Using GPT-4o for CAD-RADS feature extraction and categorization with free-text coronary CT Angiography reports (Preprint) DOI

Youmei Chen,

Jie Sun,

Mengshi Dong

et al.

Published: Jan. 8, 2025

BACKGROUND Despite the Coronary Artery Reporting and Data System (CAD-RADS) providing a standardized approach, radiologists continue to favor free-text reports. This preference creates significant challenges for data extraction analysis in longitudinal studies, potentially limiting large-scale research quality assessment initiatives. OBJECTIVE To evaluate ability of GPT-4o model convert real-world coronary CT angiography (CCTA) reports into structured automatically identify CAD-RADS categories P Categories. METHODS retrospective study analyzed CCTA from January 2024 July 2024. A subset 25 was used prompt engineering instruct LLMs extracting categories, Categories, presence myocardial bridges non-calcified plaques. Reports were processed using API custom Python scripts. The ground truth established by radiologist based on 2.0 guidelines. Model performance assessed accuracy, sensitivity, specificity, F1 score. Intra-rater reliability Cohen's Kappa coefficient. RESULTS Among 999 patients (median age 66 years, range 58-74; 650 males), categorization showed accuracy 0.98-1.00, sensitivity 0.95-1.00, specificity score 0.96-1.00. Categories demonstrated 0.97-1.00, 0.90-1.00, 0.91-0.99. Myocardial bridge detection achieved 0.98 calcified plaque accuracy. values all classifications exceeded 0.98. CONCLUSIONS efficiently accurately converts data, excelling classification, burden assessment,

Language: Английский

ChatGPT and radiology report: potential applications and limitations DOI
Marco Parillo,

Federica Vaccarino,

Bruno Beomonte Zobel

et al.

La radiologia medica, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 7, 2024

Language: Английский

Citations

5
Opportunities and challenges in the application of large artificial intelligence models in radiology DOI Creative Commons
Liangrui Pan,

Zhenyu Zhao,

Ying Lü

et al.

Meta-Radiology, Journal Year: 2024, Volume and Issue: 2(2), P. 100080 - 100080

Published: May 8, 2024

Influenced by ChatGPT, artificial intelligence (AI) large models have witnessed a global upsurge in model research and development. As people enjoy the convenience this AI model, more subdivided fields are gradually being proposed, especially radiology imaging field. This article first introduces development history of models, technical details, workflow, working principles multimodal video generation models. Secondly, we summarize latest progress education, report generation, applications unimodal radiology. Finally, paper also summarizes some challenges radiology, with aim better promoting rapid revolution field radiography.

Language: Английский

Citations

4
Engaging Preference Optimization Alignment in Large Language Model for Continual Radiology Report Generation: A Hybrid Approach DOI

Amaan Izhar,

Norisma Idris, Nurul Japar

et al.

Cognitive Computation, Journal Year: 2025, Volume and Issue: 17(1)

Published: Jan. 27, 2025

Language: Английский

Citations

0
Artificial Intelligence in Spine Imaging DOI
Kushal Patel,

P R Cooper,

Puneet Belani

et al.

Magnetic Resonance Imaging Clinics of North America, Journal Year: 2025, Volume and Issue: 33(2), P. 389 - 398

Published: Feb. 14, 2025

Language: Английский

Citations

0
Using GPT-4o for CAD-RADS feature extraction and categorization with free-text coronary CT Angiography reports (Preprint) DOI

Youmei Chen,

Jie Sun,

Mengshi Dong

et al.

Published: Jan. 8, 2025

BACKGROUND Despite the Coronary Artery Reporting and Data System (CAD-RADS) providing a standardized approach, radiologists continue to favor free-text reports. This preference creates significant challenges for data extraction analysis in longitudinal studies, potentially limiting large-scale research quality assessment initiatives. OBJECTIVE To evaluate ability of GPT-4o model convert real-world coronary CT angiography (CCTA) reports into structured automatically identify CAD-RADS categories P Categories. METHODS retrospective study analyzed CCTA from January 2024 July 2024. A subset 25 was used prompt engineering instruct LLMs extracting categories, Categories, presence myocardial bridges non-calcified plaques. Reports were processed using API custom Python scripts. The ground truth established by radiologist based on 2.0 guidelines. Model performance assessed accuracy, sensitivity, specificity, F1 score. Intra-rater reliability Cohen's Kappa coefficient. RESULTS Among 999 patients (median age 66 years, range 58-74; 650 males), categorization showed accuracy 0.98-1.00, sensitivity 0.95-1.00, specificity score 0.96-1.00. Categories demonstrated 0.97-1.00, 0.90-1.00, 0.91-0.99. Myocardial bridge detection achieved 0.98 calcified plaque accuracy. values all classifications exceeded 0.98. CONCLUSIONS efficiently accurately converts data, excelling classification, burden assessment,

Language: Английский

Citations

0