The PREMM(1,2,6) model predicts risk of MLH1, MSH2, and MSH6 germline mutations based on cancer history

Area of Science:

• Genetics
• Oncology
• Bioinformatics

Background:

• Hereditary colorectal cancer (CRC) is often linked to mutations in mismatch repair (MMR) genes.
• Accurate risk assessment for MMR gene mutations is crucial for early detection and prevention strategies.

Purpose of the Study:

• To develop and validate a predictive model (PREMM(1,2,6)) for estimating the risk of mutations in MLH1, MSH2, and MSH6.
• To incorporate personal and family cancer history into a risk prediction tool for hereditary cancer syndromes.

Main Methods:

• A multivariable polytomous logistic regression model was developed using data from 4539 individuals tested for MMR gene mutations.
• The model, PREMM(1,2,6), predicts the overall risk and gene-specific risk of MMR mutations.
• Model performance was validated using 1827 population-based colorectal cancer cases.

Main Results:

• Pathogenic mutations in MLH1, MSH2, and MSH6 were found in 12% of the cohort.
• Key predictors for MMR mutations included male sex, personal/family history of CRC, multiple CRCs, endometrial cancer, and extracolonic cancers.
• The model demonstrated strong discriminative ability with areas under the receiver operating characteristic curves ranging from 0.81 to 0.88 across different cohorts and genes.

Conclusions:

• The PREMM(1,2,6) model effectively estimates an individual's risk of carrying mutations in MLH1, MSH2, and MSH6.
• This web-based tool assists in assessing hereditary CRC risk and informs clinical management decisions.
• The model's validation confirms its utility in diverse populations for genetic counseling and cancer risk assessment.