Do nomograms predict aggressive recurrence after radical prostatectomy more accurately than biochemical recurrence...

Area of Science:

• Urology
• Oncology
• Medical Statistics

Background:

• Biochemical recurrence (BCR) after radical prostatectomy (RP) is a common outcome.
• Predicting aggressive recurrence, defined by a short prostate-specific antigen (PSA) doubling time (DT), is crucial for guiding treatment decisions.
• Existing risk stratification models aim to predict these outcomes, but their comparative accuracy requires evaluation.

Purpose of the Study:

• To compare the predictive accuracy (PA) of nine established risk stratification models.
• To assess model performance in predicting biochemical recurrence (BCR) versus aggressive recurrence (BCR with PSA doubling time <9 months) post-RP.
• To determine if model accuracy differs between preoperative and postoperative risk assessments.

Main Methods:

• A cohort of 1550 men undergoing RP between 1988 and 2007 was analyzed.
• Nine distinct risk stratification models were evaluated for their ability to predict BCR and aggressive recurrence.
• Predictive accuracy was quantified using the concordance index (c-statistic).

Main Results:

• The 10-year risks for BCR and aggressive recurrence were 47% and 9%, respectively.
• All nine models demonstrated higher PA for predicting aggressive recurrence (mean c=0.756) compared to BCR alone (mean c=0.702), with a mean difference of 0.054.
• This improvement in PA was more pronounced for preoperative models than postoperative models (0.053 vs 0.036, P=0.03).
• Sensitivity analyses using varying definitions for aggressive recurrence (PSA DT <6 or <12 months) yielded similar results.

Conclusions:

• Existing risk models offer superior predictive accuracy for aggressive recurrence compared to BCR alone after RP.
• The enhanced prediction of aggressive recurrence likely stems from models' focus on cancer biology, which correlates more strongly with aggressive disease.
• While all tested models showed comparable accuracy for predicting aggressive recurrence, their performance in this specific prediction is noteworthy.