Login to Access Video or Poster Abstract: MP64-11
Sources of Funding: None

Introduction

Because of the spread of saturation biopsy and MRI guided needle biopsy, we may often see a cancer in transition zone (TZ). To assess whether there is a difference in prognosis in TZ and peripheral zone (PZ) prostate cancer, we studied the patients with clinically localized prostate cancers who had RARP.

Methods

We studied retrospectively the clinical and pathological features of 404 consecutive patients with prostate cancer treated with RARP by one surgeon at Tokyo Medical University. We categorized the patients into predominantly PZ cancer, TZ cancer and both TZPZ cancer based on the area of cancer on whole-mount step sections of RARP specimens. Biochemical recurrence (BCR) was evaluated using the Kaplan-Meier method. With a Cox hazard regression analysis, we identified significant preoperative factors that predict BCR, and based on these results, we developed a nomogram to predict the non-BCR at 5 years after RARP. A concordance index was used to assess the value of a nomogram, and a calibration plot was used to compare the predicted values to actual values.

Results

PZ, TZ and both cancer was identified in 64%, 33% and 3% of the patients, respectively. Non-BCR rates at 2 and 5 year for patients with PZ were 88% and 78%, respectively, compared to 97% and 97% for patients with TZ cancer (p=0.0072). Also, none of 10 patients with both cancer had a BCR. While the frequency of positive surgical margins was similar between PZ and TZ groups (28% for both), the patients with PZ were more likely to have seminal vesicle invasion compared to patients with TZ (10% vs. 1%, p=0.001) and to have >=4+3/4+4 Gleason score (52% vs.40%, p=0.37). In fact, in the absence of seminal vesicle invasion and >=4+3/4+4 Gleason score, non-BCR rate at 3 year was 95% for patients with PZ and 99% for patients with TZ (p=0.40). A Cox hazard regression analysis showed that zonal distribution of cancer (p=0.004) was significantly associated with BCR after controlled with PSA, pathological stage, Gleason score and surgical margins. Based on this analysis, the postoperative momogram to predict non-BCR was constructed with an excellent concordance index of 0.89. The calibration plots also appeared to be good.

Conclusions

We concluded that the prognosis of a patient with prostate cancer is significantly more dependent upon the features of cancer in the PZ than in the TZ. Assessment of zonal distribution of cancer is important to gauge prognosis.

Funding

None