Communication With Clinicians and Relatives About Cascade Genetic Testing in Cancer Patients With Germline Pathogenic Variants

PURPOSE Germline testing is underutilized and varies by cancer diagnosis. We hypothesized that patient and clinician involvement in cascade testing of relatives varies by the cancer susceptibility (breast v gastrointestinal GI) of the affected gene. METHODS All patients diagnosed with cancer in Georgia or California during 2018-2019 and reported to SEER registries, who were linked to a pathogenic variant (PV) result from testing laboratories, were surveyed 4.5 years postdiagnosis about (1) clinician involvement in result communication to relatives, (2) attitudes about result communication, (3) result communication to relatives, and (4) relatives' testing. PVs were categorized by primary association with breast ( BRCA1, BRCA2, ATM, BARD1, BRIP1, CDH1, CHEK2, PALB2, PTEN, RAD51C, RAD51D, STK11, TP53 ) or GI cancers ( MLH1, MSH2, MSH6, PMS2, EPCAM, APC, BMPR1A, CDKN2A, GREM1, POLD1, POLE, SMAD4 ). RESULTS A total of 4,080 patients were surveyed; 2,183 responded (53.5%). Most had PVs associated with breast (83.4%) versus GI cancers (16.6%). Most (85.0%) reported a genetic counseling visit. Genetic counselors were most involved in encouraging family communication (71.5%, v oncologists 33.4%, surgeons 19.4%), advising how to talk with relatives (55.0%, versus oncologists 18.0%, surgeons 9.1%), and talking directly with relatives (33.4%, v oncologists 12.6%, surgeons 7.0%). Most patients considered sharing results their responsibility (86.8%); they notified 86.8% of first-degree and 44.9% of second-degree relatives. Nearly one third (29.0%) of patients reported that no relative was tested, whereas 18.3% reported that four or more relatives were tested. Outcomes did not differ by affected gene or cancer type ( P > .1). CONCLUSION Patients with cancer are motivated to communicate PV results to relatives. However, few clinicians are involved and relatives' testing remains low. Novel care delivery models are needed to advance cascade testing and precision risk reduction.