Clinical characteristics and associated factors of de novo and recurrent prostate cancer after kidney transplantation

<jats:p>Kidney transplant recipients experience a higher burden of several malignancies, yet the factors associated with prostate cancer presentation after transplantation remain poorly understood. Unlike malignancies strongly associated with impaired immune surveillance, prostate cancer has not consistently demonstrated an increased incidence after transplantation, suggesting that different mechanisms may underlie disease presentation. This study evaluated recipient, donor, transplant, immunologic, and immunosuppressive factors associated with prostate cancer phenotype after kidney transplantation.  A retrospective cohort study was conducted using national transplant registry data from adult kidney transplant recipients diagnosed with post-transplant prostate cancer between 2015 and 2024. Cases were classified as de novo (no pre-transplant history of prostate cancer) or recurrent (documented pre-transplant history). Multivariable Firth penalized logistic regression was used to evaluate factors associated with recurrent phenotype. Prespecified sensitivity analyses included deceased donor restricted models, incorporation of donor organ quality variables, and adjustment for time from transplantation to cancer diagnosis. Exploratory machine learning analyses included elastic net logistic regression, random forest, and extreme gradient boosting. The cohort included 660 recipients, of whom 623 (94.4%) had de novo disease and 37 (5.6%) had recurrent disease. Recipient age was the only variable consistently associated with recurrent phenotype across primary and sensitivity analyses (adjusted odds ratio per year 1.11, 95% CI 1.05–1.17; p&lt;0.001). Immunosuppressive regimen, donor characteristics, immunologic variables, and time from transplantation to cancer diagnosis were not independently associated with phenotype in the primary cohort. In deceased donor restricted analyses, alemtuzumab induction showed an exploratory association with recurrent phenotype, although estimates were imprecise. Machine learning models demonstrated modest discrimination and calibration and did not outperform penalized regression approaches. These findings suggest that, among kidney transplant recipients with prostate cancer, differences between recurrent and de novo presentation are more closely associated with recipient age and underlying disease characteristics than with transplant exposures or specific immunosuppressive regimens.</jats:p>