Significance of preoperative parameters for predicting tumor volume in nonpalpable prostate cancer

Carcinoma extent in prostate needle biopsy tissue in the prediction of whole gland tumor volume in a screening population

Increasing prostate tumor volume has been shown to correlate with numerous adverse prognostic indicators for patients with prostate carcinoma The ability to predict tumor volume from pretreatment parameters is potentially critical in the stratification of patients for different management strategies. We assessed the capacity of preoperative variables to predict tumor volume in 100 men diagnosed with prostate cancer in a prostate-specific antigen (PSA)-based screening program. Preoperative information included total serum PSA concentration and needle biopsy tissue variables, including Gleason score, number of positive cores, linear extent of carcinoma in millimeters, greatest percentage of carcinoma (in a single core), total percentage of carcinoma (all cores), presence of perineural invasion, and percentage of high-grade carcinoma. The postoperative end point was total tumor volume in radical prostatectomy tissue, calculated by image analysis. We determined independently significant factors and generated a predictive modelfor whole gland tumor volume. Total tumor volume was related significantly in multivariate analysis to 3 preoperative variables: linear extent of carcinoma, exponential number of positive cores, and serum PSA. A predictive model generated based on these 3 variables accounted for only 65% of the natural deviance of the data owing to data-point scatter for individual patients, suggesting that additional variables are needed to more accurately predict tumor volume. Findings highlight the importance of reporting quantitative measures of tumor amount in prostate needle biopsy specimens; several measures of tumor extent (vs 1 measure) provide maximal information on prostate cancer size.

Predicting prostate carcinoma volume and stage at radical prostatectomy by assessing needle biopsy specimens for percent surface area and cores positive for carcinoma, perineural invasion, Gleason score, DNA ploidy and proliferation, and preoperative serum prostate specific antigen: a report of 454 cases

BACKGROUND

DNA ploidy analysis of prostate carcinoma is a generally accepted prognostic marker, particularly when tumors are extraprostatic at the time of surgery. In the past decade, the DNA content of prostate carcinoma frequently has been assessed in needle biopsy specimens based on the assumption that ploidy, in conjunction with serum prostate specific antigen (PSA) and Gleason score, provides valuable pretreatment information.

METHODS

Between 1995 and 1998, the authors identified a consecutive series of 454 prostate carcinomas, verified by needle biopsies and followed by radical retropubic prostatectomies (RRP). Based on the needle biopsies, DNA ploidy and MIB-I immunostaining were measured by digital image analysis (DIA). The authors also quantified the percent of nuclei in four categories from the DNA histograms. The DIA data were combined with the age of the patient at diagnosis, the serum PSA, Gleason score, percent cores and percent surface area positive for carcinoma, and status of perineural invasion in multivariate models using tumor volume and risk of extraprostatic extension (EPE) at RRP as the outcome variables.

RESULTS

Joint predictors of tumor volume at RRP were the percent cores positive for carcinoma (P < 0.0001), serum PSA (P < 0.0001), the percent surface area positive for carcinoma (P < 0.0001), and the percent nuclei classified by DNA quantification to be in the "S-phase" category (P = 0.03). Joint predictors of risk of EPE were the percent cores positive for carcinoma (P = 0.0004), a Gleason score of 7 (P < 0.0001), a Gleason score of 8 or 9 (P < 0.0001), serum PSA (P = 0.006) and perineural invasion (P = 0.02).

CONCLUSIONS

After adjusting for traditional prognostic markers, DNA ploidy interpretation and MIB-I quantitation of prostate carcinoma did not appear to jointly predict either outcome variable in the multivariate models. However, a quantitative measure related to both ploidy and proliferation, the percent of nuclei in the putative "S-phase" category from the DIA histograms, was found to jointly predict for tumor volume.