Project Summary The goal of this translational project is to link the genetic and cellular features of the overarching, proving ground [cancer-permissive tumor microenvironment (TME)] hypothesis of this U54 application to imaging biomarkers that can be used clinically to detect and ultimately intercept prostate cancer. Little is known about the cellular subtypes that participate in the epithelial-stromal interactions in prostate cancer, or about the timing of their appearance during progression. We have developed a series of translational imaging agents designed to measure cancer epithelium and specific cells within the TME. We will leverage a subset of those agents to understand quantitatively the TME in prostate cancer so that it can be detected and reprogrammed early during malignant transformation. In Aim 1 we will image patients with primary prostate cancer prospectively with multi- parametric magnetic resonance imaging (mpMRI) and 18F-DCFPyL (PyL) positron emission tomography/computed tomography (PET/CT), which targets the prostate-specific membrane antigen (PSMA). We will perform standard-of-care PyL PET/CT imaging on patients with intermediate to high-risk prostate cancer prior to prostatectomy. We will quantify and correlate between imaging and pathology with the hypothesis that PSMA-low regions in otherwise high-grade tumors correlate with adverse pathology at prostatectomy in an effort to understand better the heterogeneity of primary disease. We will also compare PyL PET/CT with the cellular composition and spatial architecture of immune and stromal compartments of the prostate cancer TME. We will also retrospectively use prostatectomies from the same patients that were consented for genomic studies in Project 2. We will test the hypothesis that PyL PET/CT measured PSMA levels will directly correlate with MYC expression, MYC copy number gain, PTEN loss, and TP53 mutation. We will also determine in an unbiased manner which gene expression, genomic and epigenomic alterations best correlate with PSMA PET imaging features. We will also employ radiomic and machine-learning approaches not only to subtype patients with prostate cancer into prognostic groups, as we have done before, but also to predict who may have aggressive disease on initial PyL PET/CT staging. For this we will require merging not only the mpMRI and PyL PET/CT data, but also the histopathology, relevant genomics, and available protected historical information. Aims 2 and 3 involve longitudinal studies in the BMPC murine model, which we have developed and recapitulates human prostate cancer according to a specific timetable. We will determine the extent of macrophages and cancer- associated fibroblasts in and around the BMPC tumors as they develop over time and correlate their presence with imaging by small animal PET/MR. We will use novel imaging agents that we have developed, which target CSF1R (macrophages) and FAP (CAFs), and correlate the results with immunohistochemis...