Purpose: We used clinical variables to predict prostate cancer detection on re-biopsy among patients diagnosed with high grade prostatic intraepithelial neoplasia (PIN) or atypia on initial prostate biopsy.
Materials and methods: A total of 45 men with atypia and 43 with high grade PIN were eligible for our study. Clinical variables were tested with univariate and multivariate logistic regression to predict who would have cancer on re-biopsy. We also calculated the odds of detecting cancer with various repeat sampling strategies and determined whether the location of initial atypia or high grade PIN is correlated to that of cancer on re-biopsy.
Results: Of the patients in the atypia and high grade PIN groups 51% had cancer on re-biopsy. Cancer was diagnosed significantly earlier in the high grade PIN than in the atypia cohort (average 7.5 versus 22.9 months, respectively, p = 0.005). Multivariate logistic modeling showed that digital rectal examination and patient age were independent predictors of cancer in atypia, whereas no variables were significantly predictive for high grade PIN. Of cancers in the atypia and high grade PIN 65% and 74%, respectively, would have been detected if re-biopsy was focused only at the initial site of disease.
Conclusions: Men with atypia or high grade PIN merit close followup because 50% will have cancer on re-biopsy as will those who are older with an abnormal digital rectal examination. Although re-biopsy should focus primarily on the original site of atypia or high grade PIN, cancer detection significantly increases with the sampling of adjacent sites.