Introduction:

Prostate cancer is heterogeneous and multi-focal, and biopsy sampling often under-samples the tumor or samples different tumor foci. Such multiple sampling may cause molecular heterogeneity that leads to an underestimation of risk. Consequently, this impacts treatment decision making and management in localized prostate cancer.

Herein, we evaluate the degree of variability in genetic alterations, and genomic risk between different biopsy cores with respect to the sampling method (mpMRI vs template).

Methods:

Methods

A total of 348 biopsy samples from 156 patients from the Miami MRI selection for Active Surveillance versus Treatment (MAST) trial (NCT02242773) were used in this study. Across the whole time-course, all biopsy cores with successful gene expression profiling were categorized by mpMRI targeted (110 samples) or template (187 samples) sampling. Cores of higher Gleason score from each group were selected. Variation in expression was analyzed across biopsy samples from same patients.

Results:

The Decipher genomic score was statistically different across template or targeted biopsy in unpaired analysis (p=0.02), but no difference in paired analysis. Of the different molecular pathways, there was no difference regarding gene expression between both groups, except for E2F transcription factor.

MRI targeted biopsy Decipher genomic score did not vary according to PIRADS score (p=0.6). Angiogenesis pathway, E2F transcription factor pathway, G2/M checkpoint pathway were highly expressed in PIRADS 5 lesions, whereas estrogen response signaling was lower in PIRADS 5.

Conclusion:

mpMRI sampling could lead to could result in better genomic assessment for prostate cancer, however, radiologic lesion scoring is not correlated to the genomic risk.

Funding: N/A