asserted-by":"crossref","first-page":"1064","DOI":"10.1093\/bioinformatics\/btw749","volume":"33","author":"G Teodoro","year":"2017","unstructured":"Teodoro G, Kur\u00e7 TM, Taveira LFR, Melo ACMA, Gao Y, Kong J, et al. Algorithm sensitivity analysis and parameter tuning for tissue image segmentation pipelines. Bioinformatics. 2017;33(7):1064\u201372.","journal-title":"Bioinformatics"},{"issue":"2","key":"551_CR26","doi-asserted-by":"publisher","first-page":"97","DOI":"10.1093\/comjnl\/27.2.97","volume":"27","author":"DE Knuth","year":"1984","unstructured":"Knuth DE. Literate programming. Comput J. 1984;27(2):97\u2013111.","journal-title":"Comput J"},{"issue":"2","key":"551_CR27","doi-asserted-by":"publisher","first-page":"108","DOI":"10.3390\/info11020108","volume":"11","author":"J Howard","year":"2020","unstructured":"Howard J, Gugger S. Fastai: a layered API for deep learning. Information. 2020;11(2):108.","journal-title":"Information"},{"key":"551_CR28","unstructured":"Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, et al. PyTorch: an imperative style, high-performance deep learning library. 2019:8024\u201335."},{"key":"551_CR29","doi-asserted-by":"crossref","unstructured":"P\u00e9rez-Garc\u00eda F, Sparks R, Ourselin S. TorchIO: a Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning. arXiv:2003.04696 [cs, eess, stat]. 2020.","DOI":"10.1016\/j.cmpb.2021.106236"},{"issue":"7825","key":"551_CR30","doi-asserted-by":"publisher","first-page":"357","DOI":"10.1038\/s41586-020-2649-2","volume":"585","author":"CR Harris","year":"2020","unstructured":"Harris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D, et al. Array programming with NumPy. Nature. 2020;585(7825):357\u201362.","journal-title":"Nature"},{"key":"551_CR31","unstructured":"Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, et al. TensorFlow: large-scale machine learning on heterogeneous distributed systems. 2016;abs\/1603.04467."},{"key":"551_CR32","doi-asserted-by":"crossref","unstructured":"Ankenbrand MJ, Lohr D, Schl\u00f6telburg W, Reiter T, Wech T, Schreiber LM. A Deep Learning Based Cardiac Cine Segmentation Framework for Clinicians - Transfer Learning Application to 7T. medRxiv. 2020:2020.06.15.20131656.","DOI":"10.1101\/2020.06.15.20131656"},{"key":"551_CR33","unstructured":"Booz Allen Hamilton. Data Science Bowl Cardiac Challenge Data. https:\/\/www.kaggle.com\/c\/second-annual-data-science-bowl: kaggle.com; 2016."},{"key":"551_CR34","doi-asserted-by":"crossref","unstructured":"Bai W, Sinclair M, Tarroni G, Oktay O, Rajchl M, Vaillant G, et al. Automated cardiovascular magnetic resonance image analysis with fully convolutional networks. Journal of Cardiovascular Magnetic Resonance. 2018;20.","DOI":"10.1186\/s12968-018-0471-x"},{"issue":"1","key":"551_CR35","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1186\/s12968-016-0227-4","volume":"18","author":"SE Petersen","year":"2016","unstructured":"Petersen SE, Matthews PM, Francis JM, Robson MD, Zemrak F, Boubertakh R, et al. UK Biobank\u2019s cardiovascular magnetic resonance protocol. J Cardiovasc Magn Reson. 2016;18(1):8.","journal-title":"J Cardiovasc Magn Reson"},{"issue":"1","key":"551_CR36","doi-asserted-by":"publisher","first-page":"182","DOI":"10.1002\/mrm.28423","volume":"85","author":"M Hock","year":"2020","unstructured":"Hock M, Terekhov M, Stefanescu MR, Lohr D, Herz S, Reiter T, et al. B0 shimming of the human heart at 7T. Magn Reson Med. 2020;85(1):182\u201396.","journal-title":"Magn Reson Med"},{"key":"551_CR37","doi-asserted-by":"crossref","unstructured":"Ankenbrand MJ, Lohr D, Schreiber LM. Exploring ensemble applications for multi-sequence myocardial pathology segmentation. Myocardial Pathology Segmentation Combining Multi-Sequence Cardiac Magnetic Resonance Images. Lecture Notes in Computer Science; 2020. p. 60\u20137.","DOI":"10.1007\/978-3-030-65651-5_6"},{"key":"551_CR38","doi-asserted-by":"publisher","first-page":"88","DOI":"10.1016\/j.patrec.2008.04.005","volume":"30","author":"G Brostow","year":"2009","unstructured":"Brostow G, Fauqueur J, Cipolla RJPRL. Semantic object classes in video: a high-definition ground truth database. Pattern Recognit Lett. 2009;30:88\u201397.","journal-title":"Pattern Recognit Lett"}],"container-title":["BMC Medical Imaging"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12880-021-00551-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12880-021-00551-1\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12880-021-00551-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,21]],"date-time":"2023-10-21T05:58:17Z","timestamp":1697867897000},"score":1,"resource":{"primary":{"URL":"https:\/\/bmcmedimaging.biomedcentral.com\/articles\/10.1186\/s12880-021-00551-1"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,2,15]]},"references-count":38,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2021,12]]}},"alternative-id":["551"],"URL":"https:\/\/doi.org\/10.1186\/s12880-021-00551-1","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-97535\/v1","asserted-by":"object"}]},"ISSN":["1471-2342"],"issn-type":[{"value":"1471-2342","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,2,15]]},"assertion":[{"value":"20 October 2020","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 January 2021","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"15 February 2021","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"13 September 2021","order":4,"name":"change_date","label":"Change Date","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Update","order":5,"name":"change_type","label":"Change Type","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The original article was updated to correct a typographical error in references 12 and 29","order":6,"name":"change_details","label":"Change Details","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Ethics approval of the local ethics committee at the University Hospital W\u00fcrzburg has been granted under reference number 7\/17-sc. Written, informed consent was obtained from all human participants.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"All human volunteers gave their consent for publication using our institutional consent form.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The authors declare that they have no competing interests.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"27"}}