To the Editor: Guo et al1 reported that prostate cancer of transition zone origin lacks TMPRSS2-ERG gene fusion. In a cohort of 30 patients, the authors showed that transitional zone tumors completely lacked TMPRSS2-ERG gene fusion in contrast with the finding of 43% gene fusion in cancers from the same glands originating from the peripheral zone. They concluded that the TMPRSS2-ERG gene fusion is associated with the zonal origin of prostate cancer. This study raises several points on which we would like to comment. We note that the selected cases showed significantly larger tumor volumes in the transitional zone (mean of 4.0 cm3) than in the peripheral zone (mean of 1.2 cm3) tumors. In addition, the transitional zone cancers in this study showed primary Gleason pattern 4 in 7 of 30 (23%) cases and even showed secondary Gleason pattern 5 in 2 of 30 (7%) cases. Overall, transitional zone cancer from this study compared with the peripheral zone cancers showed higher primary Gleason pattern in 6 of 30 (20%) cases and higher overall Gleason score in 5 of 30 (17%) cases. Typically, incidental cancers of the transitional zone encountered in practice are of lower tumor volume and show lower Gleason scores than the coexisting cancers in the peripheral zone. It is unclear whether those higher-grade ‘transition zone’ tumors included in this study are biologically different from the true low-grade tumors typically observed in the transition zone.2, 3 As only the largest tumor foci were investigated, the smaller tumor foci that were not studied could have expressed the gene fusion. We and other groups 4, 5, 6 showed that the TMPRSS2-ERG exhibits heterogeneity within different cancer foci in the same prostate gland. Barry et al4 reported that 41% of multifocal cancers on prostatectomy show discordance of the TMPRSS2-ERG gene fusion. Therefore, to unequivocally conclude that transitional zone cancers completely lack TMPRSS2-ERG gene fusion, the typical transitional zone cancers also should have been included in the study and all tumor foci should have been analyzed for the gene fusion. It would also be beneficial if the authors clarified whether they required all evaluated cells (100%) to conclude that transitional zone cancers completely lack the TMPRSS2-ERG gene fusion signal or whether they established a certain cutoff point for a positive fusion status to be regarded as positive (for example, 80 or 90% of examined cells). Clarifying this methodological issue would ensure reproducibility in other studies and may clarify the differences with the findings in other cohorts.

In an unpublished study, we analyzed 50 tumors that were incidentally discovered in men who presented for benign prostatic disease and who underwent transurethral resection of the prostate (TURP). All patients had Gleason pattern of 4 or less, with only one patient showing Gleason score 4+5 on TURP, who was also found to have peripheral tumor of the same score when he underwent radical prostatectomy. We examined the ERG gene rearrangements in these presumed transitional zone tumors using break-apart fluorescent in-situ hybridization technique. We assessed a minimum of 100 nuclei, and all cells within one focus showed the same gene fusion status. Excluding the cases above, ERG gene rearrangements was present in 6 of 45 (13.3%) patients (only informative tests were analyzed). This rate of ERG gene rearrangements in unsuspected and incidental prostate cancers in TURP specimens is similar to the finding of 15%, recently shown in a large watchful waiting cohort from Sweden, which similarly included only TURP specimens.7

In summary, although we agree that biological difference may exist between transitional and peripheral zone tumors, we caution against the conclusion of complete absence of TMPRSS2-ERG gene fusion within the prostate transition zone tumors.