Transcriptomic characterization of prostate cancer in patients with HIV infection.

396 Background: Prostate cancer is the most commonly diagnosed malignancy in the aging population of men living with HIV. Despite an increasing burden of prostate cancer, limited data exist to determine whether HIV infection is associated with differences in tumor biology and clinical outcomes in prostate cancer. We leveraged a large real-world data linkage to evaluate transcriptional profiles in incident prostate cancers among persons living with HIV (PWH) versus people without HIV (PWoH). Methods: We used a secure, anonymous linkage between prostate biopsy or radical prostatectomy whole transcriptome profiling data from a 22-feature Decipher prostate genomic classifier (GC) and real-world data from insurance claims, pharmacy records, and electronic health records (Veracyte, San Diego, CA). HIV status was identified using claims-based algorithms combining diagnosis codes and antiretroviral pharmacy records. Propensity scores for HIV status were used to match PWH to PWoH (1:5) based on clinical risk group, Gleason score, and test type. We compared transcriptomic signatures and subtypes using Chi-squared and Wilcoxon rank sum descriptive statistics. Results: We identified 1,170 eligible PWH with genomic data who met inclusion criteria based on diagnosis and treatment codes, as well as 5,850 matched PWoH comparators. Transcriptomic profiling indicated that prostate tumors in PWH had a more activated tumor immune microenvironment with higher T cell inflamed, activated CD4, CD8, cytotoxic T cell, B cell activity, ESTIMATE immune bulk infiltration and a higher prevalence of the basal immune subtype (46% vs 38%, p<0.001) in PWH as compared with comparators. In addition, PWH demonstrated higher PORTOS radiation response and slightly higher Decipher scores (median 0.51 vs 0.48, p<0.01) were found as compared to PWoH. Lower androgen receptor activity, pTEN inactive and higher neuroendocrine-like, p53 mutation and pRb loss signature scores were observed in PWH. Conclusions: Using a large national linkage of prostate tumor profiling and clinical data, we found that primary tumors of PWH exhibited distinct prostate cancer transcriptomic profiles compared to PWoH counterparts, including a higher prevalence of an activated tumor immune microenvironment and basal immune subtype. These findings suggest potential biologic differences in prostate cancer linked to HIV that underscore a need for further inquiry to understand whether there is a need to tailor management in this population.