Race-Specific 3D Computational Pathology Biomarkers for Predicting Prostate Cancer-Specific Recurrence
Abstract
An estimated 31,620 men are expected to die due to prostate cancer (PCa) in 2019 alone. In addition, over 174,650 men are expected to be diagnosed with PCa in 2019. PCa is also disproportionately prevalent and more aggressive in African American (AA) men. Incidence of PCa in AA men is almost 60% higher, with a two- to three-fold increased mortality compared to Caucasian American (CA) men, as well as the general population. Not only do AA men have more aggressive variants of PCa, their risk of dying from low-grade PCa is also double that of CA men. While there is no doubt that socioeconomic factors such as access to prostate specific antigen screening play a role in this health disparity, mounting evidence based on molecular and morphologic interrogation of PCa also points to biological differences in the disease phenotype between AA and CA men. Extant nomograms (e.g., CAPRA, Kattan, Sharait, and Swanson) as well as genomic (Decipher, a 22-gene expression test to predict the likelihood of post-surgical metastasis in PCa men) risk prediction models are developed on an overwhelmingly (~95%) Caucasian population and hence do not explicitly account for the tumor-specific attributes of PCa in AA men. There is thus an urgent need for risk assessment models to prognosticate PCa outcomes that are developed specifically for AA men. Population-specific companion diagnostic tools specifically geared toward AA men could help identify those PCa patients with low-grade disease who could be spared the morbidities of additional chemotherapy, radiation, or hormonal therapy. Similarly, these population-specific models could be used to identify those AA men with more aggressive PCa who would, in turn, be ideal candidates for more aggressive or adjuvant therapy in order to prevent future metastatic disease. Our team at Case Western Reserve University (CWRU) is a world leader in computational pathology tools to quantitatively characterize the PCa phenotype from routine 2D H&E tissue slide images. We have shown that quantitative characterization of the shape, architecture, and texture of nuclei and glands on tissue images can stratify PCa patients (N>900) based on their time to recurrence into low- and high-risk groups. In addition, we have shown that stromal features of PCa on surgical excised tissue images are different between AA and CA men and that a computational prognostic model taking these features into account is strongly associated with risk of recurrence in two validation datasets of N=170 AA PCa men. The AA-specific prognostic model is almost twice as accurate compared to a model trained with a combination of AA and CA men with PCa. Recent technological advances in 3D scanning techniques such as the open-top light sheet (OTLS) microscopy technique (developed by Co-Investigator Dr. Liu) promises to provide a transformative improvement in diagnostic performance for a number of reasons: (1) vastly greater (multiple log orders) sampling of tissue specimens, (2) volumetric imaging of cell distributions and tissue structures that can enable computational elucidation of cancer-specific prognostic and predictive features, and (3) nondestructive imaging. Considering that we are already able to discern significant morphometric differences in the stroma between AA and CA men with PCa using 2D computational image analysis, this begs the question: How much more prognostic value might be derived from the computational interrogation of stromal, epithelial morphology, as well as the shape, texture, and arrangement of glands in volumetric specimens imaged via OTLS compared to routine 2D tissue images? In this Prostate Cancer Health Disparity grant, we seek to develop a computational risk-assessment tool (PRISM3D) based on computer-extracted features from surgical specimens that are scanned by the novel 3D OTLS technology. The objectives of this project are two-fold. First, to combine the power of OTLS and 3D computational pathol
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 10, 2021
- Source ID
- W81XWH2010851
Entities
People
- Anant Madabhushi
Organizations
- Case Western Reserve University
- United States Army