{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T06:17:33Z","timestamp":1788329853622,"version":"build-2803163510"},"reference-count":42,"publisher":"Ovid Technologies (Wolters Kluwer Health)","issue":"17","content-domain":{"domain":["www.ahajournals.org"],"crossmark-restriction":true},"short-container-title":["JAHA"],"published-print":{"date-parts":[[2026,9,1]]},"abstract":"<jats:sec xml:lang=\"en\">\n                    <jats:title>Background<\/jats:title>\n                    <jats:p xml:lang=\"en\">Cardiovascular disease (CVD) remains a leading cause of death, but population\u2010level screening for atherosclerosis often depends on specialized imaging that is not routinely available. We evaluated foundation models for opportunistic detection of carotid atherosclerosis from chest radiographs, assessed the prognostic value of the resulting biomarker for CVD death, and examined model explainability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec xml:lang=\"en\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p xml:lang=\"en\">Among 5785 participants with carotid sonography annotations, we compared Rad\u2010DINO, DINOv2, OpenCLIP, and CheXagent for chest radiograph\u2013based atherosclerosis detection using linear probing, selective fine\u2010tuning, and low\u2010rank adaptation. Performance was assessed by area under the receiver operating characteristic curve with 95% CIs. Explainability was evaluated using gradient\u2010weighted class activation mapping and a clinical relevance index, which quantified attention within cardiovascular anatomic masks. The selected model generated a Deep\u2010Learning Chest X\u2010Ray Atherosclerosis Score, which was tested for association with CVD death in 32\u2009524 participants using Framingham Risk Score\u2013adjusted Cox models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec xml:lang=\"en\">\n                    <jats:title>Results<\/jats:title>\n                    <jats:p xml:lang=\"en\">\n                      Rad\u2010DINO achieved the highest performance with selective fine\u2010tuning (area under the receiver operating characteristic curve, 0.74 [95% CI, 0.71\u20130.77]) and low\u2010rank adaptation (area under the receiver operating characteristic curve, 0.74 [95% CI, 0.70\u20130.76]), with the highest anatomic relevance under low\u2010rank adaptation (clinical relevance index, 0.31). Higher Deep\u2010Learning Chest X\u2010Ray Atherosclerosis Score tertiles were strongly associated with CVD death (adjusted hazard ratio, 9.36 [95% CI, 3.35\u201326.15] for high versus low;\n                      <jats:italic toggle=\"no\">P<\/jats:italic>\n                      \u2010trend&lt;0.005). The model concordance index was 0.73 (95% CI, 0.67\u20130.75).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec xml:lang=\"en\">\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p xml:lang=\"en\">Foundation models enable opportunistic detection of carotid atherosclerosis from chest radiographs and yield a biomarker that provides prognostic information for CVD death. Explainability analyses demonstrated alignment with CVD anatomic regions, supporting clinical relevance and scalable risk stratification.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1161\/jaha.125.048485","type":"journal-article","created":{"date-parts":[[2026,8,20]],"date-time":"2026-08-20T11:30:41Z","timestamp":1787225441000},"update-policy":"https:\/\/doi.org\/10.1161\/crossmarkpolicy","source":"Crossref","is-referenced-by-count":0,"title":["Opportunistic Screening of Carotid Atherosclerosis and Cardiovascular Mortality Risk Using Chest Radiographs: A\u00a0Comparative Study of Foundation Models"],"prefix":"10.1161","volume":"15","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7085-0105","authenticated-orcid":false,"given":"Azka","family":"Rehman","sequence":"first","affiliation":[{"name":"Seoul National University Graduate School","place":["Seoul, South Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7942-1286","authenticated-orcid":false,"given":"Jaewon","family":"Kim","sequence":"additional","affiliation":[{"name":"Seoul National University Graduate School","place":["Seoul, South Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-4547-0861","authenticated-orcid":false,"given":"Hyeokjong","family":"Lee","sequence":"additional","affiliation":[{"name":"Seoul National University Graduate School","place":["Seoul, South Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8586-0645","authenticated-orcid":false,"given":"Jooyoung","family":"Chang","sequence":"additional","affiliation":[{"name":"XAIMED Co. Ltd","place":["Seoul, South Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7498-4829","authenticated-orcid":false,"given":"Sang Min","family":"Park","sequence":"additional","affiliation":[{"name":"Seoul National University Graduate School","place":["Seoul, South Korea"]},{"name":"Seoul National University Hospital","place":["Seoul, South Korea"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"276","reference":[{"key":"e_1_3_2_2_2","volume-title":"Deaths from cardiovascular disease surged 60% globally over the last 30\u2009years: report","author":"World Heart Federation","year":"2023","unstructured":"World Heart Federation . Deaths from cardiovascular disease surged 60% globally over the last 30\u2009years: report. World Heart Federation; 2023. Accessed February 3, 2024. https:\/\/world\u2010heart\u2010federation.org\/news\/deaths\u2010from\u2010cardiovascular\u2010disease\u2010surged\u201060\u2010globally\u2010over\u2010the\u2010last\u201030\u2010years\u2010report\/"},{"key":"e_1_3_2_3_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.atherosclerosis.2012.11.008"},{"key":"e_1_3_2_4_2","doi-asserted-by":"publisher","DOI":"10.1001\/archneur.59.7.1162"},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1016\/s1474-4422(20)30484-1"},{"key":"e_1_3_2_6_2","doi-asserted-by":"publisher","DOI":"10.1093\/eurheartj\/eht151"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.3389\/fphys.2023.1225636"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1002\/jcsm.13144"},{"key":"e_1_3_2_9_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-39631-x"},{"key":"e_1_3_2_10_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10916-023-02030-2"},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","DOI":"10.1177\/20552076231191055"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2020.03.027"},{"key":"e_1_3_2_13_2","first-page":"134","volume-title":"International conference on medical image computing and computer\u2010assisted intervention","author":"Yang Z","year":"2024","unstructured":"Yang Z, Zhang J, Wang G, Kalra MK, Yan P. Cardiovascular disease detection from multi\u2010view chest X\u2010rays with bi\u2010mamba. International conference on medical image computing and computer\u2010assisted intervention. Springer; 2024:134\u2013144."},{"key":"e_1_3_2_14_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.atherosclerosis.2009.11.012"},{"key":"e_1_3_2_15_2","doi-asserted-by":"publisher","DOI":"10.1148\/ryct.2021200486"},{"key":"e_1_3_2_16_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.mcpdig.2025.100300"},{"key":"e_1_3_2_17_2","doi-asserted-by":"publisher","DOI":"10.1093\/ehjdh\/ztaf033"},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1109\/RBME.2025.3531360"},{"key":"e_1_3_2_19_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.102996"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/N19-1423"},{"key":"e_1_3_2_21_2","first-page":"1877","article-title":"Language models are few\u2010shot learners","volume":"33","author":"Brown T","year":"2020","unstructured":"Brown T, Mann B, Ryder N, Subbiah M, Kaplan JD, Dhariwal P, Neelakantan A, Shyam P, Sastry G, Askell A, et\u00a0al. Language models are few\u2010shot learners. Adv Neural Inf Proces Syst. 2020;33:1877\u20131901.","journal-title":"Adv Neural Inf Proces Syst"},{"key":"e_1_3_2_22_2","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2304.07193"},{"key":"e_1_3_2_23_2","first-page":"8748","volume-title":"International conference on machine learning","author":"Radford A","year":"2021","unstructured":"Radford A, Kim JW, Hallacy C, Ramesh A, Goh G, Agarwal S, Sastry G, Askell A, Mishkin P, Clark J, et\u00a0al. Learning transferable visual models from natural language supervision. International conference on machine learning. PMLR; 2021:8748\u20138763."},{"key":"e_1_3_2_24_2","first-page":"12888","volume-title":"International conference on machine learning","author":"Li J","year":"2022","unstructured":"Li J, Li D, Xiong C, Hoi S. BLIP: bootstrapping language\u2010image pre\u2010training for unified vision\u2010language understanding and generation. International conference on machine learning. PMLR; 2022:12888\u201312900."},{"key":"e_1_3_2_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"e_1_3_2_26_2","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0868"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/WACV57701.2024.00746"},{"key":"e_1_3_2_28_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4842-6168-2_6"},{"key":"e_1_3_2_29_2","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929"},{"key":"e_1_3_2_30_2","unstructured":"Dosovitskiy A Beyer L Kolesnikov A Weissenborn D Zhai X Unterthiner T Dehghani M Minderer M Heigold G Gelly S et\u00a0al. An image is worth 16\u00d716 words: transformers for image recognition at scale. Proceedings of the International Conference on Learning Representations (ICLR); 2021."},{"key":"e_1_3_2_31_2","unstructured":"Hu EJ Shen Y Wallis P Allen\u2010Zhu Z Li Y Wang S Wang L Chen W. LoRA: low\u2010rank adaptation of large language models. Proceedings of the International Conference on Learning Representations (ICLR); 2022."},{"key":"e_1_3_2_32_2","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-024-00965-w"},{"key":"e_1_3_2_33_2","first-page":"2818","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Cherti M","year":"2023","unstructured":"Cherti M, Beaumont R, Wightman R, Wortsman M, Ilharco G, Gordon C, Schuhmann C, Schmidt L, Jitsev J. Reproducible scaling laws for contrastive language\u2010image learning. Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition. IEEE; 2023:2818\u20132829."},{"key":"e_1_3_2_34_2","unstructured":"Chen Z Varma M Delbrouck JB Paschali M Blankemeier L Van Veen D Valanarasu JMJ Youssef A Cohen JP Reis EP et\u00a0al. CheXagent: towards a foundation model for chest X\u2010ray interpretation. AAAI 2024 Spring Symposium on Clinical Foundation Models March 25\u201327 2024 Stanford California; 2024."},{"key":"e_1_3_2_35_2","unstructured":"Seibold C. ChestXRayAnatomySegmentation. GitHub. 2023. Accessed February 4 2024. https:\/\/github.com\/ConstantinSeibold\/ChestXRayAnatomySegmentation."},{"key":"e_1_3_2_36_2","unstructured":"Zhang H Cisse M Dauphin YN Lopez\u2010Paz D. mixup: beyond empirical risk minimization. In: International Conference on Learning Representations (ICLR); 2018."},{"key":"e_1_3_2_37_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.echo.2007.11.011"},{"key":"e_1_3_2_38_2","doi-asserted-by":"publisher","DOI":"10.1097\/HJH.0b013e3281fc975a"},{"key":"e_1_3_2_39_2","doi-asserted-by":"publisher","DOI":"10.1056\/NEJM199901073400103"},{"key":"e_1_3_2_40_2","doi-asserted-by":"publisher","DOI":"10.21105\/joss.01317"},{"key":"e_1_3_2_41_2","first-page":"2825","article-title":"Scikit\u2010learn: machine learning in Python","volume":"12","author":"Pedregosa F","year":"2011","unstructured":"Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, et\u00a0al. Scikit\u2010learn: machine learning in Python. J Mach Learn Res. 2011;12:2825\u20132830.","journal-title":"J Mach Learn Res"},{"key":"e_1_3_2_42_2","first-page":"8024","volume-title":"Advances in neural information processing systems","author":"Paszke A","year":"2019","unstructured":"Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, Killeen T, Lin Z, Gimelshein N, Antiga L, et\u00a0al. PyTorch: an imperative style, high\u2010performance deep learning library. Advances in neural information processing systems. ACM; 2019:8024\u20138035."},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.3390\/info11020108"}],"container-title":["Journal of the American Heart Association"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.ahajournals.org\/doi\/full\/10.1161\/JAHA.125.048485","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T06:02:39Z","timestamp":1788328959000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ahajournals.org\/doi\/10.1161\/JAHA.125.048485"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,1]]},"references-count":42,"journal-issue":{"issue":"17","published-print":{"date-parts":[[2026,9,1]]}},"alternative-id":["10.1161\/JAHA.125.048485"],"URL":"https:\/\/doi.org\/10.1161\/jaha.125.048485","relation":{},"ISSN":["2047-9980"],"issn-type":[{"value":"2047-9980","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,1]]},"assertion":[{"value":"2025-12-08","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-06-29","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-20","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e048485"}}