Universal germline testing for cancer susceptibility and actionable noncancer disorders among 19,842 patients: Initial findings from the City of Hope INSPIRE study.

Abstract

10594 Background: Germline testing can be essential for risk stratification and therapeutic decision making. However, guideline-based testing misses ~50% of people with inherited cancer risk. To address this gap, we launched the INSPIRE study, offering universal germline testing (31 non-cancer and 154 cancer genes) to consented patients affected and unaffected with cancer. Methos: In 2020, we initiated a scalable in person/remote consent model without formal pre-test counseling. Consenting patients opted-in for one, both, or neither germline test. Research staff placed orders. Results were reviewed by a genetic counselor. Pathogenic/likely pathogenic (P/LP) and increased risk allele (IRA) results were disclosed by phone and scheduled for a genetic consult. Variant of uncertain significance (VUS)/negative results were disclosed by letter with a genetic consult available on request. Descriptive statistics and chi-squared tests evaluated a priori hypotheses regarding potential factors impacting genetic testing uptake and actionability. Results: We approached 26,920 patients: 22,115 (82%) enrolled, 1,830 (7%) declined, 2,975 (11%) deferred a decision. Consented patients opted for testing for germline cancer (21,871/ 99%), non-cancer (21,731/98%), or neither (160/0.7%). We have returned results to 19,842 patients: mean age 62, 69% female; 72% white, 14% Asian, 5% Black, 9% other; 24% Hispanic. Most patients (15,578) had cancer, 13% with multiple primaries. Among the 19,842 patients, 4,456 (23%) had at least one P/LP/IRA variant; 3,971 (20%) in a cancer risk gene and 589 (3%) in a non-cancer gene. Variant proportion by category was 1,466 (30%) high-penetrance, 847 (18%) moderate-penetrance, 1,127 (23%) low-penetrance, 1,282 (28%) recessive, and 34 (0.7%) preliminary evidence. Penetrance prevalence differed by sex: high penetrance variants were more common in females (33% vs. 27%; p=<0.001) and moderate penetrance variants were more common in males (21% vs. 17%; p=0.003). 16,883 (85%) individuals carry at least one VUS. Conclusions: A scalable model to increase germline genetic testing among unselected cancer and unaffected patients across >20 sites was highly feasible. A significant proportion of patients had actionable findings in cancer and non-cancer genes. Next steps include analyses to establish variant prevalence across groups, a longitudinal assessment of psychological outcomes and an evaluation of the impact of result disclosure on cost, patient care, and outcomes.