To compare the clinical performance of original and revised Bethesda guidelines for the detection of MSH2/MLH1 gene carriers in patients with colorectal cancer.METHODS
A total of 1,222 patients with newly diagnosed colorectal cancer were included in the EPICOLON study, a prospective, multicenter, nationwide epidemiology survey aimed at establishing the incidence of HNPCC in Spain (JAMA 2005; 293:1986–1994). Performance characteristics of the original and revised Bethesda guidelines were assessed with respect to the presence of MSH2/MLH1 germline mutations. Logistic regression analysis was performed to establish the most effective strategy.RESULTS
Original or revised Bethesda guidelines were equivalent strategies in terms of sensitivity (100%vs 100%; ns), specificity (98.1%vs 97.9%; ns), and overall accuracy (98.1%vs 97.9%; ns), as well as positive (25.8%vs 24.2%) and negative predictive values (100%vs 100%). The most discriminating individual variables were criteria number 1 (i.e., fulfillment of the Amsterdam criteria; RR = 34.14; 95% CI = 6.85–170.16; p < 0.001) and number 2 (i.e., individuals with two HNPCC-related neoplasms; RR = 35.63; 95% CI = 4.83–262.6; p < 0.001) of the original guidelines, and criterion number 1 of the revised guidelines (i.e., colorectal cancer diagnosed under 50 yr of age; RR = 29.34; 95% CI = 3.81–225.96; p= 0.001). The aggregation of these three criteria was equivalent to both Bethesda guidelines in terms of sensitivity (100%) and negative predictive value (100%), but superior to the revised criteria regarding specificity (98.5%; p < 0.05), overall accuracy (98.5%; p < 0.05), and positive predictive value (30.8%).CONCLUSIONS
Original and revised Bethesda guidelines are equivalent, highly effective criteria for the identification of MSH2/MLH1 gene mutation carriers in patients with newly diagnosed colorectal cancer. A new set of recommendations, based on a combination of some of their individual criteria, may provide additional advantages in terms of effectiveness.