Predicting survival of patients with node-positive prostate cancer following multimodal treatment.

BACKGROUND

According to the TNM staging system, patients with prostate cancer (PCa) with lymph node invasion (LNI) are considered a single-risk group. However, not all LNI patients share the same cancer control outcomes.

OBJECTIVE

To develop and internally validate novel nomograms predicting cancer-specific mortality (CSM)-free rate in pN1 patients.

DESIGN, SETTING, AND PARTICIPANTS: We evaluated 1107 patients with pN1 PCa treated with radical prostatectomy, pelvic lymph node dissection, and adjuvant therapy at two tertiary care centers between 1988 and 2010.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Univariable and multivariable Cox regression models tested the relationship between CSM and patient clinical and pathologic characteristics, which consisted of prostate-specific antigen (PSA) value, pathologic Gleason score, pathologic tumor stage, status of surgical margins, number of positive lymph nodes, and status of adjuvant therapy. A Cox regression coefficient-based nomogram was developed and internally validated.

RESULTS AND LIMITATIONS

All 1107 patients received adjuvant hormonal therapy (aHT). Additionally, 35% of patients received adjuvant radiotherapy (aRT). The 10-yr CSM-free rate was 84% in the entire cohort and 87% in patients treated with aRT plus aHT versus 82% in patients treated with aHT alone (p=0.08). At multivariable analyses, PSA value, pathologic Gleason score, pathologic tumor stage, surgical margin status, number of positive lymph nodes, and aRT status were statistically significant predictors of CSM (all p ≤ 0.04). Based on these predictors, nomograms were developed to predict the 10-yr CSM-free rate in the overall patient population and in men with biochemical recurrence. These models showed high discrimination accuracy (79.5-83.3%) and favorable calibration characteristics. These results are limited by their retrospective nature.

CONCLUSIONS

Some patients with pN1 PCa have favorable CSM-free rates at 10 yr. We developed and internally validated the first nomograms that allow an accurate prediction of the CSM-free rate in these patients at an individual level.