Computational Histomorphometric Predictor of Pathologic Response and Disease Recurrence for Triple-Negative Breast Cancer
Abstract
Problem Statement: There is an increased need for predictive and prognostic assays to help distinguish between more and less aggressive phenotypes of cancer due to (a) dramatic increase in cancer incidence and (b) improvements in early diagnosis. In 2018, approximately 266,000 new cases of invasive breast cancer (BC) will be diagnosed in the US. About 15%-20% of breast cancers are found to be triple-negative (TN). Triple-negative breast cancer (TNBC) is characterized by the absence of estrogen receptor (ER) or progesterone receptor (PR) expression, and the lack of human epidermal growth factor receptor 2 (HER2) gene amplification. TNBC exhibits a distinctly aggressive nature, with higher rates of recurrence and shorter overall survival times compared with the other breast cancer subtypes. Most TNBC patients are treated with neoadjuvant chemotherapy (NAC) followed by surgery. The pathologic complete response (pCR) is an important endpoint for NAC since patients who attain this status after surgery have improved survival. Unfortunately, a certain number of patients (60%-70%) will not have pCR following NAC for TNBC. More importantly, even though for the patients who have pCR, they may still experience of disease recurrence. However, there are no companion diagnostic tools that are able to predict pCR and recurrence. Identifying these patients early could allow for identifying candidates who might benefit from escalation of NAC or potentially be candidates for adjuvant therapy. Finally, increasing evidence appears to suggest that African-Americans (AA) have a higher likelihood of being diagnosed with TNBC and have worse clinical outcomes compared to CA. Even though molecular link between environmental exposures, disparities, and TNBC in AA and CA women is a complex undertaking, it is worthwhile to analyze and discover if there exist significant difference between AA and CA in terms of histology morphology. Rationale: Interestingly, for TNBC, high numbers of tumor infiltrating lymphocytes (TILs) have been shown as a prognostic marker, as well as predicting pCR, in any breast cancer subtype treated with NAC. Digitization of histological samples facilitates a quantitative approach towards evaluating disease progression and predicting outcome. Principal Investigator Dr. Lu, his mentor Dr. Madabhushi, and their team have been developing advanced computerized digital pathology image analysis methods, enabling a detailed morphologic interrogation of the tumor landscape from a tissue slide. In reality, tumor morphology reflects the sum of all temporal genetic and epigenetic changes and alterations in tumor cells, thereby providing incredible utility for predicting tumor biology and clinical behavior. The computer extracted morphologic tumor and stroma features might therefore enable better assessment of patient outcome than is currently possible by gene-expression based assays alone. Objective: In this work, we are looking to discover the morphology feature differences between AA and CA by analyzing H&E tissue samples of TNBC, and construct population-specific Quantitative Histomorphometric based Predictor (QuHbictor) to (1) predict the pCR following NAC, and identify candidates who might benefit from escalation of NAC or potentially be candidates for adjuvant therapy; (2) predict disease recurrence of TNBC who do have pCR, and identify candidates who might benefit from adjuvant therapy. QuHbictor will employ advanced image analysis and machine learning techniques for comprehensive characterization of TNBC tumor cell morphology, TIL morphology, the interaction between TILs and tumor cell, and intra-tumor heterogeneity from digitized images of H&E stained TNBC specimens. Impact: The proposed image based predictor would be a low-cost, tissue non-destructive assay that just uses the image of standard H&E slide from retrospectively treated patients. To date, there is no molecular assay for risk stratification o
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 19, 2019
- Source ID
- W81XWH1910668
Entities
People
- Cheng Lu
Organizations
- Case Western Reserve University
- United States Army