{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T17:19:55Z","timestamp":1787764795793,"version":"build-2784847793"},"reference-count":29,"publisher":"Springer Science and Business Media LLC","license":[{"start":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T00:00:00Z","timestamp":1787702400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T00:00:00Z","timestamp":1787702400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/100000052","name":"NIH Office of the Director","doi-asserted-by":"publisher","award":["1OT2OD038065-01"],"award-info":[{"award-number":["1OT2OD038065-01"]}],"id":[{"id":"10.13039\/100000052","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Mike Slive Foundation for Prostate Cancer Research Pilot Grant"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["J Digit Imaging. Inform. med."],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Deep learning (DL) has shown promise in segmenting clinically significant prostate cancer (csPCa) on MRI. However, batch effects arising from multi-site variations impact their generalizability. This study investigates the impact of training data harmonization, diversity, and few-shot fine-tuning on DL in segmenting csPCa on multi-site MRI. 3\u00a0T prostate MRI from\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u20091822 patients of four sites (public: D\u2081,\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u20091500; D\u2082,\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u2009157; institutional: D\u2083,\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u200947; D\u2084,\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u2009118) was leveraged. Bi-parametric MRI (T2-weighted, ADC) was harmonized with csPCa lesions delineated by expert radiologists. nnU-Net DL models for csPCa segmentation on MRI were trained separately (C\u2081 on D\u2081; C\u2082 on D\u2082) and jointly (C\u2083 on D\u2081-D\u2082). The jointly trained C\u2083 model was then fine-tuned on D\u2083 and D\u2084 using few-shot learning in increments of 50%, 75%, and 100% of train data. Models were primarily evaluated on holdout test sets using sensitivity primarily, in addition to AUC, DSC, precision, and Hausdorff distance (paired\n                    <jats:italic>t<\/jats:italic>\n                    -tests for evaluating significance). Batch effects persisted across datasets despite harmonization. For csPCa segmentation on D\n                    <jats:sub>1<\/jats:sub>\n                    \/D\n                    <jats:sub>2<\/jats:sub>\n                    test sets,\n                    <jats:italic>C<\/jats:italic>\n                    <jats:sub>1<\/jats:sub>\n                    achieved sensitivities of 0.51\u2009\u00b1\u20090.34\/0.11\u2009\u00b1\u20090.17,\n                    <jats:italic>C<\/jats:italic>\n                    <jats:sub>2<\/jats:sub>\n                    achieved 0.23\u2009\u00b1\u20090.31\/0.17\u2009\u00b1\u20090.22, while\n                    <jats:italic>C<\/jats:italic>\n                    <jats:sub>3<\/jats:sub>\n                    improved to 0.48\u2009\u00b1\u20090.36\/0.33\u2009\u00b1\u20090.26. On D\n                    <jats:sub>3<\/jats:sub>\n                    \/D\n                    <jats:sub>4<\/jats:sub>\n                    test sets,\n                    <jats:italic>C<\/jats:italic>\n                    <jats:sub>\n                      <jats:italic>3<\/jats:italic>\n                    <\/jats:sub>\n                    achieved zero-shot sensitivities of 0.19\u2009\u00b1\u20090.22\/0.37\u2009\u00b1\u20090.35. Few-shot fine-tuning on D\n                    <jats:sub>3<\/jats:sub>\n                    and D\n                    <jats:sub>4<\/jats:sub>\n                    improved sensitivity to 0.32\u2009\u00b1\u20090.26\/0.54\u2009\u00b1\u20090.30 on D\n                    <jats:sub>3<\/jats:sub>\n                    \/D\n                    <jats:sub>4<\/jats:sub>\n                    test sets after using 100% of available fine-tuning samples. On subset analyses, lesions\u2009&gt;\u20090.5 cm\n                    <jats:sup>3<\/jats:sup>\n                    had consistently higher segmentation performance compared to smaller lesions\u2009&lt;\u20090.5 cm\n                    <jats:sup>3<\/jats:sup>\n                    . Diverse multi-site training improves robustness of DL models for csPCa segmentation. Pre-trained models perform modestly on institutional MRI datasets under zero-shot inference, but few-shot fine-tuning on target sites enhances performance.\n                  <\/jats:p>","DOI":"10.1007\/s10278-026-02191-9","type":"journal-article","created":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T16:38:44Z","timestamp":1787762324000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Enhancing Robustness of Deep Learning to Batch Effects from Multi-site Data for Segmentation of Clinically Significant Prostate Cancer on MRI"],"prefix":"10.1007","author":[{"given":"Sumedh","family":"Sonawane","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pranav","family":"Sompalle","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dihitha","family":"Baddula","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Minoti","family":"Kedare","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Natasha","family":"Setiadi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Abhishek","family":"Midya","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sena","family":"Azamat","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zelin","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dattatraya","family":"Patil","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Allison","family":"Yang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elias","family":"Kikano","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Danielle","family":"Hill","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas","family":"Massuti","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Martin G.","family":"Sanda","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anant","family":"Madabhushi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Clinton D.","family":"Bahler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2852-3869","authenticated-orcid":false,"given":"Rakesh","family":"Shiradkar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,8,26]]},"reference":[{"key":"2191_CR1","doi-asserted-by":"publisher","unstructured":"Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA A Cancer J Clinicians. 2023 Jan;73(1):17\u201348. https:\/\/doi.org\/10.3322\/caac.21763","DOI":"10.3322\/caac.21763"},{"key":"2191_CR2","doi-asserted-by":"publisher","unstructured":"Ahmed HU, El-Shater Bosaily A, Brown LC, Gabe R, Kaplan R, Parmar MK, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study. The Lancet. 2017 Feb;389(10071):815\u201322. https:\/\/doi.org\/10.1016\/S0140-6736(16)32401-1","DOI":"10.1016\/S0140-6736(16)32401-1"},{"key":"2191_CR3","doi-asserted-by":"publisher","unstructured":"Ren H, Peng Y, Si Y, Ye Y, Gong L. Accuracy, intra-, and inter-radiologist variability of PI-RADS v2.1 scoring for clinically significant prostate cancer detection. Quant Imaging Med Surg. 2025 Aug;15(8):7080\u20139. https:\/\/doi.org\/10.21037\/qims-2025-37","DOI":"10.21037\/qims-2025-37"},{"key":"2191_CR4","doi-asserted-by":"publisher","unstructured":"Chatterjee A, Gallan A, Fan X, Medved M, Akurati P, Bourne RM, et al. Prostate Cancers Invisible on Multiparametric MRI: Pathologic Features in Correlation with Whole-Mount Prostatectomy. Cancers. 2023 Dec 13;15(24):5825. https:\/\/doi.org\/10.3390\/cancers15245825","DOI":"10.3390\/cancers15245825"},{"key":"2191_CR5","doi-asserted-by":"publisher","unstructured":"\u017burowska A, P\u0119ksa R, Bie\u0144kowski M, Skrobisz K, Sowa M, Matuszewski M, et al. Prostate Cancer and Its Mimics\u2014A Pictorial Review. Cancers. 2023 Jul 19;15(14):3682. https:\/\/doi.org\/10.3390\/cancers15143682","DOI":"10.3390\/cancers15143682"},{"key":"2191_CR6","doi-asserted-by":"publisher","unstructured":"Esteva A, Feng J, Van Der Wal D, Huang SC, Simko JP, DeVries S, et al. Prostate cancer therapy personalization via multi-modal deep learning on randomized phase III clinical trials. npj Digit Med. 2022 Jun 8;5(1):71. https:\/\/doi.org\/10.1038\/s41746-022-00613-w","DOI":"10.1038\/s41746-022-00613-w"},{"key":"2191_CR7","doi-asserted-by":"publisher","unstructured":"Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, et al. A survey on deep learning in medical image analysis. Medical Image Analysis. 2017 Dec;42:60\u201388. https:\/\/doi.org\/10.1016\/j.media.2017.07.005","DOI":"10.1016\/j.media.2017.07.005"},{"key":"2191_CR8","doi-asserted-by":"publisher","unstructured":"Zhu G, Luo J, Ouyang Z, Cheng Z, Deng Y, Guan Y, et al. The Assessment of Prostate Cancer Aggressiveness Using a Combination of Quantitative Diffusion-Weighted Imaging and Dynamic Contrast-Enhanced Magnetic Resonance Imaging. CMAR. 2021 Jul;Volume 13:5287\u201395. https:\/\/doi.org\/10.2147\/CMAR.S319306","DOI":"10.2147\/CMAR.S319306"},{"key":"2191_CR9","doi-asserted-by":"publisher","unstructured":"Glocker B, Robinson R, Castro DC, Dou Q, Konukoglu E. Machine Learning with Multi-Site Imaging Data: An Empirical Study on the Impact of Scanner Effects [Internet]. arXiv; 2019 [cited 2025 Oct 10]. Available from: http:\/\/arxiv.org\/abs\/1910.04597https:\/\/doi.org\/10.48550\/arXiv.1910.04597","DOI":"10.48550\/arXiv.1910.04597"},{"key":"2191_CR10","doi-asserted-by":"publisher","unstructured":"Feng F, Miladinovic B, Zhang K, Dignam JJ, Wang D, Yu M, et al. Early Endpoints in High-risk Localized Prostate Cancer: Exploratory Analysis of Three Radiation Therapy Oncology Group Phase 3 Studies. European Urology. 2023 Sep;84(3):331\u201340. https:\/\/doi.org\/10.1016\/j.eururo.2023.05.031","DOI":"10.1016\/j.eururo.2023.05.031"},{"key":"2191_CR11","doi-asserted-by":"publisher","unstructured":"Onofrey JA, Casetti-Dinescu DI, Lauritzen AD, Sarkar S, Venkataraman R, Fan RE, et al. Generalizable Multi-Site Training and Testing Of Deep Neural Networks Using Image Normalization. In: 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019) [Internet]. Venice, Italy: IEEE; 2019 [cited 2025 Oct 13]. p. 348\u201351. Available from: https:\/\/ieeexplore.ieee.org\/document\/8759295\/https:\/\/doi.org\/10.1109\/ISBI.2019.8759295","DOI":"10.1109\/ISBI.2019.8759295"},{"key":"2191_CR12","doi-asserted-by":"publisher","unstructured":"Saha A, Bosma JS, Twilt JJ, Van Ginneken B, Bjartell A, Padhani AR, et al. Artificial intelligence and radiologists in prostate cancer detection on MRI (PI-CAI): an international, paired, non-inferiority, confirmatory study. The Lancet Oncology. 2024 Jul;25(7):879\u201387. https:\/\/doi.org\/10.1016\/S1470-2045(24)00220-1","DOI":"10.1016\/S1470-2045(24)00220-"},{"key":"2191_CR13","doi-asserted-by":"publisher","unstructured":"Adams LC, Makowski MR, Engel G, Rattunde M, Busch F, Asbach P, et al. Prostate158 - An expert-annotated 3T MRI dataset and algorithm for prostate cancer detection. Computers in Biology and Medicine. 2022 Sep;148:105817. https:\/\/doi.org\/10.1016\/j.compbiomed.2022.105817","DOI":"10.1016\/j.compbiomed.2022.10581"},{"key":"2191_CR14","doi-asserted-by":"publisher","unstructured":"Ng ABCD, Asif A, Agarwal R, Panebianco V, Girometti R, Ghai S, et al. Biparametric vs Multiparametric MRI for Prostate Cancer Diagnosis: The PRIME Diagnostic Clinical Trial. JAMA. 2025 Oct 7;334(13):1170. https:\/\/doi.org\/10.1001\/jama.2025.13722","DOI":"10.1001\/jama.2025.13722"},{"key":"2191_CR15","doi-asserted-by":"publisher","unstructured":"Klein S, Staring M, Murphy K, Viergever MA, Pluim J. elastix: A Toolbox for Intensity-Based Medical Image Registration. IEEE Trans Med Imaging. 2010 Jan;29(1):196\u2013205. https:\/\/doi.org\/10.1109\/TMI.2009.2035616","DOI":"10.1109\/TMI.2009.2035616"},{"key":"2191_CR16","doi-asserted-by":"publisher","unstructured":"Zhu JY, Park T, Isola P, Efros AA. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks [Internet]. arXiv; 2020 [cited 2026 Jul 1]. Available from: http:\/\/arxiv.org\/abs\/1703.10593https:\/\/doi.org\/10.48550\/arXiv.1703.10593","DOI":"10.48550\/arXiv.1703.10593"},{"key":"2191_CR17","doi-asserted-by":"publisher","unstructured":"18.Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods. 2021 Feb;18(2):203\u201311. https:\/\/doi.org\/10.1038\/s41592-020-01008-z","DOI":"10.1038\/s41592-020-01008-z"},{"key":"2191_CR18","doi-asserted-by":"publisher","unstructured":"Nyul LG, Udupa JK, Xuan Zhang. New variants of a method of MRI scale standardization. IEEE Trans Med Imaging. 2000 Feb;19(2):143\u201350. https:\/\/doi.org\/10.1109\/42.836373","DOI":"10.1109\/42.836373"},{"key":"2191_CR19","doi-asserted-by":"publisher","unstructured":"21.Fortin JP, Parker D, Tun\u00e7 B, Watanabe T, Elliott MA, Ruparel K, et al. Harmonization of multi-site diffusion tensor imaging data. NeuroImage. 2017 Nov;161:149\u201370. https:\/\/doi.org\/10.1016\/j.neuroimage.2017.08.047","DOI":"10.1016\/j.neuroimage.2017.08.047"},{"key":"2191_CR20","doi-asserted-by":"publisher","unstructured":"22.McInnes L, Healy J, Saul N, Gro\u00dfberger L. UMAP: Uniform Manifold Approximation and Projection. JOSS. 2018 Sep 2;3(29):861. https:\/\/doi.org\/10.21105\/joss.00861","DOI":"10.21105\/joss.00861"},{"key":"2191_CR21","doi-asserted-by":"publisher","unstructured":"Zhang Y, Ma X, Li M, Huang K, Zhu J, Wang M, et al. Generalist medical foundation model improves prostate cancer segmentation from multimodal MRI images. npj Digit Med. 2025 Jun 18;8(1):372. https:\/\/doi.org\/10.1038\/s41746-025-01756-2","DOI":"10.1038\/s41746-025-01756-2"},{"key":"2191_CR22","doi-asserted-by":"publisher","unstructured":"McInnes L, Healy J, Melville J. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction [Internet]. arXiv; 2020 [cited 2025 Sep 17]. Available from: http:\/\/arxiv.org\/abs\/1802.03426https:\/\/doi.org\/10.48550\/arXiv.1802.03426","DOI":"10.48550\/arXiv.1802.03426"},{"key":"2191_CR23","doi-asserted-by":"publisher","unstructured":"Oerther B, Engel H, Bamberg F, Sigle A, Gratzke C, Benndorf M. Cancer detection rates of the PI-RADSv2.1 assessment categories: systematic review and meta-analysis on lesion level and patient level. Prostate Cancer Prostatic Dis. 2022 Jun;25(2):256\u201363. https:\/\/doi.org\/10.1038\/s41391-021-00417-1","DOI":"10.1038\/s41391-021-00417-1"},{"key":"2191_CR24","doi-asserted-by":"publisher","unstructured":"Alshuhri MS, Alhulail AA, Alqahtani AGM, Madkhali Y, Aljuhani M, Alghuraybi RA, et al. Stability of ADC measurements in diffusion MRI: A multicenter phantom study using multi-point b-values. Radiography. 2025 Jan;31(1):212\u20139. https:\/\/doi.org\/10.1016\/j.radi.2024.11.009","DOI":"10.1016\/j.radi.2024.11.009"},{"key":"2191_CR25","doi-asserted-by":"publisher","unstructured":"Rajagopal A, Redekop E, Kemisetti A, Kulkarni R, Raman S, Sarma K, et al. Federated Learning with Research Prototypes: Application to Multi-Center MRI-based Detection of Prostate Cancer with Diverse Histopathology. Academic Radiology. 2023 Apr;30(4):644\u201357. https:\/\/doi.org\/10.1016\/j.acra.2023.02.012","DOI":"10.1016\/j.acra.2023.02.012"},{"key":"2191_CR26","doi-asserted-by":"publisher","unstructured":"Wen J, Zhang Z, Lan Y, Cui Z, Cai J, Zhang W. A survey on federated learning: challenges and applications. Int J Mach Learn & Cyber. 2023 Feb;14(2):513\u201335. https:\/\/doi.org\/10.1007\/s13042-022-01647-y","DOI":"10.1007\/s13042-022-01647-y"},{"key":"2191_CR27","doi-asserted-by":"publisher","unstructured":"Rosenkrantz AB, Taneja SS. Radiologist, Be Aware: Ten Pitfalls That Confound the Interpretation of Multiparametric Prostate MRI. American Journal of Roentgenology. 2014 Jan;202(1):109\u201320. https:\/\/doi.org\/10.2214\/AJR.13.10699","DOI":"10.2214\/AJR.13.10699"},{"key":"2191_CR28","doi-asserted-by":"publisher","unstructured":"Zaridis DI, Mylona E, Tsiknakis N, Tachos NS, Matsopoulos GK, Marias K, et al. ProLesA-Net: A multi-channel 3D architecture for prostate MRI lesion segmentation with multi-scale channel and spatial attentions. Patterns. 2024 Jul;5(7):100992. https:\/\/doi.org\/10.1016\/j.patter.2024.100992","DOI":"10.1016\/j.patter.2024.100992"},{"key":"2191_CR29","doi-asserted-by":"publisher","unstructured":"Roge A, Hiremath A, Sobota M, Tirumani SH, Kayat Bittencourt L, Ream J, et al. Evaluating the sensitivity of deep learning to inter-reader variations in lesion delineations on bi-parametric MRI in identifying clinically significant prostate cancer. In: Iftekharuddin KM, Drukker K, Mazurowski MA, Lu H, Muramatsu C, Samala RK, editors. Medical Imaging 2022: Computer-Aided Diagnosis [Internet]. San Diego, United States: SPIE; 2022 [cited 2026 Jul 4]. p. 41. Available from: https:\/\/www.spiedigitallibrary.org\/conference-proceedings-of-spie\/12033\/2613245\/Evaluating-the-sensitivity-of-deep-learning-to-inter-reader-variations\/https:\/\/doi.org\/10.1117\/12.2613245.full 10.1117\/12.2613245","DOI":"10.1117\/12.2613245.full"}],"container-title":["Journal of Imaging Informatics in Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-026-02191-9.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10278-026-02191-9","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-026-02191-9.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T16:38:46Z","timestamp":1787762326000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10278-026-02191-9"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,26]]},"references-count":29,"alternative-id":["2191"],"URL":"https:\/\/doi.org\/10.1007\/s10278-026-02191-9","relation":{},"ISSN":["2948-2933"],"issn-type":[{"value":"2948-2933","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,26]]},"assertion":[{"value":"7 March 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"15 July 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"31 July 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"26 August 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"IRB at Indiana University approved this study. A copy of the approval letter is provided as a supplementary document.","order":1,"name":"Ethics","label":"Ethics Approval","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"IRB waived the requirement for informed consent since retrospective and de-identified datasets were used in this study.","order":2,"name":"Ethics","label":"Consent to Participate and 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"}}]}}