P2-08-03: Quantitative MRI for Noninvasive Prediction of Prognostic Markers in Breast Cancer.

Abstract

Abstract Background: Magnetic resonance imaging (MRI) is a valuable tool for assessing extent of breast cancer and monitoring treatment response. Quantitative measures by diffusion-weighted MRI (DWI) and dynamic contrast-enhanced (DCE) MRI reflect tumor cellularity and vascularity. Tumor grade and some histopathological markers, such as ER, PR, HER-2, Ki67 and P53, are prognostic factors that can also be associated with tumor cellularity and vascularity. DWI and DCE measures may therefore provide a noninvasive means for predicting disease prognosis and stratifying patients to appropriate therapies. The purpose of this study was to investigate the correlation between quantitative MRI features and prognostic pathological factors in patients with invasive breast cancer. Methods: This IRB-approved retrospective study included patients with biopsy-proven invasive cancer who underwent 1.5T breast MRI (including DCE and DWI) from October 2005 to May 2006 prior to treatment. Pathology data was obtained from pre-treatment biopsy and intrinsic subtype classification was approximated by standard immunohistochemistry characteristics. After excluding cases with missing MRI or pathology data, the final study cohort included 41 invasive cancers (36 ductal and 5 lobular carcinomas) in 36 patients. MRI measures included lesion DCE kinetic features: peak initial enhancement (PE), percent rapid enhancement (RE), and percent washout (WO), and DWI normalized apparent diffusion coefficient values (nADC). Associations between imaging features and pathology markers, cancer grades and intrinsic subtypes were evaluated by Mann-Whitney U test and multivariate logistic regression. Results: Results of univariate comparisons are summarized in Table 1. One or more DCE-MRI kinetic parameters were significantly predictive (p<0.05) of each of the histopathological markers with the exception of ER, which was marginally associated with WO (p=0.05). Each of the DCE kinetics parameters significantly discriminated grade III tumors from grades I and II and luminal A from luminal B and basal-like intrinsic subtypes. In multivariate regression, both PE and WO were significant independent predictors of tumor grade (p=0.0094, p=0.0005, respectively). WO and nADC were significant independent predictors of PR status (p=0.0054, p=0.0027), while PE was the only significant independent predictor of both Ki67 (p=0.014) and intrinsic subtype (p=0.015). Conclusion: This preliminary study suggests that quantitative MRI measures are associated with prognostic tumor markers and may provide valuable noninvasive characterization of tumor biology. Larger prospective studies are needed to validate our findings. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-08-03.