Abstract

Breast cancer stands as one of the leading cancer types which causes numerous deaths throughout the world. Mammography serves as a standard tool for detecting breast cancer at its early stages of early development and diagnosis. The various molecular subtypes of breast cancer show different reactions to treatments and have different outcomes for patients. The research aims to establish the ability of visual lesion features in digital mammography images to identify breast cancer molecular subtypes. The research involved 194 patients who received breast cancer diagnoses through biopsy and their mammography results showed BI-RADS 4A, 4B, 4C or 5 lesions with breast density C and D. ER (estrogen receptor), PR (progesterone receptor), HER2 (human epidermal growth factor receptor 2), and Ki 67 values were obtained from pathology reports. The researchers studied breast cancer lesion characteristics through right/left breast positioning and upper/outer-lower/inner quadrant placement and they evaluated margin characteristics and lesion shape and dimensions and tumor size and microcalcification types and distribution patterns. The study found that microcalcification presence linked to both ER-positive (p=0.012) and PR-positive (p=0.038) breast cancer cases. The research found that younger patients with triple-negative breast cancer had tumors that exceeded 20 mm in size and round shapes. However, shape-based findings should be interpreted cautiously due to the limited number of round lesions. The presence of indistinct margins in breast cancer images were strongly associated with triple-negative breast cancer. Visual mammographic features are generally insufficient for accurate prediction of molecular subtypes, except for certain associations observed in triple-negative breast cancer. Future studies integrating radiomic analysis, artificial intelligence-based approaches, and larger cohorts are warranted.

Key words: Breast cancer subtypes, mammography, molecular classification