Cancer Research Lab

A prediction model designed for use in men with low volume/low grade prostate cancer has been successfully validated, researchers report in the March issue of the Journal of Urology.

Early detection protocols may lead to overdiagnosis and overtreatment of tumors with volumes less than 0.5 cc, the absence of Gleason grade 4 or 5 and with organ-confined disease, Dr. R. J. Babaian of the University of Texas MD Anderson Cancer Center, Houston and colleagues note.

To evaluate these patients, the investigators previously developed a prediction model that assessed the number of positive cores and tumor length in a core. The Gleason score and prostate volume significantly enhanced the accuracy of this model for low-volume, low-grade cancer in men who had undergone extended biopsy.

In the current study, the researchers tested the model in an independent population of 170 men with prostate cancer who underwent radical prostatectomy without neoadjuvant therapy. In all cases, prostate cancer was diagnosed in only 1 positive core of a 10-core extended biopsy.

The model, which employed a tumor length of less than 2 mm, a Gleason score of no more than 4 and a prostate gland volume of more than 50 cc, had a positive predictive value of 70.4%, a negative predictive value of 71% and a diagnostic accuracy of 70.6%.

This model did better than one based on tumor length alone or another using both tumor length and Gleason score.

In light of these findings, the researchers conclude that the model can be used as a tool in clinical practice “to select men for active surveillance if an extended biopsy strategy is performed and only 1 core is positive for cancer.”