MedWire News: The inclusion of prostate cancer antigen 3 (PCA3) measurements in prostate biopsy prediction models gives a good degree of accuracy, especially in patients whose prostate-specific antigen (PSA) levels fall in the "grey area" below 10 ng/ml, report researchers.
The team adds that unnecessary biopsies can be avoided when the decision is based on two such models, without a substantial risk for missing prostate cancer.
"PCA3 holds much promise when it comes to deciding whether to biopsy," explain Riccardo Autorino (Second University of Naples, Italy) and colleagues, who compared the predictive ability of two nomograms that include PCA3 - the Prostate Cancer Prevention Trial (PCPT) calculator and Chun's nomogram - in a cohort of 218 men with PSA levels below 10 ng/ml.
The PCPT calculator is available online and comprises PCA3 plus six other variables, write Autorino and team in European Urology. These are: age, PSA, family history, digital rectal examination (DRE), prior biopsy, and prostate volume. The Chun nomogram includes age, PSA, DRE, prior biopsy, and prostate volume, in addition to PCA3.
Overall, both models predicted similar risks for the need for prostate biopsy, at 45% using the PCPT calculator and 46% with Chun's nomogram. This risk was significantly higher among the 73 men who actually harbored cancer on biopsy compared with those who did not, at 54% versus 39% for the calculator, and 54% versus 41% for the nomogram.
Receiver operating characteristic curve analysis revealed that the PCPT calculator had a significantly higher area under the curve score - where 100% equals perfect discrimination - compared with Chun's nomogram, at 80% versus 72%.
However, Chun's nomogram showed better overall calibration and a higher net benefit on decision curve analysis, say the researchers.
The team also calculated the probability of unnecessary biopsy and potential cancers missed using a threshold for biopsy of 25% for both predictive models. Using this cutoff, the PCPT calculator would save 11% of biopsies while missing no cancer, while Chun's nomogram would save double the amount of biopsies (22%) but at the cost of missing 4.1% low- and intermediate-grade cancers.
"Considering all of the points above, we believe both risk estimators to be of added value when considering the issue of biopsy in patients presenting in the grey area of PSA below 10 ng/ml," conclude Autorino et al.
MedWire (www.medwire-news.md) is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2010
Free full text