{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,24]],"date-time":"2026-10-24T13:37:31Z","timestamp":1792849051202,"version":"4.5.0"},"reference-count":27,"publisher":"Springer Science and Business Media LLC","issue":"5","license":[{"start":{"date-parts":[[2025,12,19]],"date-time":"2025-12-19T00:00:00Z","timestamp":1766102400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,12,19]],"date-time":"2025-12-19T00:00:00Z","timestamp":1766102400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/100015220","name":"Winthrop P. Rockefeller Cancer Institute, University of Arkansas for Medical Sciences","doi-asserted-by":"publisher","award":["Team Science Award"],"award-info":[{"award-number":["Team Science Award"]}],"id":[{"id":"10.13039\/100015220","id-type":"DOI","asserted-by":"publisher"}]}],"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                    Low\u2011dose CT (LDCT) screening reduces lung cancer mortality but yields high false\u2011positive findings, motivating practical AI that aligns with slice\u2011based clinical review. We evaluate a label\u2011efficient 2D pipeline that uses frozen RAD\u2011DINO embeddings from native\u2011resolution axial slices and a lightweight multilayer perceptron for patient\u2011level risk estimation via mean aggregation and isotonic calibration. Using the NLST CT arm with outcomes defined over a 0\u201324\u2011month window, we construct a fixed patient\u2011level split (one CT per patient; no cross\u2011split leakage) and perform 25 repeated imbalanced test draws (~\u20096% prevalence) to approximate screening conditions. At screening prevalence, RAD\u2011DINO\u2009+\u2009MLP achieves PR\u2011AUC\u2009=\u20090.705 (calibrated; raw 0.554) and ROC\u2011AUC\u2009=\u20090.817 (raw; 0.736 calibrated), with improved probability reliability following calibration; operating points are selected on validation and reported on test. For secondary ablations only, a near\u2011balanced cohort (\n                    <jats:italic>N<\/jats:italic>\n                    \u2009=\u20091984) yields accuracy 0.966, precision 0.974, recall 0.973, F1 0.973, and ROC\u2011AUC 0.912. Beyond classification, retrieval with triplet\u2011fine\u2011tuned embeddings attains Precision@5\u2009=\u20090.853. Interpretability analyses show that cancer cases sustain higher top\u2011k slice scores and that directional SHAP concentrates on a small subset of high\u2011probability slices; label\u2011colored t\u2011SNE provides qualitative views of embedding structure. Limitations include single\u2011cohort evaluation, lack of Lung\u2011RADS labels in public NLST, and a CXR\u2009\u2192\u2009CT pretraining shift; future work will pursue external validation and CT\u2011native self\u2011supervised continuation. Overall, frozen 2D foundation embeddings provide a strong, transparent, and computationally practical starting point for LDCT screening workflows under realistic prevalence.\n                  <\/jats:p>","DOI":"10.1007\/s10278-025-01748-4","type":"journal-article","created":{"date-parts":[[2025,12,19]],"date-time":"2025-12-19T17:33:08Z","timestamp":1766165588000},"page":"4015-4037","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Harnessing Native-Resolution 2D Embeddings for Lung Cancer Classification: A Feasibility Study with the RAD-DINO Self-supervised Foundation Model"],"prefix":"10.1007","volume":"39","author":[{"given":"Md.Enamul","family":"Hoq","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lawrence","family":"Tarbox","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"suffix":"Jr.","given":"Donald","family":"Johann","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Linda","family":"Larson-Prior","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fred","family":"Prior","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,12,19]]},"reference":[{"key":"1748_CR1","doi-asserted-by":"publisher","first-page":"395","DOI":"10.1056\/NEJMoa1102873","volume":"365","author":"The National Lung Screening Trial Research Team","year":"2011","unstructured":"The National Lung Screening Trial Research Team. Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening. N Engl J Med 2011;365:395\u2013409. https:\/\/doi.org\/10.1056\/NEJMoa1102873.","journal-title":"N Engl J Med"},{"key":"1748_CR2","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1148\/radiol.10091808","volume":"258","author":"National Lung Screening Trial Research Team","year":"2011","unstructured":"National Lung Screening Trial Research Team. The National Lung Screening Trial: Overview and Study Design. Radiology 2011;258:243\u201353. https:\/\/doi.org\/10.1148\/radiol.10091808.","journal-title":"Radiology"},{"key":"1748_CR3","doi-asserted-by":"publisher","first-page":"1045","DOI":"10.1007\/s10278-013-9622-7","volume":"26","author":"K Clark","year":"2013","unstructured":"Clark K, Vendt B, Smith K, et al. The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information Repository. J Digit Imaging 2013;26:1045\u201357. https:\/\/doi.org\/10.1007\/s10278-013-9622-7.","journal-title":"J Digit Imaging"},{"key":"1748_CR4","doi-asserted-by":"publisher","first-page":"728","DOI":"10.1056\/NEJMoa1211776","volume":"368","author":"MC Tammem\u00e4gi","year":"2013","unstructured":"Tammem\u00e4gi MC, Katki HA, Hocking WG, et al. Selection Criteria for Lung-Cancer Screening. N Engl J Med 2013;368:728\u201336. https:\/\/doi.org\/10.1056\/NEJMoa1211776.","journal-title":"N Engl J Med"},{"key":"1748_CR5","doi-asserted-by":"publisher","first-page":"S79","DOI":"10.1102\/1470-7330.2011.9020","volume":"11","author":"Y Ru Zhao","year":"2011","unstructured":"Ru Zhao Y, Xie X, De Koning HJ, et al. NELSON lung cancer screening study. Cancer Imaging 2011;11:S79\u201384. https:\/\/doi.org\/10.1102\/1470-7330.2011.9020.","journal-title":"Cancer Imaging"},{"key":"1748_CR6","doi-asserted-by":"publisher","first-page":"9286","DOI":"10.1038\/s41598-018-27569-w","volume":"8","author":"JL Causey","year":"2018","unstructured":"Causey JL, Zhang J, Ma S, et al. Highly accurate model for prediction of lung nodule malignancy with CT scans. Sci Rep 2018;8:9286. https:\/\/doi.org\/10.1038\/s41598-018-27569-w.","journal-title":"Sci Rep"},{"key":"1748_CR7","doi-asserted-by":"publisher","unstructured":"McConnell N, Vasudev P, Yamada D, et al. A computationally frugal open-source foundation model for thoracic disease detection in lung cancer screening programs 2025. https:\/\/doi.org\/10.48550\/arXiv.2507.01881.","DOI":"10.48550\/arXiv.2507.01881"},{"key":"1748_CR8","doi-asserted-by":"publisher","first-page":"1523","DOI":"10.1038\/s41467-025-56822-w","volume":"16","author":"C Niu","year":"2025","unstructured":"Niu C, Lyu Q, Carothers CD, et al. Medical multimodal multitask foundation model for lung cancer screening. Nat Commun 2025;16:1523. https:\/\/doi.org\/10.1038\/s41467-025-56822-w.","journal-title":"Nat Commun"},{"key":"1748_CR9","doi-asserted-by":"publisher","first-page":"954","DOI":"10.1038\/s41591-019-0447-x","volume":"25","author":"D Ardila","year":"2019","unstructured":"Ardila D, Kiraly AP, Bharadwaj S, et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography. Nat Med 2019;25:954\u201361. https:\/\/doi.org\/10.1038\/s41591-019-0447-x.","journal-title":"Nat Med"},{"key":"1748_CR10","doi-asserted-by":"publisher","unstructured":"Yu J, Yang B, Wang J, et al. 2D CNN versus 3D CNN for false-positive reduction in lung cancer screening. J Med Imaging 2020;7. https:\/\/doi.org\/10.1117\/1.JMI.7.5.051202.","DOI":"10.1117\/1.JMI.7.5.051202"},{"key":"1748_CR11","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.media.2017.06.015","volume":"42","author":"AAA Setio","year":"2017","unstructured":"Setio AAA, Traverso A, De Bel T, et al. Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge. Med Image Anal 2017;42:1\u201313. https:\/\/doi.org\/10.1016\/j.media.2017.06.015.","journal-title":"Med Image Anal"},{"key":"1748_CR12","doi-asserted-by":"publisher","first-page":"365","DOI":"10.1148\/radiol.2019181960","volume":"292","author":"J Choe","year":"2019","unstructured":"Choe J, Lee SM, Do K-H, et al. Deep Learning\u2013based Image Conversion of CT Reconstruction Kernels Improves Radiomics Reproducibility for Pulmonary Nodules or Masses. Radiology 2019;292:365\u201373. https:\/\/doi.org\/10.1148\/radiol.2019181960.","journal-title":"Radiology"},{"key":"1748_CR13","doi-asserted-by":"publisher","unstructured":"Avesta A, Hossain S, Lin M, et al. Comparing 3D, 2.5D, and 2D Approaches to Brain Image Auto-Segmentation. Bioengineering 2023;10:181. https:\/\/doi.org\/10.3390\/bioengineering10020181.","DOI":"10.3390\/bioengineering10020181"},{"key":"1748_CR14","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1038\/s42256-024-00965-w","volume":"7","author":"F P\u00e9rez-Garc\u00eda","year":"2025","unstructured":"P\u00e9rez-Garc\u00eda F, Sharma H, Bond-Taylor S, et al. Exploring scalable medical image encoders beyond text supervision. Nat Mach Intell 2025;7:119\u201330. https:\/\/doi.org\/10.1038\/s42256-024-00965-w.","journal-title":"Nat Mach Intell"},{"key":"1748_CR15","doi-asserted-by":"publisher","unstructured":"Yang L, Xu S, Sellergren A, et al. Advancing Multimodal Medical Capabilities of Gemini 2024. https:\/\/doi.org\/10.48550\/arXiv.2405.03162.","DOI":"10.48550\/arXiv.2405.03162"},{"key":"1748_CR16","doi-asserted-by":"publisher","DOI":"10.7937\/TCIA.HMQ8-J677","author":"National Lung Screening Trial Research Team","year":"2013","unstructured":"National Lung Screening Trial Research Team. Data from the National Lung Screening Trial (NLST) 2013. https:\/\/doi.org\/10.7937\/TCIA.HMQ8-J677.","journal-title":"Data from the National Lung Screening Trial (NLST)"},{"key":"1748_CR17","doi-asserted-by":"publisher","unstructured":"Dosovitskiy A, Beyer L, Kolesnikov A, et al. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale 2020. https:\/\/doi.org\/10.48550\/ARXIV.2010.11929.","DOI":"10.48550\/ARXIV.2010.11929"},{"key":"1748_CR18","doi-asserted-by":"publisher","unstructured":"Oquab M, Darcet T, Moutakanni T, et al. DINOv2: Learning Robust Visual Features without Supervision 2023. https:\/\/doi.org\/10.48550\/ARXIV.2304.07193.","DOI":"10.48550\/ARXIV.2304.07193"},{"key":"1748_CR19","doi-asserted-by":"publisher","unstructured":"He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition 2015. https:\/\/doi.org\/10.48550\/ARXIV.1512.03385.","DOI":"10.48550\/ARXIV.1512.03385"},{"key":"1748_CR20","doi-asserted-by":"publisher","unstructured":"Hoffer E, Ailon N. Deep metric learning using Triplet network 2014. https:\/\/doi.org\/10.48550\/ARXIV.1412.6622.","DOI":"10.48550\/ARXIV.1412.6622"},{"key":"1748_CR21","doi-asserted-by":"publisher","unstructured":"Lundberg S, Lee S-I. A Unified Approach to Interpreting Model Predictions 2017. https:\/\/doi.org\/10.48550\/ARXIV.1705.07874.","DOI":"10.48550\/ARXIV.1705.07874"},{"key":"1748_CR22","doi-asserted-by":"publisher","first-page":"1273253","DOI":"10.3389\/fpubh.2023.1273253","volume":"11","author":"M Li","year":"2023","unstructured":"Li M, Jiang Y, Zhang Y, et al. Medical image analysis using deep learning algorithms. Front Public Health 2023;11:1273253. https:\/\/doi.org\/10.3389\/fpubh.2023.1273253.","journal-title":"Front Public Health"},{"key":"1748_CR23","doi-asserted-by":"publisher","unstructured":"Pai S, Hadzic I, Bontempi D, et al. Vision Foundation Models for Computed Tomography 2025. https:\/\/doi.org\/10.48550\/ARXIV.2501.09001.","DOI":"10.48550\/ARXIV.2501.09001"},{"key":"1748_CR24","doi-asserted-by":"publisher","unstructured":"Blankemeier L, Cohen JP, Kumar A, et al. Merlin: A Vision Language Foundation Model for 3D Computed Tomography 2024. https:\/\/doi.org\/10.48550\/arXiv.2406.06512.","DOI":"10.48550\/arXiv.2406.06512"},{"key":"1748_CR25","doi-asserted-by":"publisher","unstructured":"Bommasani R, Hudson DA, Adeli E, et al. On the Opportunities and Risks of Foundation Models 2021. https:\/\/doi.org\/10.48550\/ARXIV.2108.07258.","DOI":"10.48550\/ARXIV.2108.07258"},{"key":"1748_CR26","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1038\/s41591-018-0300-7","volume":"25","author":"EJ Topol","year":"2019","unstructured":"Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med 2019;25:44\u201356. https:\/\/doi.org\/10.1038\/s41591-018-0300-7.","journal-title":"Nat Med"},{"key":"1748_CR27","doi-asserted-by":"publisher","first-page":"485","DOI":"10.7326\/M14-2086","volume":"162","author":"PF Pinsky","year":"2015","unstructured":"Pinsky PF, Gierada DS, Black W, et al. Performance of Lung-RADS in the National Lung Screening Trial: A Retrospective Assessment. Ann Intern Med 2015;162:485\u201391. https:\/\/doi.org\/10.7326\/M14-2086.","journal-title":"Ann Intern Med"}],"container-title":["Journal of Imaging Informatics in Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-025-01748-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10278-025-01748-4","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10278-025-01748-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,24]],"date-time":"2026-10-24T13:02:06Z","timestamp":1792846926000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10278-025-01748-4"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,19]]},"references-count":27,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,10]]}},"alternative-id":["1748"],"URL":"https:\/\/doi.org\/10.1007\/s10278-025-01748-4","relation":{},"ISSN":["2948-2933"],"issn-type":[{"value":"2948-2933","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,19]]},"assertion":[{"value":"30 June 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"31 October 2025","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 November 2025","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"19 December 2025","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Not applicable. This study used publicly available, de-identified data (NLST) and did not require IRB approval.","order":1,"name":"Ethics","label":"Ethics Approval","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable (public dataset, no human subjects directly involved).","order":2,"name":"Ethics","label":"Consent to Participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"All authors have reviewed and approved the final version of the manuscript for publication.","order":3,"name":"Ethics","label":"Consent for Publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","label":"Competing Interests","group":{"name":"EthicsHeading","label":"Declarations"}}]}}