Study designs of evaluations included in the review
The studies had to evaluate at least ten patients with cancer and at least ten noncancer control patients.
Specific interventions included in the review
Studies that evaluated free PSA assays were eligible. Studies that used only gamma-seminoprotein assays or the ratio of alpha-I-antichymotrypsin to total PSA were excluded. The included studies reported data for the free-to-total PSA ratio when the total PSA levels ranged from 4 to 10 ng/mL, 2.5 to 10 ng/mL, and 2.6 to 4.0 ng/mL. Assays were performed, where stated, on fresh specimens (4 studies) and specimens frozen for unreported time periods. The investigators selected optimal free-to total ratio cut points ranging from 8 to 27%. The investigators generally selected cut points to maximise sensitivity. Three studies used cut points selected to maximise accuracy, while one study used a cut point selected to maximise specificity.
Reference standard test against which the new test was compared
The inclusion criteria were not defined in terms of the reference standard test. The most appropriate reference standard was considered to be either radical prostatectomy or multiple systematic transrectal prostate needle biopsies with long-term clinical follow-up for men with negative biopsies. The included studies used needle biopsy (13 studies) or a combination of reference tests including radical prostatectomy (4 studies). No studies used long-term clinical follow-up to define the true negative results.
Participants included in the review
Patients with prostate cancer and patients with histologically confirmed noncancer were eligible. The studies used the tests in screening populations (6 studies), referred populations (most studies), or did not define the population tested.
Outcomes assessed in the review
Studies that provided data on the sensitivity and specificity were eligible. For each study where the investigators selected a single best free-to-total PSA ratio cut point, the likelihood ratios for positive and negative tests were calculated by the review authors, together with 95% confidence intervals (CIs) and the non-parametric trapezoidal area under the receiver operating characteristic (ROC) curve.
How were decisions on the relevance of primary studies made?
Two investigators reviewed all the titles and abstracts, and all potentially relevant articles were retrieved. The reviewers tried to reach consensus or used a third reviewer to referee.