Although molecular genetic approaches have become popular for tumor characterization, hematoxylin & eosin (H&E) morphology is still remarkably useful. In reality, tumor morphology reflects the sum of all molecular pathways in tumor cells, thereby providing incredible utility for predicting tumor biology, clinical behavior, nd treatment response. While the visual reading of such slides by pathologists can predict behavior, sophisticated histomorphometric analysis with computer-aided quantitation has the potential to unlock more revealing sub-visual attributes about tumors just from morphology. Some of these sub-visual features may encode for disease aggressiveness in histological (biopsy or resection) images of cancer. Additionally these sub-visual or histologic biomarkers can be correlated with disease recurrence independent of other clinical and pathologic features; and can be extracted via computerized image analysis. Recent analyses have shown that one of the major reasons for the over-diagnosis of breast cancer is the increased diagnosis of ductal carcinoma in situ (DCIS). The current rate of DCIS diagnosis is 56 per 100,000 women (NEJM 2012). The standard of care for DCIS management is surgery followed by local radiation with the addition of endocrine therapy for lesions expressing estrogen receptor. This regimen brings the recurrence rates down from 25% (in untreated patients) to around 10%. Identification of which patients may benefit from treatment has been difficult because of the low incidence of adverse events necessitating long term followup. Molecular studies recently been performed by the Badve group (in collaboration with Genomic Health, Inc) on the E-5194 clinical trial, led to the development and commercialization of a molecular assay, the DCIS Score, which predicts the likelihood of development of ipsilateral DCIS and/or invasive cancer. However, the DCIS Score does not comprehensively account for disease heterogeneity since the assay was developed in a cohort considered low risk by clinicopathological characteristics; the performance of the assay in the real-world situation is not known. The focus of this project is t optimize and evaluate a multistain computerized histomorphometric and histochemical image-based predictor (msCHIP) to identify which DCISs are likely to be clinically aggressive and hence result in an ipsilateral breast event (IBE). msCHIP employs digitized H&E and immunohistochemistry stained (Ki67, CD10 measuring cellular proliferation, vascularity) tissue sections, from which a series of features describing spatial distribution, morphology, texture and arrangement of tumor and stromal cell nuclei, and proliferative index will be extracted via advanced computer vision tools. Thus histologic biomarkers for more and less aggressive DCISs will be identified. The successful validation of msCHIP on a large cohort of 300 cases could pave the way for rapid adoption of msCHIP as an oncological decision support tool, providing critical information for more informed treatment decisions.
In this project we propose to develop and evaluate a multi-stain computerized histomorphometric and histochemical image-based predictor (msCHIP) to identify which ductal carcinomas in situ (DCIS) are likely to be clinically aggressive and hence result in an ipsilateral breast event (IBE). The successful validation of msCHIP could pave the way for its adoption as a decision support tool, allowing for the identification of patients for whom therapies could be 'de-escalated' or specifically targeted in order to maintain favorable survival rates while minimizing the substantial (and long term) treatment-related morbidity and expense from surgery, radiation, and chemotherapy.
Showing the most recent 10 out of 52 publications
