Association of BI-RADS Category 4 Subcategories and Breast Imaging Findings with Pathological Outcomes among Patients at Phetchabun Hospital, Phetchabun Province, Thailand

Abstract

Breast cancer remains an important public health problem, and breast imaging plays a key role in evaluating lesions suspected of malignancy, particularly BI-RADS category 4 lesions, which carry varying probabilities of malignancy according to their subcategories. This retrospective analytical study aimed to determine the proportion of malignant pathological outcomes, examine the association between BI-RADS category 4 subcategories and breast imaging findings with pathological outcomes, and assess the positive predictive value of BI-RADS 4A, 4B, and 4C among patients at Phetchabun Hospital. Data were collected from medical records, breast imaging reports, pathology reports, and the hospital medical database between 2017 and 2024. The study included 128 patients who underwent mammography and breast ultrasound, were classified as BI-RADS category 4, and had pathological confirmation. Data were analyzed using descriptive statistics, Chi-square test or Fisher’s exact test, Mann–Whitney U test, Kappa statistic, and odds ratios with 95% confidence intervals. The mean age of patients was 52.5 ± 10.4 years. Malignant pathological outcomes were found in 51.6% of patients, with invasive ductal carcinoma being the most common subtype. BI-RADS category 4 subcategories were significantly associated with pathological outcomes (p <0.001). The positive predictive values of BI-RADS 4A, 4B, and 4C were 30.4%, 37.0%, and 76.4%, respectively. The agreement between BI-RADS subcategories and pathological outcomes was fair, with a Kappa value of 0.302. Breast ultrasound findings significantly associated with malignancy included irregular shape, non-parallel orientation, and not circumscribed margin, whereas most mammographic findings were not significantly associated with malignancy. These findings support the use of BI-RADS category 4 subcategories together with key breast ultrasound features to assist in risk assessment and planning for pathological confirmation in patients with suspicious breast lesions.