Elsevier

Urology

Volume 60, Issue 2, August 2002, Pages 264-269
Urology

Adult urology
Morphologic and clinical significance of multifocal prostate cancers in radical prostatectomy specimens

https://doi.org/10.1016/S0090-4295(02)01728-4Get rights and content

Abstract

Objectives. To examine the histologic details of small, independent cancers compared with the largest (index) tumor and their impact on prostate-specific antigen (PSA) failure in 486 men treated only by radical retropubic prostatectomy (RRP).

Methods. The tumor volume and percentage of Gleason grade 4/5 carcinoma were recorded in 3-mm step sections. Univariate statistics were calculated for the largest, total (largest plus smaller cancers), and smaller tumor volumes, number of independent foci, patient age, and follow-up. Cox hazards model determined the relative importance of all variables in relation to failure.

Results. The mean index tumor volume was 4.16 cm3; smaller cancer volumes averaged 0.63 cm3. The index cancer volume was gaussian in distribution; smaller tumor volumes were highly skewed toward 234 carcinomas less than 0.5 cm3. Only 17% of all cases had one carcinoma. The Cox model showed similar hazard rates of PSA failure for both the index (3.43) and the total cancer (3.74) volumes. The hazard rate for the presence of any Gleason grade 4/5 carcinoma was 6.5. As the numbers of smaller tumors increased, the PSA cure rates improved.

Conclusions. The PSA failure rates (hazard ratios) were similar for the index tumor and the index plus smaller cancers, confirming that predictive estimates only need to measure the largest carcinoma. The greater the number of lesser cancers, the smaller the size of the index cancer. The extraordinary multiplicity of these small independent cancers in 3-mm step sections may explain the poor correlation between six or more biopsies with the index cancer in radical prostatectomy specimens.

Section snippets

Patients and pathologic analysis

A total of 559 untreated radical prostatectomies were performed at Stanford between January 1, 1992, and January 1, 1996. Nineteen patients were excluded because of no PSA follow-up, 6 because of hormonal therapy before radical prostatectomy, 5 because of missing information on prostate weight, 1 because of no prostate carcinoma, and 38 (6.8%) because they had two or more independent index tumors of similar size, making it impossible to determine the largest tumor. We also excluded 2 men whose

Results

The mean and median size of the largest (index) cancer was 4.16 cm3 and 2.78 cm3, respectively (Table I). On average, 2.92 additional separate cancers (mean volume 0.63 cm3) were found. The mean and median ratio of the sum of smaller cancer volumes to the index cancer volume was 15% and 6%, respectively. The mean age at radical prostatectomy (Table I) was 62.9 years (interquartile range 59 to 68); the mean and median PSA follow-up after radical prostatectomy was 3.37 and 3.2 years,

Comment

In 83% of this series of 486 radical prostatectomies, prostate carcinoma was a multifocal disease. Figure 1 shows that the secondary cancers were mostly small; 58% were less than 0.5 cm3 in volume.

The linear relationship between the index cancer volume and progression to biochemical failure is generally accepted; it is no surprise that Figure 2 shows such a relationship between the index cancer volume and PSA failure. Figure 2, however, suggests that the presence of secondary cancers appears to

Conclusions

Smaller cancers are not a significant predictor of PSA failure after radical prostatectomy; the index cancer volume is equally predictive of PSA failure after radical prostatectomy as total cancer volume (index plus smaller cancers). The significance rests solely with the largest cancer. The information in Figure 4 suggests that the largest carcinoma in the prostate may increase its volume by assimilation of adjacent independent tumors, thereby adding to the biologic importance of the index

References (15)

There are more references available in the full text version of this article.

Cited by (322)

View all citing articles on Scopus
View full text