Abstract 5980: Identification of pre-treatment tumor habitats for the prediction of neoadjuvant therapy response in triple negative breast cancer

Abstract

Abstract Introduction: Triple negative breast cancer (TNBC) patients exhibit varied levels of response to neoadjuvant chemotherapy (NAC), with only 27-51% of patients achieving pathological complete response (pCR). This diverse response can be attributed in-part to heterogeneity of the tumor microenvironment, affecting therapeutic delivery and efficacy. Multiparametric magnetic resonance imaging (MRI) can be used to spatially resolve intratumoral heterogeneity into distinct tumor subregions, or habitats. We investigated whether MRI-derived tumor habitats identified prior to initiation of NAC were predictive of pathological response in TNBC patients. Methods: Women with stage II/III TNBC who received a pre-treatment (baseline) breast MRI and NAC at our institution (2012-2019) were retrospectively identified. Pathological response was determined at surgery, with pCR defined as no residual tumor within the breast or lymph nodes. Both diffusion-weighted (DW) and dynamic contrast-enhanced (DCE) MRI data were collected prior to initiation of NAC. The apparent diffusion coefficient (a measure of cell density) was calculated for each voxel from DW-MRI. Signal enhancement ratio, percent enhancement, and wash-out slope were calculated for each voxel from DCE-MRI, providing measures of vascularity. Hierarchical clustering of voxel data was used to identify tumor habitats, with each subregion labeled in terms of “high” or “low” vascularity and cellularity based on mean parameter values for the subregion. Tumor composition was quantified as percent tumor volume comprised by each habitat. Differences between pCR and non-pCR patients were assessed using Wilcoxon rank sum test, with p<0.05 considered significant. Results: 46 women with TNBC were retrospectively identified (median age: 48, range 31-77 yrs), of which 14 (30%) achieved pCR. No significant differences between pCR and non-pCR patients were observed in baseline tumor volume or longest diameter (p>0.05). Clustering analysis yielded four tumor habitats: low-vascularity low-cellularity (LV-LC), low-vascularity high-cellularity (LV-HC), high-vascularity low-cellularity (HV-LC), and high-vascularity high-cellularity (HV-HC). Patients who achieved pCR had significantly higher fraction of the HV-HC habitat at baseline (p=0.02). No significant differences were observed for other habitats. Discussion & Conclusion: Our findings suggest multiparametric MRI can identify physiologically-distinct tumor habitats prior to NAC, which are predictive of response. A higher fraction of HV-HC habitat was associated with pCR, potentially suggestive of increased therapeutic delivery/sensitivity. Clinical translation of this approach would enable more specific characterizations of tumor heterogeneity and prediction of response, which could aid in personalizing regimens for optimal outcomes. Citation Format: Anum S. Kazerouni, Laura C. Kennedy, Shaveta Vinayak, Suzanne Dintzis, Habib Rahbar, Savannah C. Partridge. Identification of pre-treatment tumor habitats for the prediction of neoadjuvant therapy response in triple negative breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5980.