Interpreting PSA Kinetics Using GLMM Technique

Abstract

PSA doubling time (DT) was used to stratify patients into groups at low and high risk of progression. The PSA kinetics in these two groups was modeled. Evolution of serial PSA measurements over time was estimated from a general linear mixed model (GLMM) of ln(PSA). The corresponding average and individual PSA DTs were also calculated. Since November 1995, 231 patients had at least 6 months of follow-up and at least 3 PSA measurements. Based on PSA DT and repeat biopsy, 93 fulfilled the criteria for “high risk of disease progression” and 138 were defined as low risk of disease progression. Given the baseline status of individuals, two reference average lines (high risk and low risk) were derived to model the evolution of PSA levels and permit more rational decision-making regarding the need for definitive intervention. The average PSA DT was 2.97 years (95% CI: 2.2–4.4 years) in patients who were allocated to the high-risk group and 6.54 years (95% CI: 4.8–12.3 years) in those who were low risk. By applying the dynamic prognostic rule in combination with serial biopsy, a rational decision for definitive intervention based on the risk of progression could be optimally recommended about 2.3 years after the initiated surveillance. We believe that the GLMM approach provides real advantages compared to more widely used methods of calculating PSA kinetics. Our web site, www.Asure.ca, provides this approach to anyone who wishes to calculate PSA kinetics using this technique.KeywordsGeneral Linear Mixed Model (GLMM)Prostate-specific Antigen (PSA)GLMM ApproachRepeat BiopsyRandom Linear SlopeThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.