Skip to main content
Premium Trial:

Request an Annual Quote

Phenome-Wide Association Study Links Hereditary Cancer Genes to Additional Phenotypes

NEW YORK — Pathogenic variants in hereditary cancer genes can affect other malignant, but also nonmalignant, traits, a new phenome-wide association study has found.

Examining additional phenotypes associated with hereditary cancer genes could lead to a better understanding of disease etiology and improved treatments, researchers from the National Human Genome Research Institute and elsewhere said.

As they reported Thursday in JAMA Oncology, the researchers conducted a PheWAS of genetic and phenotypic data from more than 214,000 individuals, replicating known gene-phenotype associations and uncovering new ones. These new associations were not only with other neoplastic conditions but also with a range of other traits like chronic gastritis or pancreatitis.

"These findings suggest that PheWAS in EHR data sets has the potential to expand our knowledge of the phenotypes and disease processes in patients with pathogenic and likely pathogenic variants in hereditary cancer genes," senior author Joshua Denny from NHGRI and colleagues wrote in their paper.

Using data from three cohorts — the Electronic Medical Records and Genomics Sequencing cohort, the UK Biobank, and the Hereditary Cancer Registry — the researchers examined associations between germline pathogenic or likely pathogenic variants in about two dozen cancer-related genes and different phenotypes. The analysis included genes like BRCA1, BRCA2, and PTEN as well as genes like ATM, CHEK2, and PALB and drew upon ICD-9-CM and ICD-10-CM codes from participants' health records as phenotypes.

Through their analysis the researchers confirmed 38 of the 38 primary gene-cancer associations and nearly 70 percent of gene-phenotype associations documented in the Online Mendelian Inheritance in Man database. They further identified 19 new associations not in OMIM, which affected 13 hereditary cancer genes.

The novel associations include new ties to other neoplastic conditions. Pathogenic or likely pathogenic variants in CHEK2, for instance, were associated with both leukemia and plasma cell neoplasm, and pathogenic or likely pathogenic variants in ATM were linked to gastric cancer and pancreatic cancer. They additionally noted that biallelic pathogenic or likely pathogenic variants in MUTYH were associated with kidney cancer.

Other associations, meanwhile, were with nonmalignant conditions. Pathogenic or likely pathogenic variants in BRCA1 and BRCA2 were associated with ovarian cysts, while PTEN variants were linked to chronic gastritis and MEN1 variants with acute pancreatitis. These associations held, the researchers noted, after removing individuals with prior cancer diagnoses from the analysis.

Through further investigation of participants' electronic health records and other data, the researchers found that some of these associations could stem from symptoms of underlying disease. For instance, the researchers noted a link between pathogenic variants in RET and diplopia but suggested that that double vision could be due to neuroendocrine disorders including tumors.

Additional studies are needed to confirm these new associations, Denny and colleagues noted. Still, they said their findings indicate that PheWAS could help uncover additional phenotypes associated with variants in hereditary cancer genes.

In a related commentary in JAMA Oncology, Kailin Yang and Jacob Scott, both from the Cleveland Clinic, said that the impact of this study may extend past its expansion of phenotypes associated with pathogenic variants in hereditary cancer genes. The study, they added, showed that new information could be gleaned by integrating sequencing and electronic health record data.

Such an approach could enable the inclusion of more diverse populations of research participants, they said. More than half the participants in Denny and colleagues' analysis were women, though they noted that the cohorts included had varied representation of non-European ancestry groups, ranging from 21.9 percent to 6 percent.

"Developing strategies to involve patients from a diverse ethnic background and apply these in low- and middle-income countries will greatly expand the impact of PheWAS on a global scale," Yang and Scott added.