Predictive models for mutations in mismatch repair genes: implication for genetic counseling in developing countries

Area of Science:

• Genetics
• Oncology
• Bioinformatics

Background:

• Lynch syndrome (LS) is the primary inherited cause of colorectal cancer (CRC), responsible for 2-5% of cases.
• LS arises from mutations in DNA mismatch repair genes (MLH1, MSH2, PMS1, PMS2, MSH6).
• Predictive models assist in identifying individuals for genetic testing, crucial in resource-limited settings.

Purpose of the Study:

• To evaluate the sensitivity and specificity of five predictive models for detecting germline mutations in individuals with suspected Lynch syndrome.
• To compare the performance of PREMM, Barnetson, MMRpro, Wijnen, and Myriad models in a clinical setting.

Main Methods:

• Sequencing of MLH1, MSH2, and MSH6 genes in 88 patients with suspected LS.
• Calculation of mutation detection probability using five distinct predictive models.
• Construction of receiver operating characteristic (ROC) curves to assess model performance (sensitivity and specificity).

Main Results:

• Thirty-one mutations were identified among 88 patients: 16 in MSH2 and 15 in MLH1; no mutations were found in MSH6.
• The Area Under the Curve (AUC) for PREMM (0.846), Barnetson (0.850), MMRpro (0.821), and Wijnen (0.807) showed no significant statistical difference.
• The Myriad model exhibited a lower AUC (0.704) compared to the other four models.

Conclusions:

• Barnetson, PREMM, MMRpro, and Wijnen models demonstrate comparable performance in predicting Lynch syndrome-associated mutations.
• The Myriad model's predictive capability is statistically inferior to the other evaluated models.
• These findings support the clinical utility of specific predictive models for Lynch syndrome genetic testing.