Combining PSA density and imaging data with transcript markers from urinary cells to improve prediction of prostate cancer reclassification in patients on active surveillance.

Objectives: To investigate selected transcripts from urinary cells regarding their potential to predict risk reclassification in patients with prostate cancer (PCa) on active surveillance (AS). Patients and methods: Urine samples were prospectively collected from 74 patients before control biopsy and immediately processed to obtain urinary cells. A panel of 29 PCa-associated transcripts with PPIA, RPLP0 and TBP as reference genes was determined by quantitative PCR following multiplexed reverse transcription of RNA with a pre-amplification step. Receiver operating characteristic (ROC) curve analyses including the calculation of the area under the curve (AUC) were conducted to assess the predictive ability for reclassification of the respective parameters. A multivariate regression analysis was performed to identify independent predictors suitable for combination. Results: Overall, 31 patients (42%) showed PCa risk reclassification (defined by ISUP group 1 with PSA > 10 ng/mL or ISUP group 2-5 with any PSA level) at control biopsy. The AUC values and accuracies of PSA, PSA density, PSA velocity and maxPI-RADS for indicating PCa risk reclassification ranged between 0.738 and 0.767 and 66% and 73%, respectively. Of the investigated transcripts, a predictive potential could be shown for MALAT1, MCL1, NEAT1, STAT3 and STAT5B (AUC = 0.622-0.713, accuracies 64%-73%). A model combining PSA density, maxPI-RADS and the transcript MCL1 resulted in an AUC of 0.828 with a sensitivity, specificity and accuracy of 58%, 93% and 78%, respectively. Conclusion: These findings highlight the potential of urinary transcripts, particularly in combinations with clinical parameters, for the non-invasive monitoring of PCa during AS.