INTRODUCTION AND OBJECTIVE: Prostate cancer is a multifocal disease with distinct molecular subtypes. However, the utility of genomic subclassification has been challenged due to the high degree of interand intra-focal… Click to show full abstract
INTRODUCTION AND OBJECTIVE: Prostate cancer is a multifocal disease with distinct molecular subtypes. However, the utility of genomic subclassification has been challenged due to the high degree of interand intra-focal heterogeneity. We sought to analyze the prevalence of adjacent, molecularly different tumors in a single focus e termed collision tumors, as a model for intrafocal genomic heterogeneity. METHODS: At our institution, we identified 98 patients from the Early Detection Research Network (EDRN) cohort, who had prospectively collected radical prostatectomies with information on Gleason score, pathologic staging, PSA, ERG, SPINK1 and SPOP status from 2008 to 2012. We used dual ERG/SPINK1 immunohistochemistry (IHC) to detect two adjacent but clonally distinct, tumor populations at one tumor focus identified by hematoxylin and eosin (H&E). Tumor foci with intense and diffuse (>90% of cells) nuclear ERG staining were considered ERGþ and intense cytoplasmic SPINK1 staining in >10% cancer cells were considered SPINK1þ. SPOPmutation status was screened by High Resolution Melting curve and confirmed by Sanger Sequencing. RESULTS: Our cohort had median age of 62, median PSA of 5, and the majority had Gleason grade group 3 or less (89.8%), pT2 (76.5%), and pN0 (92.9%) disease (Table 1). While 40% of patients had one nodule on prostatectomy specimen H&E examination, the remaining 60% had two or more nodules. Overall, 3% of patients had collision tumors that were identified on dual ERG/SPINK1 IHC and SPOP screening (Figure 1). Amongst two additional datasets from the multi-institutional EDRN cohort there was 9.48% (22/232) frequency of collision tumors. CONCLUSIONS: Collision tumors are present in up to 10% of men with prostate cancer. Intra-focal heterogeneity observed in primary prostate cancer may be related to this phenomenon, suggesting that genomic heterogeneity from biopsy samples may be partly due to sampling error rather than the futility of prostate cancer genomics.
               
Click one of the above tabs to view related content.