Congress:
ECR25
Poster Number:
C-26740
Type:
Poster: EPOS Radiologist (scientific)
DOI:
10.26044/ecr2025/C-26740
Authorblock:
J. M. D. Fonseca1, S. Kolenda Zloić2, C. I. Ayogu3, K. Marole4, S. V. Moreira5, G. Capello Ingold6, M. H. Yoshikawa7, E. Finnegan8, M. A. Soato Ratti9; 1Gainesville, FL/US, 2Zagreb/HR, 3Liverpool/UK, 4Saint George/GD, 5Itaperuna RJ/BR, 6Pilar/AR, 7Boston/US, 8Dublin/IE, 9São Paulo/BR
Disclosures:
João Martins Da Fonseca:
Nothing to disclose
Sanda Kolenda Zloić:
Nothing to disclose
Chukwudi Isaac Ayogu:
Nothing to disclose
Karabo Marole:
Nothing to disclose
Sarah Verdan Moreira:
Nothing to disclose
Gianluca Capello Ingold:
Nothing to disclose
Marcia Harumy Yoshikawa:
Nothing to disclose
Emma Finnegan:
Nothing to disclose
Marco Aurélio Soato Ratti:
Nothing to disclose
Keywords:
Abdomen, Artificial Intelligence, Emergency, Ultrasound, Computer Applications-Detection, diagnosis, Trauma
In this systematic review and meta-analysis of six studies and over 2,000 patients, we evaluated the diagnostic accuracy of AI models for the detection of abdominal free fluid using ultrasonography in emergency cases. We found a pooled sensitivity of 0.916 (95% CI: 0.784-0.970), pooled specificity of 0.941 (95% CI: 0.878-0.972) and an AUC of 0.965 (95% CI: 0.906 - 0.979) across the included studies. Our study suggests that AI models provide reliable diagnostic accuracy for detecting abdominal free fluid, though further studies are needed to address specific subgroups.