Skip to main content

28-10-2009 | Oncology | Article

Epstein criteria predict organ-confined but not insignificant prostate cancer


Free abstract

MedWire News: Radical prostatectomy (RP) patients have a high likelihood of organ-confined disease and 5-year freedom from disease recurrence if their initial biopsy results fit the Epstein criteria, study findings show.

However, the Epstein criteria cannot accurately predict the presence of biologically insignificant disease in these men, using either a classical or liberal definition of insignificant disease.

The criteria were designed to “give prostate-specific antigen (PSA)-based and needle biopsy-based criteria for identifying potentially biologically insignificant tumors that might be safely managed by initial surveillance,” write Michael Lee, from the Glickman Urological and Kidney Institute in Cleveland, Ohio, USA, and colleagues.

The researchers correlated the initial biopsy findings of 268 men with the pathologic features of their RP specimens to evaluate the accuracy of the Epstein criteria at predicting insignificant disease.

The Epstein criteria include a PSA density of less than 0.15 ng/ml, a biopsy Gleason score of 6 or less, the presence of tumor in two biopsy cores or less, and no more than 50% tumor involvement in any core.

Insignificant disease was defined classically as organ-confined disease with a Gleason score below 6 and tumor volume less than 0.5 cm3, or liberally as organ-confined disease with a Gleason score below 6, with any volume tumor.

The presence of any volume organ-confined tumor, and biochemical relapse-free survival (bRFS) rates were also examined.

In all, 51% of patients met the Epstein criteria, 62% had organ-confined disease, and classically and liberally defined insignificant disease was present in 25% and 34% of patients, respectively.

More patients who met the Epstein criteria had insignificant disease using either definition than who did not meet the criteria (48.1% vs 23.5%).

The researchers calculated the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for each endpoint.

Sensitivity and NPV were best using the classical definition of insignificant disease (74% and 86%, respectively), while specificity and PPV were best for organ-confined disease (74% and 92%, respectively). The researchers believe these varied results indicate the criteria are not “robust enough” to be used to predict biologically insignificant disease.

The 1- and 5-year probabilities of bRFS in patients who met the Epstein criteria were both 100% compared with 98% and 83% for those who did not meet the criteria.

Lee et al suggest in the journal European Urology that “men who do meet the Epstein criteria have an excellent chance of having organ-confined disease and minimal risk for biochemical failure within 6 years of RP, suggesting that most such men are curable at diagnosis.”

However, they conclude: “Use of these criteria does not… exclude the potential for biologic progression using a strategy of active surveillance and delayed intervention.”

MedWire ( is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2009

By Sarah Guy

Related topics