Purpose: In recent years the role of pelvic lymph node dissection before or at radical prostatectomy has been questioned. We compared the outcome of performing or omitting pelvic lymph node dissection in patients undergoing radical retropubic prostatectomy by formal decision analysis.
Materials and methods: We created a decision tree for patients undergoing radical retropubic prostatectomy that accounts for all possible branch points and outcomes. Outcome probabilities applied to the model were derived from published data. Utility values were determined for each outcome state by a panel of experts. Decision analysis was done using decision analysis computer software. Sensitivity analysis was performed to determine which elements in the model were most important and to calculate threshold values.
Results: Using probability data from the literature and our assigned outcome values, decision analysis initially favored omitting pelvic lymph node dissection. Sensitivity analysis revealed that the threshold values for the incidence of positive lymph nodes and the sensitivity of lymph node dissection and frozen section analysis at which outcomes were equivalent were 18% and 80%, respectively. Analysis was insensitive to the pelvic lymph node dissection complication rate.
Conclusions: We performed formal decision analysis to determine the incidence of lymph node metastasis below which pelvic lymph node dissection is not warranted at radical retropubic prostatectomy. Our results suggest that lymph node dissection is unnecessary in the subset of patients in which the risk of lymph node involvement is less than 18%.