Prostate specific antigen standardization: Making sense of the controversy
Efforts to standardize PSA assays were undertaken a decade ago, but PSA standardization remains controversial.
To improve diagnostic sensitivity and specificity in discriminating between benign prostatic hyperplasia and prostate cancer, different approaches have been proposed: PSA velocity, which measures time-related increase in PSA values; PSA density, which compares PSA levels with gland dimension evaluated by ultrasound; and free PSA, which measures the amount of PSA unbound to proteins such as chimotripsin.
PSA standardization, interchangeability
Significant variation in PSA results among non-equimolar assays was a major factor behind the desire for PSA assay standardization. Equimolar recognition of free and complexed PSA forms is critical to accurate PSA testing, especially at the 4.0 ng/mL cutoff: Inaccurate quantification of PSA concentrations at this critical level can result in false-positive or false-negative results.
In 1995, the molar extinction coefficient and PSA mass were recalculated with quantitative amino acid analysis by Stamey et al (Prostate 1995; 27:198-203). Stamey's work led to the proposal of a PSA standard (90:10 ratio of complexed to free PSA) to mitigate the non-equimolar response of some PSA assays. This, in turn, became the basis for the PSA definition adopted by the World Health Organization in 1999: WHO 96/670 for total PSA and WHO 96/668 for free PSA comprises 100% fPSA.
Developing an official definition of equimolarity for PSA is not easily achieved. In spite of the efforts to standardize PSA cutoff, the problem remains unsolved: in a recent study, Stephan et al note insufficient interchangeability among tPSA, fPSA, and %fPSA values obtained in patients with BPH and prostate cancer among five different commercial assays, all claimed as equimolars (Clin Chem 2006; 52:59-64). Stephan used as a reference method the Access Hybritech PSA because of "equimolarity well characterized." Values obtained by using the other methods ranged between 87% and 115%.
At the same time, Roddam et al reported on a study of 11 different assay methods using samples that were not from patients, but obtained with different water dilutions of the WHO tPSA and free PSA standards in samples from 200 laboratories participating in the UK National External Quality Assessment Service scheme (Ann Clin Biochem 2006; 43: 35-48). The authors concluded that none of the assays was without bias and none was equimolar. Debate surrounds the two studies, demonstrating that the problem of PSA standardization persists and raising the real possibility that clinicians using PSA values from different assays may perceive them as interchangeable.
