Assessing the predictive accuracy of hMLH1 and hMSH2 mutation probability models

Area of Science:

• Genetics
• Oncology
• Medical Diagnostics

Background:

• Hereditary Nonpolyposis Colorectal Cancer (HNPCC) increases susceptibility to colorectal and extra-colonic cancers.
• Existing guidelines for HNPCC family identification lack sufficient sensitivity and specificity.
• The Wijnen pre-test probability model and Myriad Genetics Laboratory prevalence table were developed to improve HNPCC detection accuracy.

Purpose of the Study:

• To evaluate the predictive accuracy of the Wijnen model and Myriad table for HNPCC genetic mutations.
• To compare the performance of these models against existing guidelines in identifying individuals with HNPCC.

Main Methods:

• Analysis of 49 patients who underwent genetic testing for hMLH1/hMSH2 germline mutations.
• Evaluation of the revised Bethesda guidelines, Wijnen model, and Myriad table performance metrics (sensitivity, specificity).
• Assessment of predictive accuracy using receiver operator characteristic (ROC) curves.

Main Results:

• Revised Bethesda guidelines showed high sensitivity (94.4%) but low specificity (12.9%) for germline mutations.
• At a 10.0% mutation probability threshold, Wijnen model and Myriad table had sensitivities of 55.6% and 60.0%, respectively.
• Specificities for the Wijnen model and Myriad table were 54.8% and 23.8%, respectively; both were poor predictors of mutation prevalence (ROC AUCs 0.616 and 0.400).

Conclusions:

• Neither the Wijnen model nor the Myriad table demonstrate adequate sensitivity or specificity for HNPCC genetic testing decisions.
• Current predictive models and guidelines require improvement for accurate HNPCC diagnosis.
• Further research is needed to develop more reliable tools for identifying individuals eligible for HNPCC genetic testing.