{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T04:50:00Z","timestamp":1791521400926,"version":"4.3.4"},"reference-count":23,"publisher":"MDPI AG","issue":"19","license":[{"start":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T00:00:00Z","timestamp":1791244800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T00:00:00Z","timestamp":1791244800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J. Clin. Med.","JCM"],"abstract":"<jats:p>Background\/Objectives: Multimodal large language models are emerging tools for medical image interpretation, but their performance on isolated diffusion-weighted imaging (DWI) for acute ischemic stroke remains largely unexplored. We evaluated GPT-5 for stroke detection and vascular territory classification on preselected representative DWI images using a structured two-step prompting protocol. Methods: This single-center, retrospective, balanced case\u2013control study included 220 cases (110 stroke patients and 110 stroke-negative controls). Four representative axial DWI slices per case were provided without apparent diffusion coefficient maps. Step 1 queried stroke presence and vascular territory; Step 2 queried the middle cerebral artery (MCA) segment-associated territory (M1\u2013M4) following an MCA response. The reference standard was radiologist\u2013neurologist consensus based on full multimodal magnetic resonance imaging, blinded to GPT-5 outputs. Diagnostic performance metrics and Cohen\u2019s kappa (\u03ba) were calculated. Results: Stroke detection showed a sensitivity of 100% (95% confidence interval [CI], 96.6\u2013100%), specificity of 83.6% (95% CI, 75.6\u201389.4%), accuracy of 91.8%, and \u03ba = 0.836. Accuracy was 93.6% for anterior\/posterior circulation (\u03ba = 0.838) and 90.9% for vascular territory (\u03ba = 0.799). Conditional MCA segment-associated territory accuracy was 71.6% (53\/74; \u03ba = 0.602); end-to-end accuracy was 66.3% (53\/80). Exploratory analysis of the 18 false-positive cases identified four overlapping error categories: artifactual and physiologic signal misinterpretation (72.2%), chronic lesion misclassification (27.8%), terminological inconsistency or unsupported report statements (77.8%), and etiological scenarios or unhedged definitive diagnoses (38.9%). Conclusions: GPT-5 achieved high sensitivity and a specificity of 83.6% for stroke classification on preselected representative DWI images using a vascular territory-based diagnostic framework. However, the 16.4% false-positive rate indicates that positive findings require radiologist confirmation. These findings support further evaluation of GPT-5 as a potential clinician-assisted image-interpretation tool but do not establish a clinical decision-support role.<\/jats:p>","DOI":"10.3390\/jcm15197716","type":"journal-article","created":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T08:28:57Z","timestamp":1791361737000},"page":"7716","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Diagnostic Accuracy of GPT-5 in Detecting and Localizing Acute Ischemic Stroke on Isolated Diffusion-Weighted Imaging: A Retrospective Case\u2013Control Study"],"prefix":"10.3390","volume":"15","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6018-6105","authenticated-orcid":false,"given":"Emre","family":"G\u00f6k\u00e7en","sequence":"first","affiliation":[{"name":"Department of Emergency Medicine, Faculty of Medicine, Yozgat Bozok University, 66100 Yozgat, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4075-8601","authenticated-orcid":false,"given":"Mikail","family":"Ku\u015fdo\u011fan","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Faculty of Medicine, Yozgat Bozok University, 66100 Yozgat, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6266-5177","authenticated-orcid":false,"given":"Mustafa Fatih","family":"Erkoc","sequence":"additional","affiliation":[{"name":"Department of Radiology, Faculty of Medicine, Yozgat Bozok University, 66100 Yozgat, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4288-8170","authenticated-orcid":false,"given":"Levent","family":"Albayrak","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Faculty of Medicine, Yozgat Bozok University, 66100 Yozgat, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"G\u00fcne\u015f Seda","family":"Albayrak","sequence":"additional","affiliation":[{"name":"Department of Neurology, Faculty of Medicine, Yozgat Bozok University, 66100 Yozgat, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Deniz","family":"Tanr\u0131verdi","sequence":"additional","affiliation":[{"name":"Department of Emergency Medicine, Dr. Nevruz Erez State Hospital, 76000 I\u011fd\u0131r, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,10,6]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"523","DOI":"10.1016\/j.mayocp.2018.02.013","article-title":"Diagnosis and management of acute ischemic stroke","volume":"93","author":"Hasan","year":"2018","journal-title":"Mayo Clin. Proc."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"1230","DOI":"10.1161\/CIRCRESAHA.121.319948","article-title":"Advances in acute ischemic stroke therapy","volume":"130","author":"Xiong","year":"2022","journal-title":"Circ. Res."},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Koyun, M., and Taskent, I. (2025). Evaluation of advanced artificial intelligence algorithms\u2019 diagnostic efficacy in acute ischemic stroke: A comparative analysis of ChatGPT-4o and Claude 3.5 Sonnet models. J. Clin. Med., 14.","DOI":"10.20944\/preprints202412.2282.v1"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Alis, C., Ay, E., Genc, G., and Bulut, S. (2025). Evaluating deep learning-based commercial software for detecting ischemic lesions on DWI in stroke patients. Diagnostics, 15.","DOI":"10.3390\/diagnostics15182357"},{"key":"ref_5","first-page":"187","article-title":"A retrospective evaluation of the potential of ChatGPT in the accurate diagnosis of acute stroke","volume":"31","author":"Kuzan","year":"2025","journal-title":"Diagn. Interv. Radiol."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"1231","DOI":"10.1007\/s11604-024-01619-y","article-title":"Diagnostic performances of GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro in \u201cDiagnosis Please\u201d cases","volume":"42","author":"Sonoda","year":"2024","journal-title":"Jpn. J. Radiol."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"345","DOI":"10.1016\/j.diii.2025.04.006","article-title":"Comparison between multimodal foundation models and radiologists for the diagnosis of challenging neuroradiology cases with text and images","volume":"106","author":"Bruge","year":"2025","journal-title":"Diagn. Interv. Imaging"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"1111","DOI":"10.1007\/s00330-024-11115-6","article-title":"Revolution or risk?\u2014Assessing the potential and challenges of GPT-4V in radiologic image interpretation","volume":"35","author":"Huppertz","year":"2025","journal-title":"Eur. Radiol."},{"key":"ref_9","first-page":"27","article-title":"Artificial intelligence in radiology: Diagnostic sensitivity of ChatGPT for detecting hemorrhages in cranial computed tomography scans","volume":"32","author":"Dev","year":"2026","journal-title":"Diagn. Interv. Radiol."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Salehi, S., Singh, Y., Horst, K.K., Hathaway, Q.A., and Erickson, B.J. (2025). Agentic AI and large language models in radiology: Opportunities and hallucination challenges. Bioengineering, 12.","DOI":"10.3390\/bioengineering12121303"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"330","DOI":"10.1038\/s43856-025-01021-3","article-title":"Multi-model assurance analysis showing large language models are highly vulnerable to adversarial hallucination attacks during clinical decision support","volume":"5","author":"Omar","year":"2025","journal-title":"Commun. Med."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"265","DOI":"10.1007\/s00330-025-11807-7","article-title":"Motion artifacts and image quality in stroke MRI: Associated factors and impact on AI and human diagnostic accuracy","volume":"36","author":"Krag","year":"2026","journal-title":"Eur. Radiol."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"765","DOI":"10.1016\/j.clinimag.2015.06.001","article-title":"Readout-segmented echo-planar imaging in diffusion-weighted MR imaging of acute infarction of the brainstem and posterior fossa: Comparison of single-shot echo-planar diffusion-weighted sequences","volume":"39","author":"Byeon","year":"2015","journal-title":"Clin. Imaging"},{"key":"ref_14","first-page":"1434","article-title":"False-negative diffusion-weighted MR findings in acute ischemic stroke","volume":"21","author":"Oppenheim","year":"2000","journal-title":"AJNR Am. J. Neuroradiol."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"123","DOI":"10.2463\/mrms.mp.2016-0031","article-title":"CSF pulsation artifacts on ADC maps obtained with readout-segmented EPI","volume":"16","author":"Ohashi","year":"2017","journal-title":"Magn. Reson. Med. Sci."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"381","DOI":"10.1016\/S0895-6111(03)00013-2","article-title":"DWI findings of periventricular ischemic changes in patients with leukoaraiosis","volume":"27","author":"Calli","year":"2003","journal-title":"Comput. Med. Imaging Graph."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"139","DOI":"10.1111\/jon.12493","article-title":"Neuroimaging of dilated perivascular spaces: From benign and pathologic causes to mimics","volume":"28","author":"Rudie","year":"2018","journal-title":"J. Neuroimaging"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"815","DOI":"10.1007\/s13244-018-0637-y","article-title":"Magnetic resonance imaging of arterial stroke mimics: A pictorial review","volume":"9","author":"Adam","year":"2018","journal-title":"Insights Into Imaging"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"2456550","DOI":"10.1155\/2022\/2456550","article-title":"Deep learning in ischemic stroke imaging analysis: A comprehensive review","volume":"2022","author":"Cui","year":"2022","journal-title":"BioMed Res. Int."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"2154","DOI":"10.1038\/s41598-022-06021-0","article-title":"Development and clinical application of a deep learning model to identify acute infarct on magnetic resonance imaging","volume":"12","author":"Bridge","year":"2022","journal-title":"Sci. Rep."},{"key":"ref_21","first-page":"100287","article-title":"Large language models in radiology reporting\u2014A systematic review of performance, limitations, and clinical implications","volume":"12","author":"Artsi","year":"2025","journal-title":"Intell.-Based Med."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"190","DOI":"10.1038\/s41746-024-01185-7","article-title":"Hidden flaws behind expert-level accuracy of multimodal GPT-4 vision in medicine","volume":"7","author":"Jin","year":"2024","journal-title":"npj Digit. Med."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"61","DOI":"10.1038\/s43856-021-00062-8","article-title":"Deep learning-based detection and segmentation of diffusion abnormalities in acute ischemic stroke","volume":"1","author":"Liu","year":"2021","journal-title":"Commun. Med."}],"container-title":["Journal of Clinical Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2077-0383\/15\/19\/7716\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T04:14:35Z","timestamp":1791519275000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2077-0383\/15\/19\/7716"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,6]]},"references-count":23,"journal-issue":{"issue":"19","published-online":{"date-parts":[[2026,10]]}},"alternative-id":["jcm15197716"],"URL":"https:\/\/doi.org\/10.3390\/jcm15197716","relation":{},"ISSN":["2077-0383"],"issn-type":[{"value":"2077-0383","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,6]]}}}