Breast Cancer MRI Shows High Subtype Accuracy

MULTIPARAMETRIC MRI could help distinguish between luminal and non-luminal invasive ductal carcinoma (IDC), the most common type of breast cancer, with high accuracy, according to a prospective study, offering clinicians a potential non-invasive adjunct to histopathology when planning personalised treatment.

The study found that combining specific MRI features achieved an area under the curve (AUC) of 0.92 for differentiating luminal and non-luminal tumours, with 87% sensitivity and 86% specificity. The findings suggest multiparametric MRI may provide clinically useful imaging biomarkers, although external validation is needed before routine implementation.

A Potential MRI Biomarker in Breast Cancer Assessment

Researchers evaluated whether multiparametric breast MRI could non-invasively distinguish luminal (A and B) from non-luminal (HER2-enriched and triple-negative) IDC subtypes.

The prospective single-centre study included 102 patients with newly diagnosed IDC who underwent 3T multiparametric MRI before treatment. Tumours were classified according to immunohistochemistry. MRI protocols included T1-weighted imaging, T2-weighted imaging, subtraction imaging, dynamic contrast-enhanced MRI (DCE-MRI), diffusion-weighted imaging (DWI), and MR spectroscopy (MRS). Investigators assessed morphological characteristics alongside semiquantitative and quantitative DCE-MRI parameters, apparent diffusion coefficient (ADC) values, and MRS choline peaks.

Distinct Imaging Features Linked to Non-Luminal Disease

Among the 102 patients, 48 (47.1%) had luminal tumours and 54 (52.9%) had non-luminal disease.

Luminal tumours were more likely to appear irregular with spiculated margins and heterogeneous enhancement. In contrast, non-luminal tumours more frequently demonstrated circumscribed margins, rim enhancement, perilesional oedema and axillary lymphadenopathy.

Non-luminal cancers also showed higher Ktrans values, signal enhancement ratio (SER) and maximum slope of increase, alongside lower extracellular extravascular volume fraction (Ve) and ADC values. A choline peak on MR spectroscopy was observed more frequently in the non-luminal group.

Multivariable regression identified perilesional oedema, low Ve and high SER as independent predictors of non-luminal breast cancer. When combined, these parameters achieved an AUC of 0.92, with 87% sensitivity and 86% specificity.

Adjunctive Role Requires Further Validation

The findings support multiparametric MRI as a reliable adjunct to histopathology for differentiating breast cancer subtypes and may help inform personalised treatment strategies and prognostication. However, as the study was conducted at a single centre with 102 patients, the authors noted that external validation is required before these imaging markers can be adopted in routine clinical practice.

Reference

Mohakud S et al. Evaluating the utility of qualitative and quantitative multiparametric MRI parameters in distinguishing luminal and non-luminal breast cancer subtypes. Clin Radiol. 2026;DOI:10.1016/j.crad.2026.107426.

Featured image: David A Litman on Adobe Stock