Document Type : Original Article
Authors
1
Department of Radiodiagnosis, RNT medical college, Udaipur, Rajasthan, India
2
Independent researcher, Jaipur, Rajasthan, India
3
Department of General Medicine, Government Medical College, Kota, Rajasthan, India
4
Independent researcher, Jhalawar Medical College, Jhalawar, Rajasthan, India
10.30485/ijsrdms.2026.593451.1711
Abstract
Background and aim: Breast cancer comprises distinct molecular subtypes, Luminal A, Luminal B, HER2-enriched, and Triple-Negative Breast Cancer (TNBC), with varied biological behaviour and treatment response. This study aimed to correlate multimodality imaging features with molecular subtypes to improve non-invasive diagnostic accuracy.
Material and methods: This cross-sectional observational study included 50 female patients with breast lumps. All underwent mammography, ultrasonography (USG), and magnetic resonance imaging (MRI), followed by histopathological and immunohistochemical classification. Imaging features including lesion shape, margins, calcifications, posterior acoustic characteristics, and enhancement patterns were analyzed.
Results: The mean age was 51.56 ± 10.3 years, with the highest incidence in the 41–50-year group (34%). Luminal A was the most common subtype (30%), followed by Luminal B (28%). Mammographic calcifications were most frequent in HER2-enriched tumors (77.8%) and associated with this subtype (OR: 3.33, 95% CI: 0.62–18.0, AUC: 0.63, sensitivity: 77.8%, specificity: 48.7%, p=0.162). Posterior acoustic enhancement strongly predicted TNBC (OR: 20.9, 95% CI: 10.5–415, AUC: 0.88, sensitivity: 75.0%, specificity: 100.0%, p<0.001). Circumscribed margins also predicted TNBC (OR: 18.75, 95% CI: 3.46–101.6, AUC: 0.86, sensitivity: 83.3%, specificity: 89.5%, p<0.001), while posterior acoustic shadowing was associated with Luminal A (OR: 5.33, 95% CI: 1.27–22.32, AUC: 0.69, sensitivity: 80.0%, specificity: 57.1%, p=0.035).
Conclusion: Distinct multimodality imaging features were associated with specific breast cancer molecular subtypes and may support non-invasive preoperative subtype prediction and personalized treatment planning. However, the single-centre design and small sample require validation in larger prospective multicentre studies.
Keywords
Subjects
