{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T17:01:46Z","timestamp":1793379706555,"version":"4.5.0"},"reference-count":16,"publisher":"Springer Science and Business Media LLC","issue":"4","license":[{"start":{"date-parts":[[2025,11,8]],"date-time":"2025-11-08T00:00:00Z","timestamp":1762560000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,11,8]],"date-time":"2025-11-08T00:00:00Z","timestamp":1762560000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"University of California, San Francisco-Stanford CERSI\/Gilead Postdoctoral Fellowship in Regulatory Science"},{"name":"National Library of Medicine of the National Institutes of Health","award":["K99LM014099"],"award-info":[{"award-number":["K99LM014099"]}]},{"DOI":"10.13039\/100000062","name":"National Institute of Diabetes and Digestive and Kidney Diseases","doi-asserted-by":"publisher","award":["K23DK139455"],"award-info":[{"award-number":["K23DK139455"]}],"id":[{"id":"10.13039\/100000062","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Dig Dis Sci"],"published-print":{"date-parts":[[2026,4]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Metabolic dysfunction-associated steatohepatitis (MASH) is a leading cause of cirrhosis. Vibration Controlled Transient Elastography (VCTE) measurements are often captured in text-based reports and not readily accessible for clinical research. Large language models (LLMs) show promise for curating information from unstructured documentation, but their efficiency for MASH and VCTE extraction are unclear.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We used a cohort of 493 patients with compensated MASH cirrhosis. We compared the abilities of GPT-4o and Claude 3.5 Sonnet for identifying the presence of MASH and extracting maximum VCTE stiffness and Controlled Attenuation Parameter (CAP) measurements from clinical documentation. We ran a cost analysis of the LLMs. As exploratory analysis, we used LASSO-Cox to associate LLM-extracted features with death or decompensation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      For identifying MASH in clinical notes, GPT-4o and Claude 3.5 achieved F1-scores of 90.5% and 80.0%. For identifying peak VCTE measurements, GPT-4o achieved 99.3% and 99.1% accuracies for stiffness and CAP, while Claude 3.5 achieved 93.3% and 94.1% accuracies. LLM extraction of one variable required\u2009~\u20092000 tokens per note, with a cost of\u2009~\u2009$0.012\/note for GPT-4o and\u2009~\u2009$0.014\/note for Claude 3.5. In LASSO-Cox regressions, VCTE stiffness (HR 1.03, 95% CI 1.01\u20131.05,\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.016) and CAP score (HR 0.99, 95% CI 0.99\u20131.00,\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.029) were statistically significant predictive variables for death or decompensation.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>LLMs can extract MASH presence and VCTE parameters from documentation with high accuracy and low cost. When incorporated into survival analyses, LLM-extracted variables are associated with important clinical outcomes. Given the growing availability of LLMs, liver diseases researchers should incorporate these methods to facilitate real-world studies.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1007\/s10620-025-09539-1","type":"journal-article","created":{"date-parts":[[2025,11,8]],"date-time":"2025-11-08T18:12:13Z","timestamp":1762625533000},"page":"1505-1511","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":4,"title":["Large Language Models can Identify the Presence of MASH and Extract VCTE Measurements from Unstructured Documentation"],"prefix":"10.1007","volume":"71","author":[{"given":"Aryana T.","family":"Far","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aryan","family":"Ayati","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jordan","family":"Guillot","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shadera","family":"Azzam","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vivek A.","family":"Rudrapatna","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jin","family":"Ge","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,11,8]]},"reference":[{"key":"9539_CR1","doi-asserted-by":"publisher","first-page":"1335","DOI":"10.1097\/HEP.0000000000000004","volume":"77","author":"ZM Younossi","year":"2023","unstructured":"Younossi ZM, Golabi P, Paik JM, Henry A, Van Dongen C, Henry L. The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review. Hepatology. 2023;77:1335\u20131347. https:\/\/doi.org\/10.1097\/HEP.0000000000000004.","journal-title":"Hepatology."},{"key":"9539_CR2","doi-asserted-by":"publisher","first-page":"348","DOI":"10.1080\/13696998.2023.2184967","volume":"26","author":"EB Tapper","year":"2023","unstructured":"Tapper EB, Bonafede M, Fishman J et al. Healthcare resource utilization and costs of care in the United States for patients with non-alcoholic steatohepatitis. J Med Econ. 2023;26:348\u2013356. https:\/\/doi.org\/10.1080\/13696998.2023.2184967.","journal-title":"J Med Econ."},{"key":"9539_CR3","doi-asserted-by":"publisher","first-page":"e000572","DOI":"10.1136\/bmjgast-2020-000572","volume":"8","author":"KL Hayward","year":"2021","unstructured":"Hayward KL, Johnson AL, Horsfall LU, Moser C, Valery PC, Powell EE. Detecting non-alcoholic fatty liver disease and risk factors in health databases: accuracy and limitations of the ICD-10-AM. BMJ Open Gastroenterol. 2021;8:e000572. https:\/\/doi.org\/10.1136\/bmjgast-2020-000572.","journal-title":"BMJ Open Gastroenterol."},{"key":"9539_CR4","doi-asserted-by":"publisher","DOI":"10.2196\/66910","volume":"27","author":"TM Seinen","year":"2025","unstructured":"Seinen TM, Kors JA, van Mulligen EM, Rijnbeek PR. Using Structured Codes and Free-Text Notes to Measure Information Complementarity in Electronic Health Records: Feasibility and Validation Study. J Med Internet Res. 2025;27:e66910. https:\/\/doi.org\/10.2196\/66910.","journal-title":"J Med Internet Res."},{"key":"9539_CR5","doi-asserted-by":"publisher","first-page":"334","DOI":"10.1016\/j.ijmedinf.2019.06.028","volume":"129","author":"TT Van Vleck","year":"2019","unstructured":"Van Vleck TT, Chan L, Coca SG et al. Augmented intelligence with natural language processing applied to electronic health records for identifying patients with non-alcoholic fatty liver disease at risk for disease progression. Int J Med Inform. 2019;129:334\u2013341. https:\/\/doi.org\/10.1016\/j.ijmedinf.2019.06.028.","journal-title":"Int J Med Inform."},{"key":"9539_CR6","doi-asserted-by":"publisher","DOI":"10.1016\/j.gastha.2025.100701","volume":"4","author":"J Guillot","year":"2025","unstructured":"Guillot J, Williams CYK, Azzam S et al. Risk prediction in patients with metabolic dysfunction\u2013associated steatohepatitis using natural language processing. Gastro Hep Advances. 2025;4:100701. https:\/\/doi.org\/10.1016\/j.gastha.2025.100701.","journal-title":"Gastro Hep Advances."},{"key":"9539_CR7","doi-asserted-by":"publisher","first-page":"1753","DOI":"10.1097\/HEP.0000000000001115","volume":"81","author":"AT Far","year":"2025","unstructured":"Far AT, Bastani A, Lee A et al. Evaluating the positive predictive value of code-based identification of cirrhosis and its complications utilizing GPT-4. Hepatology. 2025;81:1753\u20131763. https:\/\/doi.org\/10.1097\/HEP.0000000000001115.","journal-title":"Hepatology."},{"key":"9539_CR8","doi-asserted-by":"publisher","first-page":"1966","DOI":"10.1097\/HEP.0000000000000520","volume":"78","author":"ME Rinella","year":"2023","unstructured":"Rinella ME, Lazarus JV, Ratziu V et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. Hepatology. 2023;78:1966\u20131986. https:\/\/doi.org\/10.1097\/HEP.0000000000000520.","journal-title":"Hepatology."},{"key":"9539_CR9","doi-asserted-by":"publisher","first-page":"1705","DOI":"10.1016\/j.ultrasmedbio.2003.07.001","volume":"29","author":"L Sandrin","year":"2003","unstructured":"Sandrin L, Fourquet B, Hasquenoph J-M et al. Transient elastography: a new noninvasive method for assessment of hepatic fibrosis. Ultrasound Med Biol. 2003;29:1705\u20131713. https:\/\/doi.org\/10.1016\/j.ultrasmedbio.2003.07.001.","journal-title":"Ultrasound Med Biol."},{"key":"9539_CR10","doi-asserted-by":"publisher","first-page":"1825","DOI":"10.1016\/j.ultrasmedbio.2010.07.005","volume":"36","author":"M Sasso","year":"2010","unstructured":"Sasso M, Beaugrand M, de Ledinghen V et al. Controlled attenuation parameter (CAP): a novel VCTE\u2122 guided ultrasonic attenuation measurement for the evaluation of hepatic steatosis: preliminary study and validation in a cohort of patients with chronic liver disease from various causes. Ultrasound Med Biol. 2010;36:1825\u20131835. https:\/\/doi.org\/10.1016\/j.ultrasmedbio.2010.07.005.","journal-title":"Ultrasound Med Biol."},{"key":"9539_CR11","doi-asserted-by":"publisher","first-page":"7236","DOI":"10.3748\/wjg.v22.i32.7236","volume":"22","author":"I Mikolasevic","year":"2016","unstructured":"Mikolasevic I, Orlic L, Franjic N, Hauser G, Stimac D, Milic S. Transient elastography (FibroScan(\u00ae)) with controlled attenuation parameter in the assessment of liver steatosis and fibrosis in patients with nonalcoholic fatty liver disease - Where do we stand? World J Gastroenterol. 2016;22:7236\u20137251. https:\/\/doi.org\/10.3748\/wjg.v22.i32.7236.","journal-title":"World J Gastroenterol."},{"key":"9539_CR12","doi-asserted-by":"publisher","first-page":"2410","DOI":"10.1111\/ajt.14239","volume":"17","author":"K Atiemo","year":"2017","unstructured":"Atiemo K, Skaro A, Maddur H et al. Mortality Risk Factors Among Patients With Cirrhosis and a Low Model for End-Stage Liver Disease Sodium Score (\u226415): An Analysis of Liver Transplant Allocation Policy Using Aggregated Electronic Health Record Data. Am J Transplant. 2017;17:2410\u20132419. https:\/\/doi.org\/10.1111\/ajt.14239.","journal-title":"Am J Transplant."},{"key":"9539_CR13","doi-asserted-by":"publisher","first-page":"923","DOI":"10.1016\/j.jbi.2009.07.007","volume":"42","author":"A Mykowiecka","year":"2009","unstructured":"Mykowiecka A, Marciniak M, Kup\u015b\u0107 A. Rule-based information extraction from patients\u2019 clinical data. J Biomed Inform. 2009;42:923\u2013936. https:\/\/doi.org\/10.1016\/j.jbi.2009.07.007.","journal-title":"J Biomed Inform."},{"key":"9539_CR14","unstructured":"Anthropic. Introducing Claude 3.5 Sonnet. Claude 3.5 Sonnet. June 20, 2024. Accessed July 3, 2025. https:\/\/www.anthropic.com\/news\/claude-3-5-sonnet"},{"key":"9539_CR15","unstructured":"OpenAI. Hello GPT-4o. OpenAI. May 13, 2024. Accessed July 3, 2025. https:\/\/openai.com\/index\/hello-gpt-4o\/"},{"key":"9539_CR16","doi-asserted-by":"publisher","first-page":"326","DOI":"10.1053\/j.gastro.2025.02.034","volume":"169","author":"W-Y Liu","year":"2025","unstructured":"Liu W-Y, Huang S, Ji H et al. From \u201cBurnt-Out\u201d to \u201cBurning-Out\u201d: Capturing Liver Fat Loss in Patients With Advanced Metabolic Dysfunction-Associated Steatotic Liver Disease From a Dynamic Perspective. Gastroenterology. 2025;169:326-336.e4. https:\/\/doi.org\/10.1053\/j.gastro.2025.02.034.","journal-title":"Gastroenterology."}],"container-title":["Digestive Diseases and Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10620-025-09539-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10620-025-09539-1","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10620-025-09539-1.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,5]],"date-time":"2026-05-05T08:42:53Z","timestamp":1777970573000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10620-025-09539-1"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,8]]},"references-count":16,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,4]]}},"alternative-id":["9539"],"URL":"https:\/\/doi.org\/10.1007\/s10620-025-09539-1","relation":{},"ISSN":["0163-2116","1573-2568"],"issn-type":[{"value":"0163-2116","type":"print"},{"value":"1573-2568","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,8]]},"assertion":[{"value":"19 September 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 November 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"8 November 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors of this manuscript have the following potential conflicts of interest to disclose as described by DDS:\u2014Dr. Jin Ge previously received research support from Merck and Co. He previously served on an advisory board for Gilead Sciences and previously consulted for Astellas Pharmaceuticals\/Iota Biosciences and Madrigal Pharmaceuticals.\u2014Dr. Vivek Rudrapatna\u2019s research is partially funded by grants to UCSF from the following for-profit entities: Janssen, Merck, Genentech, Alnylam, Takeda, Mitsubishi Tanabe, Stryker, and Blueprint Medicines.  He is also a shareholder in ZebraMD, DataUnite, and Acucare AI.  He has received honoraria from Natera and Ironwood Pharmaceuticals.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}}]}}