Despite a body of evidence demonstrating the importance of the tumor-mediated immune response to the outcome of prostate cancer patients, no studies have adequately evaluated the tumor-mediated immune response in African American patients. Given that the African American race is a risk factor for more aggressive and lethal prostate cancer, the purpose of this study is to test the overall hypothesis that difference in the quantity and localization of tumor-infiltrating lymphocytes into the prostate gland correlat with PCa racial health disparities and that androgen signaling regulates this process. For this study, we will utilize a prostate tissue biorepository collected between 2003 and 2013, containing over 6,500 archival prostate cancer specimens, 25% (1,630) of which are from African American men. To test our hypothesis in a definitive way, we will utilize the concept of an "immunoscore." The immunoscore, a combined score based on the quantitation of CD3+ and CD8+ T lymphocytes in a tumor, has already been found to be a strong predictive indicator in patients with colorectal cancer and is only starting to be explored for other solid tumor types. Inour Specific Aim 1, we will adapt existing immunoscoring methodology to be compatible with and applicable to prostate cancer, and will test the reproducibility and robustness of calculating the prostate cancer immunoscore in a small, preliminary panel of African American and Caucasian patients from our tissue bank. In our Specific Aim 2, we will determine the prostate cancer immunoscore and androgen receptor status in a large cohort of African American (1,630) and Caucasian (1,630) patients, for which we also have 10-year follow-up data. We will then be able to determine if differences exist between African Americans and Caucasians with respect to: prostate cancer immunoscores, the spectrum/quantity of tumor-infiltrating T lymphocytes (CD3+, CD4+, and CD8+), and/or androgen receptor status; and if race is predictive of any of these variables. In our Specific Aim 3, we will perform Cox regression analysis and statistical modeling to determine if the prostate cancer immunoscore and/or AR status are predictors of long-term clinical outcomes (biochemical recurrence and/or disease-specific mortality) in African American and/or Caucasian patients with localized prostate cancer. If so, we will be able to construct an immunoscore-based classifier that clinicians can use to better predict outcome in their African American patients with prostate cancer-something that currently doesn't exist and is desperately needed. This work will also provide the clinical justification for future mechanisti studies surrounding race, tumor-infiltrating T lymphocytes, and androgen signaling in prostate cancer.