To study the rate of change in prostate specific antigen (PSA velocity) in patients with prostate cancer initially managed by 'watchful waiting'.Patients and methods
Serial PSA levels were determined in 141 patients with prostate cancer confirmed by biopsy, who were initially managed expectantly and enrolled between May 1990 and December 1995. Sixty-seven patients eventually underwent surgery (mean age 59 years) because they chose it (the decision for surgery was not based on PSA velocity). A cohort of 74 patients remained on 'watchful waiting' (mean age 69 years). Linear regression and logarithmic transformations were used to segregate those patients who showed a rapid rise, defined as a > 50% rise in PSA per year (or a doubling time of <2 years) and designated 'rapid risers'.Results
An initial analysis based on a minimum of two PSA values showed that 31% were rapid risers. Only 15% of patients with more than three serial PSA determinations over ≥ 6 months showed a rapid rise in PSA level. There was no advantage of log-linear analysis over linear regression models.Conclusion
Three serial PSA determinations over ≥ 6 months in patients with clinically localized prostate cancer identifies a subset (15%) of patients with a rapidly rising PSA level. Shorter PSA surveillance with fewer PSA values may falsely identify patients with rapid rises in PSA level. However, further follow-up is required to determine if a rapid rise in PSA level identifies a subset of patients with an aggressive biological phenotype who are either still curable or who have already progressed to incurability through metastatic disease.