Residual cancer in radical prostatectomy specimens from men with biopsy-proven adenocarcinoma occasionally may be difficult, or even impossible, to identify. Although this finding was recently described as "minimal residual cancer" or the "vanishing cancer phenomenon," there are no data on the incidence of this phenomenon in surgical pathology practice. We evaluated 3,038 consecutive radical prostatectomies performed at the Johns Hopkins Hospital between 1988 and 1995, excluding cases with a history of transurethral resection, prior therapy with a luteinizing hormone-releasing hormone agonist, focal Gleason grade 4 or 5, capsular penetration, or a positive surgical margin. Of this group, 84 cases with minimal or no residual cancer were identified. In 60 of these cases, residual cancer was "difficult to find" (mean total volume, 0.03 cc; range, 0.01-0.08 cc); in 20 cases, residual cancers were classified as "minute" (mean total volume, 0.07 cc; range, 0.03-0.13 cc). In four cases, no residual cancer could be identified, including two cases in which the diagnosis of cancer on needle biopsy was confirmed, one case in which review of the diagnostic needle biopsy revealed only high-grade prostatic intraepithelial neoplasia, and one case in which molecular analysis demonstrated mislabeling of the needle biopsy specimen. The annual incidence of minimal residual cancer increased from 0.5% in 1988 to 4% in 1993 and has begun to plateau at 3 to 4% since 1993 (p = 0.0016 for increasing trend). These data confirm the general impression that with more vigilant screening of men for prostate cancer, there has been an associated increase in cancer with little or no residual cancer at radical prostatectomy.