-
1
Ductal carcinoma in situ (DCIS) represents about 25% of new breast cancer cases and has a 20% to 50% risk of upstaging to invasive cancer.
-
2
Mammography, the primary imaging method for DCIS, has limitations in sensitivity, particularly in dense breasts and low-grade lesions.
-
3
Artificial intelligence (AI) shows promise in enhancing DCIS detection, classification, and risk stratification, with AUCs between 0.70 and 0.97.
-
4
AI models outperformed traditional 2D approaches, demonstrating improved sensitivity of 80–96% and specificity up to 93% in DCIS assessment.
-
5
Limitations of the reviewed studies include retrospective designs, small DCIS-specific datasets, and concerns regarding generalizability and external validation.
