The study aims to evaluate ChatGPT ability to interpret uroflowmetry studies and formulate appropriate management plans, comparing its performance to that of expert humans.

We included patients aged 18-85 who had undergone a uroflowmetry study between March and June 2024. A standardized template, which included patient demographics, presenting symptoms, and a printout of the uroflowmetry study, was inputted to ChatGPT (version 4o) and was prompted to analyze the uroflow shape pattern, readings of the void volume, Qmax, flow time, average flow (Qavg), time to maximum flow, and the PVR obtained using u/s ultrasound at the end of the uroflowmetry study. The uroflowmetry study, clinical impression and management plans were compared with those of 2 expert urologists to assess the concordance. Institutional board review approval was obtained.

Data were collected from 100 patients, with the majority being female 60% and a mean age of 38 years. ChatGPT demonstrated a concordance rate of 94.3% for the uroflowmetry graph’s shape pattern. The voided volume, Qmax, flow time, Qavg, and time to maximum flow showed a concordance rate of 91.2%. The concordance rate for the clinical impression was 88.5%, and for management plans was 85.8%. Overall concordance including shape, parameters, impression and management was 90.85%.

ChatGPT showed excellent accuracy in interpreting uroflowmetry graphs and key parameters, closely matching expert urologists' assessments in the majority of cases. However, its performance was slightly lower in providing clinical impressions and formulating management plans, indicating room for improvement. These findings suggest that AI tools like ChatGPT have significant potential to support urologists in interpreting uroflowmetry studies and developing clinical decisions, and further evaluation of their role in clinical practice is encouraged.

ChatGPT achieved high concordance rates when compared to expert interpretations of uroflowmetry studies, clinical impressions, and management plans. These findings highlight the potential of AI software to support clinical decision-making in urology and warrant further research to explore its broader applications and integration into clinical workflows.

. Xu Y, Peng C, Wei Z, Liao F, Xiao M. Can Chat-GPT be a substitute for urological resident physicians in diagnosing diseases? A preliminary conclusion from an exploratory investigation. World J Urol. 2023 Sep;41(9):2569-71. doi: 10.1007/s00345-023-04539-0. PMID: 37505265.
Davis R, Eppler M, Ayo-Ajibola O, Loh-Doyle JC, Nabhani J, Samplaski M, Gill I, Cacciamani GE. Evaluating the effectiveness of artificial intelligence-powered large language models application in disseminating appropriate and readable health information in urology. J Urol. 2023 Oct;210(4):688-94. doi: 10.1097/JU.0000000000003615. PMID: 37428117.
. zczesniewski JJ, Tellez Fouz C, Ramos Alba A, Diaz Goizueta FJ, García Tello A, Llanes González L. ChatGPT and most frequent urological diseases: analyzing the quality of information and potential risks for patients. World J Urol. 2023 Nov;41(11):3149-53. doi: 10.1007/s00345-023-04563-0. PMID: 37632558.