Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases

Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases View ORCID Profile Ryo Kurokawa , View ORCID Profile Yuji Ohizumi , View ORCID Profile Jun Kanzawa , View ORCID Profile Mariko Kurokawa , Takao Kiguchi , View ORCID Profile Wataru Gonoi , View ORCID Profile Osamu Abe doi: https://doi.org/10.1101/2024.04.11.24305622 Ryo Kurokawa 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Ryo Kurokawa Yuji Ohizumi 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Yuji Ohizumi Jun Kanzawa 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jun Kanzawa Mariko Kurokawa 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Mariko Kurokawa Takao Kiguchi 2 Department of Radiology, Ichinomiyanishi Hospital , 1 Hira Kaimei, Ichinomiya-shi, Aichi, 494-0001, Japan MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Wataru Gonoi 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Wataru Gonoi For correspondence: gonoiw{at}gmail.com Osamu Abe 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo , 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Osamu Abe Abstract Full Text Info/History Metrics Data/Code Preview PDF Abstract Backgrounds Large language artificial intelligence models have showed its diagnostic performance based solely on textual information from clinical history and imaging findings. However, the extent of their performance when utilizing radiological images and providing differential diagnoses has yet to be investigated. Purpose We employed the latest version of Claude 3, Opus, released on March 4, 2024, to investigate its diagnostic performance in answering Radiology’s Diagnosis Please quiz questions under three conditions: (1) when provided with clinical history alone; (2) when given clinical history along with imaging findings; and (3) when supplied with clinical history and key images. Furthermore, we evaluated the diagnostic performance of the model when instructed to list differential diagnoses. Materials and Methods Claude 3 Opus was tasked with listing the primary diagnosis and two differential diagnoses for 322 quiz questions from Radiology’s “Diagnosis Please” cases, which included cases 1 to 322, published from 1998 to 2023. The analyses were carried out under the following input conditions: Condition 1: Submitter-provided clinical history (text) alone Condition 2: Submitter-provided clinical history and imaging findings (text) Condition 3: Submitter-provided clinical history (text) and key images (PDF files) We applied McNemar’s tests to evaluate differences in correct response rates for primary diagnoses across Conditions 1, 2, and 3. Results The correct primary diagnoses rates were 62/322 (19.3%), 178/322 (55.3%), and 93/322 (28.8%) for Conditions 1, 2, and 3, respectively. Additionally, Claude 3 Opus accurately provided the correct answer as a differential diagnosis in up to 22/322 (6.8%) of cases. There were statistically significant differences in correct response rates for primary diagnoses between all combinations of Conditions 1, 2, and 3 (p<0.001). Conclusion Claude 3 Opus demonstrated significantly improved diagnostic performance by inputting key images in addition to clinical history. The ability to list important differential diagnoses was also confirmed. Key Results This study investigated Claude 3 Opus’s performance in Radiology Diagnosis Please Cases using clinical history, key images, and imaging findings. Key images or imaging findings inputs significantly improved correct primary diagnoses from 19.3% to 28.8% or 55.5%, respectively. By having two additional differential diagnoses presented, total correct responses improved by 3.1–6.8%. Summary statement Large language AI model Claude 3 Opus demonstrated significantly improved diagnostic accuracy by adding key images with clinical history compared with clinical history alone. Introduction Large language artificial intelligence (AI) models are designed to understand and generate human-like text based on the input they receive, and have demonstrated remarkable capabilities across various domains[ 1 ]. In the field of diagnostic imaging, Ueda et al.[ 2 ] conducted a study using Chat GPT’s GPT-4 model[ 3 ] and reported that it correctly answered 54% (170 of 313) of Radiology’s Diagnosis Please quiz questions based solely on textual information from clinical history and imaging findings. This highlights the potential use of language AI in diagnostic imaging. However, the diagnostic accuracy of the AI models using radiological images remains unclear. Moreover, the ability to provide a list of highly probable differential diagnoses and their respective accuracies, which is crucial in daily clinical practice, has not been previously investigated using AI models. Claude 3[ 4 ], a large-language AI model developed by Anthropic (California, United States), possesses the ability to read and analyze not only input text but also image data. In this study, we employed the latest version of Claude 3, Opus, released on March 4, 2024, to examine its diagnostic performance in answering Radiology’s Diagnosis Please quiz questions under three conditions: (1) when provided with clinical history alone; (2) when given clinical history along with imaging findings; and (3) when supplied with clinical history and key images. Furthermore, we evaluated the diagnostic performance of the model when instructed to list differential diagnoses. Materials and Methods An overview of this study is presented in Figure ( Figure ). We used Claude 3 Opus (Anthropic, California, United States, released on 4th Mar 2024, accessed on 15th Mar 2024, https://claude.ai/ ) to list the primary diagnosis and two differential diagnoses for 322 quiz questions from Radiology’s Diagnosis ( https://dxp.rsna.org/ ), spanning from Cases 1 to 322, published between 1998 and 2023. Analyses were conducted under the following three input data conditions: Download figure Open in new tab Figure Overview of this study Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases. • This study investigated Claude 3 Opus’s performance in Radiology Diagnosis Please Cases using clinical history, key images, and imaging findings. • Key images or imaging findings inputs significantly improved correct primary diagnoses from 19.3% to 28.8% or 55.5%, respectively. • By having two additional differential diagnoses presented, total correct responses improved by 3.1-6.8%. Condition 1: Submitter-provided clinical history (text) alone Condition 2: Submitter-provided clinical history and imaging findings (text) Condition 3: Submitter-provided clinical history (text) and key images (PDF files) Each condition was examined using separate threads. When extracting the submitter-identified imaging findings, two trainee radiologists and one board-certified diagnostic radiologist with 11 years of experience meticulously removed sentences containing answers to ensure the integrity of the analysis. The accuracy of the primary diagnosis and the two differential diagnoses generated by Claude 3 Opus were determined by consensus among three board-certified diagnostic radiologists with 8, 11, and 19 years of experience. Ethical approval was not required because this study exclusively utilized published articles. McNemar’s tests were used to assess the difference in correct response rates for the primary diagnoses under Conditions 1, 2, and 3. Two-sided P values <0.05 were considered statistically significant. Statistical analyses were performed using the R software (version 4.1.1; R Foundation for Statistical Computing, Vienna, Austria). Results The results are summarized in Table ( Table ). The primary diagnoses were correct in 62/322 (19.3%), 178/322 (55.3%), and 93/322 (28.8%) cases under conditions 1, 2, and 3, respectively. Claude 3 Opus provided the correct answer as a differential diagnosis in up to 22/322 (6.8%) cases. There were significant differences in the correct response rates for primary diagnoses between any combination of conditions 1, 2, and 3 (ps<0.001). View this table: View inline View popup Download powerpoint Table Summary of the results Discussion In this study, we evaluated the diagnostic performance of the language AI model Claude 3 Opus in answering Radiology’s Diagnosis Please quiz questions based on three types of input information: clinical history alone, clinical history accompanied by imaging findings, and clinical history coupled with key images. Ueda et al. [ 2 ] demonstrated that using the Chat GPT GPT-4 model, a large language AI model, with the input of text information from clinical history and imaging findings, Chat GPT correctly answered 54% of questions. Similarly, Li et al. [ 5 ] showed a 17% improvement in diagnostic accuracy when using the GPT-4 model compared to the GPT-3.5 model of Chat GPT when inputting text information from clinical history and imaging findings. Large language AI models continue to evolve, enabling them to read the image information provided by users. However, the diagnostic performances of AI models that use radiological images in conjunction with textual information have not yet been investigated. The results of this study demonstrated that Claude 3 Opus could present the correct diagnosis in 28.8% (93/322) of cases by inputting key images in addition to the clinical history. This shows a 9.5% increase with a statistically significant difference in accuracy compared with providing clinical history alone, potentially reflecting the imaging diagnostic performance of AI models. However, it was inferior to the accuracy achieved with clinical history and submitter-provided imaging findings (28.8% vs 55.3%; respectively, with a significant difference in accuracy of 26.5%). The 55.3% (178/322) accuracy of primary diagnoses obtained by Claude 3 Opus from clinical history plus imaging findings was nearly equivalent to the 54% shown by Ueda et al. using the GPT-4 model. Owing to real-world complexity, there may be more than one possible diagnosis in daily clinical practice, necessitating the reporting of potential differential diagnoses. In this study, the AI model presented two differential diagnoses, in addition to the primary diagnosis. It succeeded in presenting the correct answer as a differential diagnosis in up to 6.8% (22/322) of cases, resulting in an accuracy of up to 62.1 % (200/322) when the clinical history and imaging findings were input. The results of this study suggest that large language AI models, in their current state, may be more effectively utilized as a supportive tool for radiologists in formulating differential diagnoses based on accurate image interpretation and clinical history, rather than as a standalone solution for replacing radiologists in the image interpretation process. Although these AI models demonstrate a certain level of usefulness in assisting radiologists, their ability to interpret image findings solely from the images themselves and provide meaningful contributions to diagnosis still requires improvement and is still in the early stages of development. However, as the performance of large language AI models continues to advance and innovations in the provided information emerge, there is potential for further improvements in their diagnostic capabilities. To fully realize this potential, additional research and validation are necessary to explore the evolving role of AI models in the field of diagnostic imaging. Data Availability All data produced are available online at https://pubs.rsna.org/journal/radiology . Abbreviations AI artificial intelligence References 1. ↵ Eloundou T , Manning S , Mishkin P , Rock D. GPTs are GPTs: an early look at the labor market impact potential of large language models . arXiv, 2023 . Available from: http://econ.GN . Available from: http://arxiv.org/abs/2303.10130 . 2. ↵ Ueda D , Mitsuyama Y , Takita H , Horiuchi D , Walston SL , Tatekawa H , Miki Y. ChatGPT’s diagnostic performance from patient history and imaging findings on the diagnosis please quizzes . Radiology 2023 ; 308 ( 1 ): e231040 . doi: 10.1148/radiol.231040 . OpenUrl CrossRef 3. ↵ GPT-4 . Available from: https://openai.com/gpt-4 . accessed March 17, 2024 . 4. ↵ Claude \ anthropic . Available from: https://www.anthropic.com/claude . accessed March 17, 2024 . 5. ↵ Li D , Gupta K , Bhaduri M , Sathiadoss P , Bhatnagar S , Chong J. Comparing GPT-3.5 and GPT-4 accuracy and drift in radiology diagnosis please cases . Radiology 2024 ; 310 ( 1 ): e232411 . doi: 10.1148/radiol.232411 . OpenUrl CrossRef View the discussion thread. Back to top Previous Next Posted April 14, 2024. Download PDF Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases Ryo Kurokawa , Yuji Ohizumi , Jun Kanzawa , Mariko Kurokawa , Takao Kiguchi , Wataru Gonoi , Osamu Abe medRxiv 2024.04.11.24305622; doi: https://doi.org/10.1101/2024.04.11.24305622 Share This Article: Copy Citation Tools Diagnostic Performance of Claude 3 from Patient History and Key Images in Diagnosis Please Cases Ryo Kurokawa , Yuji Ohizumi , Jun Kanzawa , Mariko Kurokawa , Takao Kiguchi , Wataru Gonoi , Osamu Abe medRxiv 2024.04.11.24305622; doi: https://doi.org/10.1101/2024.04.11.24305622 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Radiology and Imaging Subject Areas All Articles Addiction Medicine (573) Allergy and Immunology (865) Anesthesia (302) Cardiovascular Medicine (4453) Dentistry and Oral Medicine (444) Dermatology (383) Emergency Medicine (609) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1515) Epidemiology (15242) Forensic Medicine (30) Gastroenterology (1131) Genetic and Genomic Medicine (6615) Geriatric Medicine (669) Health Economics (1001) Health Informatics (4552) Health Policy (1372) Health Systems and Quality Improvement (1614) Hematology (543) HIV/AIDS (1270) Infectious Diseases (except HIV/AIDS) (15929) Intensive Care and Critical Care Medicine (1106) Medical Education (624) Medical Ethics (147) Nephrology (670) Neurology (6625) Nursing (346) Nutrition (999) Obstetrics and Gynecology (1148) Occupational and Environmental Health (957) Oncology (3344) Ophthalmology (979) Orthopedics (369) Otolaryngology (421) Pain Medicine (436) Palliative Medicine (130) Pathology (665) Pediatrics (1696) Pharmacology and Therapeutics (693) Primary Care Research (714) Psychiatry and Clinical Psychology (5461) Public and Global Health (9252) Radiology and Imaging (2207) Rehabilitation Medicine and Physical Therapy (1371) Respiratory Medicine (1197) Rheumatology (597) Sexual and Reproductive Health (715) Sports Medicine (530) Surgery (714) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a02bb72828628f53',t:'MTc3OTk1NTExOQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();