MALIGNANT CYTOLOGICAL WASHINGS FROM PROSTATE SPECIMENS: : AN INDEPENDENT PREDICTOR OF BIOCHEMICAL PROGRESSION AFTER RADICAL PROSTATECTOMY

Abstract

Malignant cells have previously been identified in the cytological washings of prostate specimens obtained at radical prostatectomy for clinically localized prostate cancer. We investigated whether malignant cells in the cytological washings of radical prostatectomy specimens predict biochemical progression. The affect of total androgen blockade on cytological washings was also examined. Cytological washings were obtained from radical prostatectomy specimens in 147 consecutive patients undergoing the procedure for clinically localized prostate cancer between November 1993 and April 1998. Of the 147 patients 54 were randomly selected to receive 1 month of total androgen blockade immediately before prostatectomy. To obtain the cytological specimen the extirpated prostate was subjected to a normal saline bath, as previously described. The cytology specimen was examined by a single cytopathologist blinded to preoperative and pathological findings. Biochemical progression, defined as prostate specific antigen 0.15 ng./ml. or greater, was determined using the Kaplan-Meier method. We also performed multivariate analysis of factors related to progression, including prostate specific antigen, pathological stage, margin status, Gleason grade and cytology status. Median followup was 37 months (range 13 to 66). Followup was available in 146 of 147 cases. Cytological washings were malignant in 14 of 92 patients (15%) who did not receive total androgen blockade preoperatively. In this group without androgen blockade the biochemical progression rate was significantly higher in those with positive cytology (p < 0.001). Positive cytology was an independent predictor of progression on multivariate analysis and a stronger predictor of progression than Gleason grade. No malignant cells were observed in cases of preoperative total androgen blockade (p < 0.001). However, biochemical progression was similar in the groups with and without androgen blockade (p = 0.355). Malignant cells may be identified in the cytological washings of radical prostatectomy specimens and they are an independent predictor of biochemical progression. One month of total androgen blockade preoperatively significantly decreases the rate of positive cytology but does not appear to change the rate of early biochemical failure.