Prostate cancer detection: relationship to prostate size.

OBJECTIVES

It has been suggested that the lower detection rate for cancer in large prostates is due to the smaller proportion of tissue sampled. To examine this hypothesis, we evaluated whole-mount radical prostatectomy specimens in which the volume of cancer had been determined. We correlated cancer volume to overall gland volume. In addition, we performed stochastic computer simulations of parasagittal sextant biopsies on the same group of radical prostatectomy specimens. We correlated the likelihood of a positive cancer biopsy simulation with tumor volume and gland size.

METHODS

Six hundred seven tumor foci from 180 serially sectioned whole-mount prostatectomy specimens were mapped and digitized. Tumor volume was calculated by a step-section planimetry algorithm. Before sectioning, each gland was weighed. Systematic parasagittal sextant biopsies were computer simulated for each case. For each prostate, 40 simulations were performed, with random variations in biopsy location programmed for each run. Overall cancer detection by biopsy was considered positive if 90% of the 40 simulation runs were positive for cancer. Chi-square tests were used to evaluate statistical significance.

RESULTS

Small-volume cancers (0.5 cc or less) were twice as frequent in large glands greater than 50 g (P = 0.03). These small-volume tumors comprised 33% (13 of 40) of cancers in prostates greater than 50 g, 16% (5 of 31) in glands less than 30 g, and 14% (15 of 109) in glands 30 to 50 g. The rate of positive sextant biopsy simulation was lower in glands greater than 50 g than in glands 50 g or less (48% versus 67%, P<0.03). Smaller cancers were much less likely to be detected in the simulations. The simulation detection rate for cancers 0.5 cc or less was 18% (6 of 33), compared with a detection rate of 73% (107 of 147) for cancers greater than 0.5 cc (P<0.00001).

CONCLUSIONS

The observed lower cancer detection rate in large glands is a result of the higher proportion of low-volume cancers in these glands. This suggests that large prostates are more likely to be biopsied because of an elevated prostate-specific antigen value resulting from benign elements of the gland and not from a significant cancer. Increasing the number of cores solely to compensate for increased prostate size risks a disproportionate increased detection of small-volume tumors with a low clinical likelihood of progression.