Quantitative histopathology for cancer prognosis using quantitative phase imaging on stained tissues

点击翻译按钮获取中文摘要

Project Summary About 1 in 8 U.S. women will develop invasive breast cancer over the course of her lifetime. Early diagnosis and prognosis are key to improving health outcomes. Prognostic markers in tissue biopsies help clinicians make treatment decisions and refine the patient risk stratification. New research expands the current prognostic markers to better deliver personalized treatment regimens. However, the variability of preanalytical factors (biopsy collection, processing and storage) can have a significant impact on biomarkers evaluation which can result in potentially serious consequences in terms of patient care. There is an identified need for developing clinically relevant biomarkers that are invariant to biospecimen preparation. This project proposes a technical solution to extracting intrinsic tissue morphology information, unaffected by variability in tissue staining, slice thickness, or sectioning errors. Spatial Light Interference Microscopy (SLIM) was shown to provide prognostic markers derived from tumor microenvironment using nanoscale organization of the non-malignant tissue adjacent to cancer cells, i.e., the stromal response to cancer. Preliminary results indicate that SLIM can distinguish between pairs of “matched” patients (good vs. bad outcome) and has the capability to eliminate false positives and help the clinician assign the appropriate treatment. For this project, we will validate color SLIM (cSLIM) capabilities as a prognostic tool for existing, stained histopathology slides. cSLIM will render simultaneously bright field and quantitative phase images, in a single scan. cSLIM will be implemented in a whole slide imaging (WSI) instrument with the color bright field image familiar to pathologists, while maintaining a stain-independent signal, which has intact prognosis value. The WSI instrument’s high sensitivity to stroma and collagen fibers will be used to develop robust markers for breast prognosis, which are independent of tissue slice thickness, color variability within the same stain type (say, H & E), and across stains (H & E, various immunochemical stains, etc). With this new instrument, we will test the staining-invariance performance on 196 TMA cases and validate with 300 biopsies. The work is the results of combining expertise in imaging, pathology, and image processing across four sites: UIUC Beckman Institute, the Mills Breast Cancer Institute in Urbana, UIC Pathology, and U. Wisconsin.