Prostate cancer : population-based screening and markers for long-term clinical outcome

Abstract

In 1988 and 1989 a large screening study for prostate cancer was launched in Stockholm, Sweden. At the time approximately 27 000 men between 55 and 70 years of age resided within a defined area of southern Stockholm. 2400 men were randomly selected to participate in the trial and those accepting (n=1782), were examined with digital rectal exam (DRE), transrectal ultrasound (TRUS) and a PSA test. If DRE or TRUS indicated suspicious findings or if PSA levels were 10 ng/mL or greater quadrant core biopsies of the prostate were performed. Additionally, the screening algorithm employed stipulated reexamination with DRE and TRUS if PSA concentrations were between 7 ng/L and 10 ng/mL. The initial screening yielded 65 cases of prostate cancer. In this thesis the screening material have been assessed after 20 years (paper I) and 30 years (paper II-IV).

In paper I the result of the one-time screening was evaluated after linking the background population, the participants of the study and the invited but not participating cohort to the Swedish cause of death registry and the Swedish cancer registry. Estimating the possible cancer-specific mortality reduction using the Poisson regression model resulted in no difference in prostate cancer-specific mortality between the screened population and the unscreened population, IRR= 0.97 (0.71-1.23; 95% CI).

Paper II evaluated the association between the androgen DHT and prostate cancer incidence and mortality. High levels of DHT protected from lethal prostate cancer HR= 0.44 (0.25‐0.77; 95% CI), p=0.004 after 30 years of follow up. The association remained significant both for men seemingly heathy at time of inclusion HR=0.25 (0.07‐0.88; 95% CI), p= 0.032 and for those with a recently diagnosed cancer HR= 0.50 (0.26‐0.94; 95% CI), p=0.031.

In paper III the threshold for PSA was examined and the proportions of its isoforms – free/bound PSA that is indicative for low, or negligible risk for prostate cancer death. The associations between both PSA and the ratio free/bound PSA and lethal prostate cancer were strong at long-term follow up. A baseline PSA of 2 ng/mL or less combined with ratio free/bound PSA of 0.25 or greater indicated a very low long-term risk for prostate cancer death and further screening in this cohort can be abstained or continued with lower frequency.

In paper IV thawed serum from 330 men including 36 men with lethal prostate cancer was analysed. The aim was to estimate association between elevated levels of the enzyme Thymidine kinase (TK1), a phosphorylation enzyme important in DNA synthesis, and future risk for prostate cancer-specific mortality and overall mortality. The analyses were performed with a commercially available western blot kit. Preliminary estimates indicate that high levels of TK1 is associated with an increased risk for overall mortality irrespective of whether death occurred shortly after blood draw or after a period of follow up.