{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T11:34:48Z","timestamp":1780659288727,"version":"3.54.1"},"posted":{"date-parts":[[2023,8,28]]},"group-title":"Ophthalmology","reference-count":28,"publisher":"openRxiv","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2023,8,28]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Introduction<\/jats:title>\n                  <jats:p>Assessing the capabilities of ChatGPT-4.0 and ChatGPT-3.5 for diagnosing corneal eye diseases based on case reports and compare with human experts.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We randomly selected 20 cases of corneal diseases including corneal infections, dystrophies, degenerations, and injuries from a publicly accessible online database from the University of Iowa. We then input the text of each case description into ChatGPT-4.0 and ChatGPT3.5 and asked for a provisional diagnosis. We finally evaluated the responses based on the correct diagnoses then compared with the diagnoses of three cornea specialists (Human experts) and evaluated interobserver agreements.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>The provisional diagnosis accuracy based on ChatGPT-4.0 was 85% (17 correct out of 20 cases) while the accuracy of ChatGPT-3.5 was 60% (12 correct cases out of 20). The accuracy of three cornea specialists were 100% (20 cases), 90% (18 cases), and 90% (18 cases), respectively. The interobserver agreement between ChatGPT-4.0 and ChatGPT-3.5 was 65% (13 cases) while the interobserver agreement between ChatGPT-4.0 and three cornea specialists were 85% (17 cases), 80% (16 cases), and 75% (15 cases), respectively. However, the interobserver agreement between ChatGPT-3.5 and each of three cornea specialists was 60% (12 cases).<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>The accuracy of ChatGPT-4.0 in diagnosing patients with various corneal conditions was markedly improved than ChatGPT-3.5 and promising for potential clinical integration.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Key summary points<\/jats:title>\n                  <jats:list list-type=\"simple\">\n                    <jats:list-item>\n                      <jats:label>-<\/jats:label>\n                      <jats:p>The aim of this work was to evaluate the performance of ChatGPT-4 and ChatGPT-3.5 for providing the provisional diagnosis of different corneal eye diseases based on case descriptions and compared them with three cornea specialists.<\/jats:p>\n                    <\/jats:list-item>\n                    <jats:list-item>\n                      <jats:label>-<\/jats:label>\n                      <jats:p>The accuracy of ChatGPT-4.0 in diagnosing patients with various corneal conditions was significantly better than ChatGPT-3.5 based on the specific cases.<\/jats:p>\n                    <\/jats:list-item>\n                    <jats:list-item>\n                      <jats:label>-<\/jats:label>\n                      <jats:p>The interobserver agreement between ChatGPT-4.0 and ChatGPT-3.5 was 65% while the interobserver agreement between ChatGPT-4.0 and three cornea specialists were 85%, 80%, and 75%, respectively.<\/jats:p>\n                    <\/jats:list-item>\n                  <\/jats:list>\n                <\/jats:sec>","DOI":"10.1101\/2023.08.25.23294635","type":"posted-content","created":{"date-parts":[[2023,8,28]],"date-time":"2023-08-28T17:30:10Z","timestamp":1693243810000},"source":"Crossref","is-referenced-by-count":22,"title":["Performance of ChatGPT in Diagnosis of Corneal Eye Diseases"],"prefix":"10.64898","author":[{"given":"Mohammad","family":"Delsoz","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yeganeh","family":"Madadi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wuqaas M","family":"Munir","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Brendan","family":"Tamm","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shiva","family":"Mehravaran","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mohammad","family":"Soleimani","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ali","family":"Djalilian","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Siamak","family":"Yousefi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"issue":"1","key":"2023083109450871000_2023.08.25.23294635v1.1","first-page":"13","article-title":"Corneal Innervation and Sensation: The Eye and Beyond","volume":"91","year":"2018","journal-title":"Yale J Biol Med"},{"key":"2023083109450871000_2023.08.25.23294635v1.2","unstructured":"Clinic C. doi:https:\/\/my.clevelandclinic.org\/health\/diseases\/8586-corneal-disease"},{"key":"2023083109450871000_2023.08.25.23294635v1.3","doi-asserted-by":"publisher","DOI":"10.1016\/j.ophtha.2022.07.012"},{"key":"2023083109450871000_2023.08.25.23294635v1.4","doi-asserted-by":"publisher","DOI":"10.3352\/jeehp.2019.16.25"},{"key":"2023083109450871000_2023.08.25.23294635v1.5","doi-asserted-by":"publisher","DOI":"10.1007\/s11596-021-2474-3"},{"key":"2023083109450871000_2023.08.25.23294635v1.6","doi-asserted-by":"publisher","DOI":"10.1136\/bjophthalmol-2019-315651"},{"key":"2023083109450871000_2023.08.25.23294635v1.7","doi-asserted-by":"publisher","DOI":"10.1080\/08820538.2022.2139625"},{"key":"2023083109450871000_2023.08.25.23294635v1.8","doi-asserted-by":"publisher","DOI":"10.18502\/jovr.v18i1.12730"},{"key":"2023083109450871000_2023.08.25.23294635v1.9","doi-asserted-by":"publisher","DOI":"10.3390\/diagnostics13101689"},{"key":"2023083109450871000_2023.08.25.23294635v1.10","doi-asserted-by":"publisher","DOI":"10.1016\/j.jtos.2023.01.005"},{"key":"2023083109450871000_2023.08.25.23294635v1.11","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0205998"},{"key":"2023083109450871000_2023.08.25.23294635v1.12","doi-asserted-by":"publisher","DOI":"10.1136\/bmjopen-2019-031313"},{"key":"2023083109450871000_2023.08.25.23294635v1.13","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2016.17216"},{"key":"2023083109450871000_2023.08.25.23294635v1.14","doi-asserted-by":"publisher","DOI":"10.1167\/iovs.16-19964"},{"key":"2023083109450871000_2023.08.25.23294635v1.15","doi-asserted-by":"publisher","DOI":"10.1016\/j.ophtha.2020.03.008"},{"key":"2023083109450871000_2023.08.25.23294635v1.16","doi-asserted-by":"publisher","DOI":"10.1016\/j.ogla.2020.04.012"},{"key":"2023083109450871000_2023.08.25.23294635v1.17","doi-asserted-by":"publisher","DOI":"10.1136\/bjophthalmol-2022-321141"},{"key":"2023083109450871000_2023.08.25.23294635v1.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06291-2"},{"key":"2023083109450871000_2023.08.25.23294635v1.19","unstructured":"Ramponi M. How ChatGPT actually works. AssemblyAI. Dec 23, 2022;"},{"key":"2023083109450871000_2023.08.25.23294635v1.20","doi-asserted-by":"publisher","DOI":"10.1016\/j.xops.2023.100324"},{"key":"2023083109450871000_2023.08.25.23294635v1.21","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000198"},{"key":"2023083109450871000_2023.08.25.23294635v1.22","doi-asserted-by":"publisher","DOI":"10.3389\/frai.2023.1169595"},{"key":"2023083109450871000_2023.08.25.23294635v1.23","doi-asserted-by":"publisher","DOI":"10.1016\/j.jfop.2023.100005"},{"key":"2023083109450871000_2023.08.25.23294635v1.24","unstructured":"Ortiz S. What is ChatGPT and why does it matter? ZDNET Tech. Aug 2023;"},{"key":"2023083109450871000_2023.08.25.23294635v1.25","unstructured":"Lubbad M. GPT-4 Parameters: Unlimited guide NLP\u2019s Game-Changer. March 2023;"},{"key":"2023083109450871000_2023.08.25.23294635v1.26","doi-asserted-by":"publisher","DOI":"10.1016\/j.jfop.2023.100005"},{"key":"2023083109450871000_2023.08.25.23294635v1.27","doi-asserted-by":"publisher","DOI":"10.3390\/ijerph20054541"},{"issue":"12","key":"2023083109450871000_2023.08.25.23294635v1.28","first-page":"1","article-title":"Survey of hallucination in natural language generation","volume":"55","year":"2023","journal-title":"ACM Computing Surveys"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2023.08.25.23294635","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T19:39:49Z","timestamp":1768505989000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2023.08.25.23294635"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,8,28]]},"references-count":28,"URL":"https:\/\/doi.org\/10.1101\/2023.08.25.23294635","relation":{"is-preprint-of":[{"id-type":"doi","id":"10.1097\/ICO.0000000000003492","asserted-by":"subject"}]},"subject":[],"published":{"date-parts":[[2023,8,28]]},"subtype":"preprint"}}