European Urology

3. Results

Our data are derived from a prescreened population of 965 men aged 64–77 and screened for the third time (n = 451), the fourth time (n = 502), or the fifth time (n = 12). One-third of the men had a history of previous negative biopsies. Of the 965 men, 721 (74.7%) were biopsied utilising either a PSA cut-off value of ≥3.0 ng/ml, following the ERSPC protocol, or a PCA3 score ≥10 with the intention of evaluating the performance characteristics of both markers in a setting where most men would present with a biopsy indication. A total of 122 PCa cases (16.9%) were detected.

Table 1 summarises the baseline characteristics for the total cohort and for men who were actually biopsied. The differences between the two cohorts are very small and have not been tested for statistical significance.

In Table 2, the distribution of biopsy indications by PCA3 (cut-off score: ≥10) and PSA (cut-off ≥3.0 ng/ml) are shown for the total cohort and for those men who were actually biopsied. With this algorithm, 87.5% (n = 844) of all men had a biopsy indication and 74.7% (n = 721) actually underwent a prostate biopsy.

Only 32 of the 721 men actually biopsied had a biopsy indication based on PSA alone, and 492 were based on PCA3 alone. Compliance with biopsy recommendations was very similar in all subgroups and varied between 84.3% and 88.3%. Correlation between the two markers was poor (ρ = 0.14; p < 0.0001).

Data on sensitivity and specificity of PSA (Table 3) are very much in line with published information 8 x I.M. Thompson, D.P. Ankerst, C. Chi, et al. Operating characteristics of prostate-specific antigen in men with an initial PSA level of 3.0 ng/ml or lower. JAMA. 2005;294:66-70 Crossref. , and 11 x R. Kranse, P.M.M. Beemsterboer, J.B.W. Rietbergen, D. Habbema, J. Hugosson, F.H. Schröder. Predictors for biopsy outcome in the European Randomized Study of Screening for Prostate Cancer (Rotterdam region). Prostate. 1999;39:316-322 Crossref. . A PSA cut-off value of 3.0 ng/ml, as utilised in the ERSPC study, identifies 35.2% of all PCa cases and correctly identifies 69% of those men who do not have PCa. Compared with the recommended cut-off value of 35 for the PCA3 score, the sensitivity of PSA turns out to be higher; the specificity is comparable to the PSA cut-off value of 2.0 ng/ml. No cut-off value can be identified that matches a high sensitivity and specificity, with the possible exception of the PCA3 score of 35. To maximise detection to >80% of the detectable PCa, one would have to choose a PSA cut-off of >1.0 ng/ml and a PCA3 score >20. The areas under the curve (AUCs) are indicated in Fig. 1. Based on the ROC analyses, PCA3 performs marginally better than PSA (Fig. 1A; p = 0.143).

In Table 4, PCa detection with different PCA3 cut-off scores was compared with several PSA cut-off values. The low prevalence of serious disease (n = 19; 15.6%), of which almost half was present at low PSA values, reflects the prescreened study population. The positive predictive value (PPV) of the PCA3 test rises slightly with an increasing cut-off, contrary to PSA cut-offs, as can be expected in any PSA-based prescreened population. It is remarkable, however, that 62 of the 90 men (68.9%) with a PCA3 score ≥100 did not have a positive biopsy. A PCA3 score ≥20 identifies 18 of 19 serious PCa cases and avoids 26.1% of biopsies. A PCA3 score ≥35 misses five serious PCa cases, but 48.3% of all biopsies could have been avoided. It obviously remains unknown whether the missed serious cases still would have been curable and whether those cases might have escaped notice with all screening efforts anyway.

The data evaluating the performance of PCA3 in men with low PSA values and in men with a previous negative biopsy are presented in Table 5. If the small numbers could be considered trustworthy, 4 of the 11 serious PCa cases would be missed with a PCA3 score ≥35 and 10 would be missed with a PCA3 score ≥100. These results, together with the only marginal increase in the PPV with increasing PCA3 scores, do not support our working hypothesis that PCA3 might be useful in identifying aggressive cancers in the low PSA ranges, where only 13% of men were previously biopsied. However, it is encouraging to see rising PPVs with different PCA3 cut-off values and to see that most serious PCa cases are detected with a cut-off of 35.

Looking at 212 men who had undergone previous biopsy driven by a PSA ≥3.0 ng/ml, findings do not seem to be more encouraging. The numbers, however, are small and need to be confirmed.