Bridging MRI and Oncotype DX: Multimodal Keys to Breast Cancer Prognosis

Background and Purpose: Breast magnetic resonance imaging (MRI) excels in preoperative staging but its utility in predicting Oncotype DX Recurrence Score (RS), a validated genomic assay guiding adjuvant chemotherapy is underexplored. This study rigorously assesses whether MRI-derived, clinicopathologic, and demographic features reliably predict RS, enabling non-invasive risk stratification in early-stage breast cancer. Methods: In this retrospective analysis of 261 patients with preoperative MRI and Oncotype DX testing, we evaluated anatomical (e.g., tumor size, multicentricity), functional (e.g., peak signal enhancement ratio (SER)), pathologic (e.g., Nottingham grade), and clinical parameters (e.g., age, nodal status). Univariable and multivariable linear regression models determined correlations with continuous RS (0–100), with p<0.05 denoting significance. Model performance was quantified by adjusted R². Results: Conventional MRI features (tumor size/volume, density) showed negligible RS correlation. Axillary lymphadenopathy and tumor peak SER emerged as significant imaging predictors. Pathologic factors dominated including T-stage, mitotic index, and triple-negative subtype .Age exhibited U-shaped nonlinearity. A multivariable model integrating MRI, pathology, and clinical data achieved robust RS prediction (adjusted R²=0.375, p<0.001), outperforming imaging alone. Conclusion: Multimodal MRI-pathologic-clinical integration powerfully predicts Oncotype DX RS, with nodal involvement and SER as pivotal imaging markers. This approach refines personalized therapy, potentially reducing genomic testing reliance in resource-limited settings.