{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,23]],"date-time":"2026-08-23T16:30:52Z","timestamp":1787502652613,"version":"build-2736575974"},"reference-count":27,"publisher":"Public Library of Science (PLoS)","issue":"9","license":[{"start":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T00:00:00Z","timestamp":1757635200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:sec id=\"sec001\">\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>This study evaluates the diagnostic performance of three multimodal large language models (LLMs)\u2014ChatGPT-4o, Gemini 2.0, and Claude 3.5\u2014in identifying pneumothorax from chest radiographs.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec id=\"sec002\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>In this retrospective analysis, 172 pneumothorax cases (148 patients aged &gt;12 years, 24 patients aged \u226412 years) with both chest radiographs and confirmatory thoracic CT were included from a tertiary emergency department. Patients were categorized by age and pneumothorax size (small\/large). Each radiograph was presented to all three LLMs accompanied by basic symptoms (dyspnea or chest pain), with each model analyzing each image three times. Diagnostic accuracy was evaluated using overall accuracy (all three responses correct), strict accuracy (\u22652 responses correct), and ideal accuracy (\u22651 response correct), alongside response consistency assessment using Fleiss\u2019 Kappa.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec id=\"sec003\">\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In patients older than 12 years, ChatGPT-4o demonstrated the highest overall accuracy (69.6%), followed by Claude 3.5 (64.9%) and Gemini 2.0 (57.4%). Performance was significantly poorer in pediatric patients across all models (20.8%, 12.5%, and 20.8%, respectively). For large pneumothorax in adults, ChatGPT-4o showed significantly higher accuracy compared to small pneumothorax (81.6% vs. 42.2%; p\u2009&lt;\u20090.001). Regarding consistency, Gemini 2.0 demonstrated excellent reliability for large pneumothorax (Kappa\u2009=\u20091.00), while Claude 3.5 showed moderate consistency across both pneumothorax sizes.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec id=\"sec004\">\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This study, the first to evaluate these three current multimodal LLMs in pneumothorax identification across different age groups, demonstrates promising results for potential clinical applications, particularly for adult patients with large pneumothorax. However, performance limitations in pediatric cases and with small pneumothoraces highlight the need for further validation before clinical implementation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1371\/journal.pone.0331962","type":"journal-article","created":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T17:42:54Z","timestamp":1757698974000},"page":"e0331962","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":5,"title":["New frontiers in radiologic interpretation: evaluating the effectiveness of large language models in pneumothorax diagnosis"],"prefix":"10.1371","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5629-3143","authenticated-orcid":true,"given":"Bensu","family":"Bulut","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Medine Akkan","family":"\u00d6z","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Murat","family":"Gen\u00e7","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9521-1120","authenticated-orcid":true,"given":"Ay\u015fenur","family":"G\u00fcr","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mehmet","family":"Yortanl\u0131","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bet\u00fcl \u00c7i\u011fdem","family":"Yortanl\u0131","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Oguz","family":"Sariyildiz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ramiz","family":"Yaz\u0131c\u0131","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1930-3293","authenticated-orcid":true,"given":"H\u00fcseyin","family":"Mutlu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mustafa Sirri","family":"Kotanoglu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eray","family":"Cinar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zekeriya","family":"Uykan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2025,9,12]]},"reference":[{"issue":"2","key":"pone.0331962.ref001","doi-asserted-by":"crossref","first-page":"121","DOI":"10.1159\/000135932","article-title":"Pneumothorax","volume":"76","author":"M Noppen","year":"2008","journal-title":"Respiration"},{"issue":"8","key":"pone.0331962.ref002","doi-asserted-by":"crossref","first-page":"666","DOI":"10.1136\/thorax.55.8.666","article-title":"Epidemiology of pneumothorax in England","volume":"55","author":"D Gupta","year":"2000","journal-title":"Thorax"},{"issue":"12","key":"pone.0331962.ref003","doi-asserted-by":"crossref","first-page":"868","DOI":"10.1056\/NEJM200003233421207","article-title":"Spontaneous pneumothorax","volume":"342","author":"SA Sahn","year":"2000","journal-title":"N Engl J Med"},{"issue":"7506","key":"pone.0331962.ref004","doi-asserted-by":"crossref","first-page":"1493","DOI":"10.1136\/bmj.330.7506.1493","article-title":"Radiological review of pneumothorax","volume":"330","author":"AR O\u2019Connor","year":"2005","journal-title":"BMJ"},{"issue":"1","key":"pone.0331962.ref005","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pone.0227652","article-title":"Early diagnosis of sepsis in emergency departments, time to treatment, and association with mortality: An observational study","volume":"15","author":"G Husab\u00f8","year":"2020","journal-title":"PLoS One"},{"issue":"10","key":"pone.0331962.ref006","first-page":"1290","article-title":"Prevalence, clinical manifestations and mortality rate in patients with spontaneous pneumothorax in thammasat university hospital","volume":"96","author":"N Saiphoklang","year":"2013","journal-title":"J Med Assoc Thai"},{"issue":"5","key":"pone.0331962.ref007","doi-asserted-by":"crossref","first-page":"1332","DOI":"10.1378\/chest.11-0348","article-title":"Thoracic ultrasonography for the pulmonary specialist","volume":"140","author":"SJ Koenig","year":"2011","journal-title":"Chest"},{"key":"pone.0331962.ref008","article-title":"Management of spontaneous pneumothorax: british thoracic society pleural disease guideline 2010","author":"A MacDuff","year":"2010","journal-title":"Thorax"},{"key":"pone.0331962.ref009","doi-asserted-by":"crossref","first-page":"102996","DOI":"10.1016\/j.media.2023.102996","article-title":"On the challenges and perspectives of foundation models for medical image analysis","volume":"91","author":"S Zhang","year":"2024","journal-title":"Med Image Anal"},{"issue":"7","key":"pone.0331962.ref010","doi-asserted-by":"crossref","first-page":"4096","DOI":"10.1097\/JS9.0000000000001359","article-title":"Step into the era of large multimodal models: a pilot study on ChatGPT-4V(ision)\u2019s ability to interpret radiological images","volume":"110","author":"L Zhu","year":"2024","journal-title":"Int J Surg"},{"key":"pone.0331962.ref011","doi-asserted-by":"crossref","first-page":"102611","DOI":"10.1016\/j.artmed.2023.102611","article-title":"Medical visual question answering: A survey","volume":"143","author":"Z Lin","year":"2023","journal-title":"Artif Intell Med"},{"issue":"5","key":"pone.0331962.ref012","doi-asserted-by":"crossref","first-page":"1532","DOI":"10.1109\/TMI.2022.3232411","article-title":"Medical Visual Question Answering via Conditional Reasoning and Contrastive Learning","volume":"42","author":"B Liu","year":"2023","journal-title":"IEEE Trans Med Imaging"},{"issue":"1","key":"pone.0331962.ref013","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1038\/s43856-023-00370-1","article-title":"The future landscape of large language models in medicine","volume":"3","author":"J Clusmann","year":"2023","journal-title":"Commun Med (Lond)"},{"issue":"3","key":"pone.0331962.ref014","article-title":"Comparative performance evaluation of multimodal large language models, radiologist, and anatomist in visual neuroanatomy questions","volume":"50","author":"YC G\u00fcne\u015f","year":"2024","journal-title":"Uludag Univ Med J"},{"issue":"2","key":"pone.0331962.ref015","doi-asserted-by":"crossref","first-page":"590","DOI":"10.1378\/chest.119.2.590","article-title":"Management of spontaneous pneumothorax: an American College of Chest Physicians Delphi consensus statement","volume":"119","author":"MH Baumann","year":"2001","journal-title":"Chest"},{"issue":"1","key":"pone.0331962.ref016","doi-asserted-by":"crossref","first-page":"181","DOI":"10.1038\/s41598-017-00284-8","article-title":"The Size of Spontaneous Pneumothorax is a Predictor of Unsuccessful Catheter Drainage","volume":"7","author":"T-M Tsai","year":"2017","journal-title":"Sci Rep"},{"key":"pone.0331962.ref017","article-title":"Management of spontaneous pneumothorax: British Thoracic Society Pleural Disease Guideline 2010","author":"A MacDuff","year":"2010","journal-title":"Thorax"},{"issue":"3","key":"pone.0331962.ref018","first-page":"1","article-title":"Pneumothorax In Pediatric Patients: Management Strategies To Improve Patient Outcomes","volume":"14","author":"M Harris","year":"2017","journal-title":"Pediatr Emerg Med Pract"},{"key":"pone.0331962.ref019","article-title":"Evaluation of ChatGPTs Performance in T\u00fcrkiye\u2019s First Emergency Medicine Sub-Specialization Exam","author":"H Mutlu","year":"2025","journal-title":"eajem"},{"key":"pone.0331962.ref020","doi-asserted-by":"crossref","first-page":"109545","DOI":"10.1016\/j.compbiomed.2024.109545","article-title":"Generating credible referenced medical research: A comparative study of openAI\u2019s GPT-4 and Google\u2019s gemini","volume":"185","author":"M Omar","year":"2025","journal-title":"Comput Biol Med"},{"key":"pone.0331962.ref021","article-title":"The evaluation of ChatGPT-4\u2019s capacity to provide information on febrile seizures","author":"E \u00d6zdemir Ka\u00e7er","year":"2025","journal-title":"eajem"},{"issue":"5","key":"pone.0331962.ref022","doi-asserted-by":"crossref","DOI":"10.1148\/radiol.230582","article-title":"Performance of ChatGPT on a Radiology Board-style Examination: Insights into Current Strengths and Limitations","volume":"307","author":"R Bhayana","year":"2023","journal-title":"Radiology"},{"issue":"5","key":"pone.0331962.ref023","doi-asserted-by":"crossref","first-page":"498","DOI":"10.1002\/ird3.102","article-title":"Exploring the feasibility of integrating ultra\u2010high field magnetic resonance imaging neuroimaging with multimodal artificial intelligence for clinical diagnostics","volume":"2","author":"Y Yuan","year":"2024","journal-title":"iRADIOLOGY"},{"issue":"3","key":"pone.0331962.ref024","doi-asserted-by":"crossref","first-page":"2727","DOI":"10.1097\/JS9.0000000000002234","article-title":"Multimodal large language models address clinical queries in laryngeal cancer surgery: a comparative evaluation of image interpretation across different models","volume":"111","author":"B Liang","year":"2025","journal-title":"Int J Surg"},{"issue":"1","key":"pone.0331962.ref025","article-title":"Evaluating ChatGPT\u2019s performance in classifying pertrochanteric fractures based on arbeitsgemeinschaft f\u00fcr osteosynthesefragen\/Orthopedic Trauma Association (AO\/OTA) Standards","volume":"17","author":"M Noda","year":"2025","journal-title":"Cureus"},{"issue":"3","key":"pone.0331962.ref026","doi-asserted-by":"crossref","first-page":"1111","DOI":"10.1007\/s00330-024-11115-6","article-title":"Revolution or risk?-Assessing the potential and challenges of GPT-4V in radiologic image interpretation","volume":"35","author":"MS Huppertz","year":"2025","journal-title":"Eur Radiol"},{"issue":"2","key":"pone.0331962.ref027","doi-asserted-by":"crossref","first-page":"434","DOI":"10.1002\/mrm.20321","article-title":"Transmit and receive transmission line arrays for 7 Tesla parallel imaging","volume":"53","author":"G Adriany","year":"2005","journal-title":"Magn Reson Med"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0331962","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T17:42:58Z","timestamp":1757698978000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0331962"}},"subtitle":[],"editor":[{"given":"Ahmet","family":"\u00c7a\u011flar,","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2025,9,12]]},"references-count":27,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2025,9,12]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0331962","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,9,12]]}}}