Objective: Because different PSA assays still show a wide inter-assay variation, we wondered what influence these discrepancies could have on the individual tumour characteristics of the cancers that each of these assays detect in a critical low PSA range. We analysed five different PSA assays in a biopsy simulation with PSA cut-offs of 3.0 and 4.0 ng/ml.
Materials and methods: Randomly taken samples of 360 men with prostate cancer and 96 with benign prostatic disease from a screened population with PSA range of 1.0-6.0 ng/ml (Tandem-E) were investigated. In all cases the diagnosis was confirmed by sextant biopsies. One hundred and thirty-seven men (38%) underwent radical prostatectomy. Variability amongst assays was illustrated in terms of missed cancers and unnecessary biopsies, and in terms of pathologic features of detected cancers at both PSA cut-offs.
Results: Compared to Tandem-E, all assays, except Access, showed significant differences in PSA measurements. Furthermore, none of the assays discriminated significantly between benign and malignant prostatic disease (p>0.05). Tandem-E and Elecsys lead significantly more frequently to the detection of cancers at the cost of more unnecessary biopsies compared to the other assays. Yet, at both PSA cut-offs the proportion of cancers with a certain pathologic grade or stage that were detected by each assay were approximately the same.
Conclusions: Our study shows that the use of different PSA assays only have consequences for the number, and not for the tumour characteristics of the prostate cancers that are detected. Thus, different PSA assays detect prostate cancers with the same tumour features.