Predictors of Malignancy in Patients Presenting with Breast Lumps: A Multivariable Analysis from a Tertiary Surgical Center

Open Access

Abstract

Background: Breast cancer is the prevalent type of cancer among women globally and in Bangladesh. While the majority of breast lumps are non-cancerous, identifying malignant tumors can be difficult. Clinical and ultrasound characteristics can aid in predicting cancer, whereas histopathology offers a conclusive diagnosis. Data on independent factors predicting cancer in breast lump patients in Bangladesh is still scarce. This research sought to determine clinical and imaging factors independently linked to malignancy through multivariable analysis. Methods & Materials: An analytical observational study based in a hospital took place at Bangladesh Medical University from March 2025 to February 2026, involving 244 patients aged 18 years and older with noticeable breast lumps. Demographic, clinical, imaging, and histopathological information was gathered. Histopathological evaluation confirmed malignancy, and independent predictors were determined through multivariable logistic regression. Statistical significance was established at p<0.05. Results: Among 244 patients, 70 (28.7%) had malignant lesions, with invasive carcinoma NST being the predominant histological type. Multivariable analysis identified older age, positive family history, larger lump size, skin/nipple changes, and BI-RADS 4-5 as independent predictors of malignancy. BI-RADS 4-5 showed the strongest association (aOR 11.21, 95% CI: 4.72-26.62; p<0.001), followed by age >50 years (aOR 4.57) and lump size ≥5 cm (aOR 3.91). Conclusion: Older age, positive family history, larger breast lump size, skin/nipple changes, and BI-RADS 4-5 were independent predictors of malignancy. Integrating these clinical and imaging features may improve early identification and timely diagnosis of breast cancer among patients presenting with breast lumps.