Evaluating GPT4 on Impressions Generation in Radiology Reports
2023; Radiological Society of North America; Volume: 307; Issue: 5 Linguagem: Inglês
10.1148/radiol.231259
ISSN1527-1315
AutoresZhaoyi Sun, Hanley Ong, Patrick Kennedy, Liyan Tang, Shirley Chen, Jonathan Elias, Eugene Lucas, George Shih, Yifan Peng,
Tópico(s)Radiology practices and education
ResumoHomeRadiologyVol. 307, No. 5 PreviousNext Original ResearchComputer ApplicationsEvaluating GPT-4 on Impressions Generation in Radiology ReportsZhaoyi Sun*, Hanley Ong*, Patrick Kennedy, Liyan Tang, Shirley Chen, Jonathan Elias, Eugene Lucas, George Shih, Yifan Peng Zhaoyi Sun*, Hanley Ong*, Patrick Kennedy, Liyan Tang, Shirley Chen, Jonathan Elias, Eugene Lucas, George Shih, Yifan Peng Author AffiliationsFrom the Departments of Population Health Sciences (Z.S., Y.P.), Radiology (H.O., P.K., S.C., G.S.), and Primary Care (J.E.), Weill Cornell Medicine, 425 E 61st St, Suite 301, New York, NY 10065; School of Information, The University of Texas at Austin, Austin, Tex (L.T.); and Comprehensive Weight Control Center, New York–Presbyterian/Weill Cornell Medical Center, New York, NY (E.L.).Address correspondence to Y.P. (email: [email protected]).Zhaoyi Sun*Hanley Ong*Patrick KennedyLiyan TangShirley ChenJonathan EliasEugene LucasGeorge ShihYifan Peng Published Online:Jun 27 2023https://doi.org/10.1148/radiol.231259MoreSectionsFull textPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In * Z.S. and H.O. contributed equally to this work.AbstractRadiologist-generated Impressions from radiology report Findings outperformed GPT-4-generated ones in coherence, comprehensiveness, and factual consistency, highlighting areas for future artificial intelligence improvements in radiology report generation.Download as PowerPointReferences1. Hartung MP, Bickle IC, Gaillard F, Kanne JP. How to Create a Great Radiology Report. RadioGraphics 2020;40(6):1658–1670. Link, Google Scholar2. Gundogdu B, Pamuksuz U, Chung JH, et al. Customized Impression Prediction from Radiology Reports Using BERT and LSTMs. IEEE Transactions on Artificial Intelligence. 2021; 1–1. Crossref, Google Scholar3. GPT-4. OpenAI. https://openai.com/gpt-4. Accessed May 25, 2023. Google Scholar4. Tang L, Sun Z, Idnay B, et al. Evaluating Large Language Models on Medical Evidence Summarization. medRxiv 2023.04.22.23288967. Posted April 24, 2023. Accessed May 15, 2023. Google Scholar5. Bhayana R, Krishna S, Bleakney RR. Performance of ChatGPT on a Radiology Board-style Examination: Insights into Current Strengths and Limitations. Radiology 2023;307(5):e230582. Link, Google Scholar6. Wang X, Peng Y, Lu L, Lu Z, Bagheri M, Summers RM. ChestX-Ray8: Hospital-Scale Chest X-Ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017; 3462–3471. Google ScholarArticle HistoryReceived: May 16 2023Revision requested: May 17 2023Revision received: May 31 2023Accepted: June 12 2023Published online: June 27 2023 FiguresReferencesRelatedDetailsCited ByThe Need to Re-evaluate the Role of GPT-4 in Generating Radiology ReportsPartha Pratim Ray, 1 August 2023 | Radiology, Vol. 308, No. 2Recommended Articles RSNA Education Exhibits RSNA Case Collection Vol. 307, No. 5 Supplemental MaterialMetrics Altmetric Score PDF download
Referência(s)