Logo: to the web site of Uppsala University

Background and objective

There are several methods for estimating prostate-specific antigen (PSA) doubling time (DT), and many men exhibit long periods of undetectable PSA below 0.1 ng/ml after radical prostatectomy (RP). We quantified between-method differences in estimated PSA-DT at PSA relapse after RP and evaluated how these differences affect discrimination of prostate cancer (PCa) death.

Methods

Men who underwent RP in 2007–2024 and subsequently developed a PSA relapse, defined as two consecutive values above 0.2 ng/ml, were included. We estimated PSA-DT from the date of the first undetectable PSA after RP until PSA relapse using four methods: (1) all detectable PSAs only, (2) two most recent detectable PSAs, (3) three most recent PSAs (with the third forced to 0.1 ng/ml if ≤0.1 ng/ml), and (4) all PSAs using a missing data approach. Discrimination of 10-yr PCa death was assessed using concordance index (C-index).

Key findings and limitations

A total of 3826 men were included. Median PSA-DT ranged from 4.4 to 8.6 mo, with a wide range in PSA-DT across different estimation methods. All four methods discriminated PCa death to a similar level, with C-index ranging from 0.70 to 0.74, which was slightly lower than the C-index (0.77) for the absolute PSA level at relapse. This study was restricted to men with PSA relapse after RP and may not be generalizable to other patient categories.

Conclusions and clinical implications

PSA-DT after RP was sensitive to the choice of estimation method, and all methods had comparable prognostic value to the last measured PSA level.