Purpose: We addressed whether Gleason score 3 + 4 = 7 and 4 + 3 = 7 cancers on needle biopsy behave differently and whether this behavior is independent of the number of cores involved by cancer. If it is not an independent predictor of prognosis, one may report Gleason score 7 cancer with the number of positive cores without regard to whether the primary pattern was 3 or 4. This practice would remove a source of poor interobserver reproducibility when grading prostate cancer on needle biopsy.
Materials and methods: We identified 537 patients with Gleason score 7 tumors on biopsy. The results of patient preoperative digital rectal examination, serum prostate specific antigen (PSA) measurement and age were used to predict 4 outcomes based on assessment of the corresponding radical prostatectomy specimens, including 1) pathological stage (organ confined, focal extraprostatic extension, nonfocal extraprostatic extension or seminal vesicle-lymph node involvement), 2) organ confinement (yes/no), 3) Gleason score and 4) surgical margin status (positive/negative)
Results: Multivariate regression of postoperative Gleason score groups against all 5 input variables (3 + 4 versus 4 + 3, number of positive cores, PSA, age and digital rectal examination) yielded a statistically significant positive correlation with preoperative PSA (p <0.001) and preoperative Gleason scores of 4 + 3 versus 3 + 4 on biopsy (p <0.001). Pathological stage correlated with preoperative PSA (p <0.001), Gleason score 4 + 3 disease (p = 0.016), positive digital rectal examination (p <0.001) and 3 or more positive cores (p = 0.016). Positive surgical margins were predicted only by preoperative PSA (p = 0.001).
Conclusions: Because the biological behavior of biopsy Gleason score 3 + 4 or 4 + 3 of Gleason score 7 cancer differs regardless of the number of cores involved, future nomograms predicting pathological stage would benefit from examining 3 + 4 and 4 + 3 disease separately.