SUMMARY: Hereditary factors play an important role in the risk of developing several cancers. Therefore, identification of a germline predisposition can have important implications for treatment decision making, risk-reducing interventions, cancer screening for early diagnosis, and germline testing and targeted surveillance of unaffected relatives. Previously published studies have been biased by estimating the prevalence of germline cancer susceptibility in patients with breast, prostate, and colorectal cancer from registry populations, genetic testing companies, and high-risk cancer clinics.
With the widespread adoption of Next Generation Sequencing (NGS), multiple genes can be tested simultaneously (MultiGene Panel Testing-MGPT), rather than sequential single-gene testing, making MultiGene Panel Testing cheaper, faster and more efficient. Further, single-test multigene multiplexing strategy analyzes numerous cancer susceptibility genes and frequently detects highly penetrant, clinically actionable Pathogenic Germline Variants (PGV) in individuals whose clinical histories fail to fulfill syndrome-specific testing criteria. This is clinically relevant, as it has become increasingly complex to determine which individuals warrant germline testing.
Several risk assessment models have been developed to provide probability of an individual carrying a germline mutation. However, these models only provide syndrome-specific risk assessment for Lynch Syndrome, Hereditary Breast and Ovarian Cancer syndrome (HBOC)), etc., and there is significant need for a risk assessment model tailored toward MultiGene Panel Testing.
The PREMM (PREdiction Model for gene Mutations) model has been rigorously tested, and is widely recognized and recommended by several professional societies, including the National Comprehensive Cancer Network (NCCN), the American College of Gastroenterology, and the U.S. Multi-Society Task Force on Colorectal Cancer. The PREMMplus model is a clinical prediction algorithm (tool) that estimates the cumulative probability of an individual carrying a germline mutation in 19 genes linked to cancer. Individuals are considered to be high risk if they have a risk score greater than 2.5%, and are eligible for genetic evaluation to determine if they indeed harbor germline mutations, and these individuals in turn could benefit from measures to prevent the cancer, or detect cancer early.
This aim of this study was to develop and validate PREMMplus clinical risk assessment tool (clinical prediction model) that could be used to identify individuals who are likely to have Pathogenic Germline Variant and should undergo MultiGene Panel Testing. PREMMplus was designed to identify individuals carrying Pathogenic Germline Variants in 19 cancer susceptibility genes broadly categorized by phenotypic overlap and/or relative penetrance, and they included 11 Category A genes (APC, BRCA1/2, CDH1, EPCAM, MLH1, MSH2, MSH6, biallelic MUTYH, PMS2, and TP53) and 8 Category B genes (ATM, BRIP1, CDKN2A, CHEK2, PALB2, PTEN, RAD51C, and RAD51D). Assessment of germline variant pathogenicity was based on the most recent classification made by the clinical laboratory, performing MultiGene Panel Testing. This clinical prediction model was designed to achieve both high sensitivity and Negative Predictive Value (NPV) across a diverse spectrum of syndromes, used clinical data only, did not require tumor tissue thus facilitating scalability, and was adaptable to allow for future expansion, as new genes became incorporated into routine MultiGene Panel Testing.
Clinical predictors for this model included demographics (sex, ancestry, and age at testing), as well as personal, and family history of specific cancers in first- and second-degree relatives. EIGHTEEN cancer types were selected for PREMMplus development, including both common malignancies such as breast cancer and colorectal cancer and uncommon malignancies such as sarcoma and adrenocortical carcinoma, associated with inherited risk. Individuals were excluded from analysis if a personal and family history of any of these 18 cancers were not available, and/or if age at MultiGene Panel Testing was missing. Individuals with 2 or more Pathogenic Germline Variants were excluded from the development cohort.
The performance of this clinical model was validated in nonoverlapping data sets of 8,691 and 14,849 individuals with prior MultiGene Panel Testing ascertained from clinic and laboratory-based settings, respectively.
PREMMplus demonstrated high sensitivity and high Negative Predictive Value for identifying individuals with Pathogenic Germline Variants in the 19 different cancer susceptibility genes. PREMMplus demonstrated a sensitivity of 93.9%, 91.7%, and 89.3% and Negative Predictive Value of 98.3%, 97.5%, and 97.8% for identifying Category A gene Pathogenic Germline Variants carriers, in the development and validation cohorts, respectively. PREMMplus demonstrated a sensitivity of 89.9%, 85.6%, and 84.2% and Negative Predictive Value of 95.0%, 93.5%, and 93.5% for identifying Category A/B gene Pathogenic Germline Variants carriers in the development and validation cohorts, respectively. Overall, 9.4%, 10.8%, and 9.2% of the development, clinic-based validation, and laboratory-based validation cohorts, respectively, harbored a Pathogenic Germline Variant in one of the 19 PREMMplus genes.
It was concluded that PREMMplus accurately identifies individuals with Pathogenic Germline Variants in a diverse spectrum of cancer susceptibility genes, with high sensitivity and Negative Predictive Value. PREMMplus represents a new evidence-based approach and can be used to identify individuals who should undergo MultiGene Panel Testing.
Development and Validation of the PREMMplus Model for Multigene Hereditary Cancer Risk Assessment. Yurgelun MB, Uno H, Furniss CS, et al. DOI: 10.1200/JCO.22.00120 Journal of Clinical Oncology