Redefining early-onset cancer and risk of hereditary cancer predisposition.

Abstract

10510 Background: For most solid tumors, early age of onset, a marker for potential increased inherited risk, has been arbitrarily defined as a diagnosis of cancer at age < 50. However, this disregards the variation in mean age of diagnoses across solid tumors resulting in the potential misclassification of patients at increased risk of germline alterations. Methods: 25,305 prospectively ascertained cancer patients, representing 32 solid tumor types, underwent sequencing analysis of up to 90 cancer predisposition genes using next-generation genetic analysis with reporting of germline likely pathogenic (LP) and pathogenic variants (P). Mean age of diagnosis by cancer type, with standard deviations (STD), were calculated according to Surveillance, Epidemiology, and End Results Program (SEER) data. Early-onset (EO-CA), average-onset (AO-CA), and late-onset (LO-CA) cancers were defined as cancers diagnosed < -1 STD, -1 < STD > 1, and > 1 STD around the mean age at diagnosis for the specific cancer type, respectively. Prevalence of LP/P variants was classified according to age of onset. Statistical analyses were performed using STATA. Results: Among 25,305 patients with 32 solid tumor types classified as per SEER data, 31% (n = 7,854) of patients were categorized as having EO-CA. Overall prevalence of germline alterations varied as a function of age at diagnosis with 19.2%(n = 1509), 15.9% (n = 2588/16,299), and 12.9% (149/1152) of EO-CA, AO-CA, and LO-CA patients harboring a germline LP/P variant (p < 0.0001). The enrichment of variants in EO-CAs was driven by high- and moderate-penetrance genes with 13.2%, 9.2% and 5.1% of EO-CA, AO-CA, and LO-CA patients having an LP/P variant (p < 0.0001); low-penetrance germline alterations were static across the age groups (6%, 6.7%, 7.8%). An age cutoff of 50 for early-onset cancer was > 1 STD below the mean age at diagnosis in 20 cancer types resulting in 2,226 patients who would not have been classified as having EO-CA inclusive of 394 patients who tested positive for a germline LP/P, representing 26.1% of all germline EO-CA positives. Conclusions: Early-onset cancer should be defined taking into consideration the mean age at diagnosis for a specific cancer type as an arbitrary cutoff of age < 50 misses over 25% of patients with an inherited cancer predisposition syndrome. Due to the high prevalence of inherited predisposition syndromes, genetic testing of all patients with EO-CA, irrespective of tumor type, is warranted.