09 Radiologists’ and Ultrasound Artificial Intelligence Decision-Support Assessment of Benign and Malignant Cystic Breast Lesions
Publish Date: 2026-07-13 18:05:00
Source Domain: www.cancernetwork.com
-
Study Objective: The evaluation aimed to compare the diagnostic performance of an ultrasound AI decision-support system (KOIOS DS™) with that of breast radiologists for classifying cystic breast lesions.
-
Study Population: The retrospective study included 77 breast ultrasound cases conducted between 2015 and 2025 at an academic institution, consisting of 50 biopsy-proven cystic carcinomas and 27 biopsy-proven benign cystic lesions.
-
Inclusion Criteria: Patients aged 18 years or older with cystic components in their breast lesions were included; purely solid lesions were excluded.
-
Methods: Each ultrasound lesion was interpreted by a fellowship-trained breast radiologist and analyzed using the KOIOS DS™ AI tool. The AI provided classifications of benign, probably benign, suspicious, or probably malignant. Final diagnosis was confirmed by pathology, which served as the gold standard for accuracy assessment.
-
Statistical Analysis: The study utilized 2-tailed tests and ANOVA for its statistical analysis, and diagnostic accuracy measures were calculated to evaluate the performance of both the AI system and the radiologists.
