A population-based assessment of the National Comprehensive Cancer Network practice guideline indications for pelvic lymph node dissection at radical prostatectomy

OBJECTIVES To examine the ability of the threshold recommended by the National Comprehensive Cancer Network (NCCN) in correctly predicting histologically-confirmed lymph node invasion (LNI). The 2010 NCCN practice guidelines for prostate cancer recommend a pelvic lymph node dissection (PLND) at radical prostatectomy in all individuals with a nomogram predicted LNI risk of >= 2%. PATIENTS AND METHODS We assessed 20 877 patients who were treated with radical prostatectomy and PLND between 2004 and 2006, within the Surveillance, Epidemiology and End Results database. The 2% nomogram threshold, as well as other threshold values (range 1-10%) were tested. Finally, we externally validated the NCCN guideline nomogram. RESULTS Overall, 2.5% of patients had LNI. The use of the 2% threshold would allow the avoidance of 23% of PLNDs, at the cost of missing 1.7% of patients with LNI. Conversely, the use of a 3% threshold would allow the avoidance of 58% of PLNDs, at the cost of missing 15% of patients with LNI vs 72% and 26%, respectively, for the 4% threshold. Overall, the accuracy of the NCCN guideline nomogram quantified according to the receiver-operator characteristics-derived area under the curve was 82%. CONCLUSIONS In a population-based sample, the NCCN guideline nomogram is highly accurate. However, the 2% threshold will permit the avoidance of only 23% of PLNDs, instead of the 48% intended by the NCCN guidelines. The use of a 3% threshold may allow a lower rate of PLND overtreatment, although it will miss more patients with LNI.