Performance of artificial intelligence-assisted ultrasound elastography in classifying benign and malignant breast tumors: a systematic review and meta-analysis.
Authors
Affiliations (8)
Affiliations (8)
- School of Medical Imaging, North Sichuan Medical College, Nanchong, 637000, China.
- School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, 610054, China.
- Institute of Intelligent Chinese Medicine, Chongqing University of Chinese Medicine, Chongqing, 402760, China.
- Department of Cardiovascular Ultrasound & Noninvasive Cardiology, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 610072, China.
- Ultrasound Medicine and Computational Cardiology Key Laboratory of Sichuan Province, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 610072, China.
- Department of Cardiovascular Ultrasound & Noninvasive Cardiology, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 610072, China. [email protected].
- Ultrasound Medicine and Computational Cardiology Key Laboratory of Sichuan Province, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 610072, China. [email protected].
- Department of Ultrasound, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, 610072, China. [email protected].
Abstract
Precise benign and malignant breast tumors classification is essential for effective treatment planning and outcome prognostication. Medical imaging's capability to classify breast tumors has been greatly improved by the accelerated advancement of artificial intelligence (AI). This research presents a comprehensive evaluation of the efficiency of AI-assisted ultrasound elastography (UE) specifically applied to classify benign and malignant breast tumors for the first time. We conducted extensive literature search in PubMed, Embase, IEEE, Web of Science, Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang database, and China Biology Medicine disc (CBM) to identify relevant studies that applied or developed AI algorithms for classifying benign and malignant breast masses employing UE. We used bivariate mixed-effects model for statistical analysis, obtaining binary diagnostic accuracy data to generate pooled estimates (e.g., sensitivity and specificity). The Quality Assessment of Diagnostic Accuracy Studies-AI (QUADAS-AI) tool was applied to assess the methodological quality of the included research. Sensitivity analysis was conducted to verify the robustness of the findings, and Deeks' funnel plot was employed to examine potential publication bias. Meta-regression analysis was used to investigate the sources of heterogeneity. Clinical applicability was evaluated by Fagan nomogram. The meta-analysis comprised sixteen relevant studies. Summary estimates indicated high diagnostic accuracy: the pooled sensitivity was 0.90 (95% CI: 0.85-0.94), the pooled specificity was 0.88 (0.81-0.93), the positive likelihood ratio (PLR) was 7.5 (4.7-11.9), and the negative likelihood ratio (NLR) was 0.11 (0.07-0.18). The diagnostic odds ratio (DOR) was 67 (33-137), and the area under the summary receiver operating characteristic curve (AUC) was 0.95 (0.93-0.97). AI-assisted UE demonstrates outstanding performance in benign and malignant breast tumors classification. This study was registered with PROSPERO (CRD42024590031).