Background: Emerging roles of artificial intelligence support in breast imaging leads to improved radiologists’ performance Objective: To assess the diagnostic performance of the artificial intelligence decision support in the evaluation of breast masses using automated breast ultrasound (ABUS). Methods: 182 patients (415 breast masses) who received ABUS were included. Two readers, including the experienced breast radiologist (reader1) and the breast imaging fellow (reader2), separately reviewed the ABUS images and the artificial intelligence decision support according to the ACR BI-RADS 5th edition. Results: All of the 415 masses; 395 masses (95.2%) were benign and 20 masses (4.8%) were malignant. The area under receiver operating curve (AUC) of the artificial intelligence decision support was 0.74 (95%CI 0.72-0.77) with sensitivity of 100.0% and specificity of 48.6%. The integration of artificial intelligence decision support significantly increased the AUC for both readers, from 0.82 (95% CI 0.74-0.91) to 0.85 (95% CI 0.76-0.93) for reader1 (P < 0.001) and from 0.79 (95% CI 0.71-0.88) to 0.81 (95% CI 0.73-0.89) for reader2 (P < 0.001). The artificial intelligence decision support led to a 14.2% and 16.9% alteration in BI-RADS with a 22.2% and 10.7% reduction of biopsies of the benign masses for reader1 and reader2, respectively. Conclusion: The AI decision support demonstrates diagnostic performance comparable to radiologists, exhibiting high sensitivity and high negative predictive value. Integrating AI into the diagnostic workflow will potentially enhance diagnostic performance for radiologists across various experience levels and contribute to a reduction in unnecessary biopsies of benign masses.
Hasdiseve et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: