Liquid biopsy at an NCI comprehensive cancer center: Patterns of utilization, performance, and outcomes.

Abstract

e20696 Background: Next generation sequencing is a fundamental diagnostic tool in non-small cell lung cancer (NSCLC). In tandem with tissue genomic testing (TGT), liquid biopsy (LB) has been more recently adopted in the detection of actionable genomic alterations (AGAs). We explore utilization patterns and performance of LB testing at an NCI Comprehensive Cancer Center. Methods: Retrospective chart review was performed on all NSCLC patients with LB between 2019 and 2024. Data reviewed included histology, stage, payor, sex, race, age, ECOG, and Charlson Comorbidity Index (CCI). Comprehensive TGT of 67 mutations, fusions and copy number variants was performed reflexively by institutional pathology. A registry of all diagnosed NSCLC patients without liquid biopsy (control cohort) was queried for propensity score (PS) matched controls. A total of 184 controls were matched to the LB cohort to evaluate for turnaround time (TAT) and time to treatment initiation (TTI), as well as overall survival (OS). PS matching method was performed with eight variables: age, histology, smoking status, race, gender, ECOG, provider (academic vs. community), and actionable mutation. OS was computed from date of NSCLC diagnosis to death. Survival distributions were estimated using Kaplan-Meier methods, and the log-rank test examined differences in OS between groups. Results: A total of 257 LB patients were captured and 755 were evaluated in the control cohort. When compared to control, LB patients tended to be younger (age 67 vs 69, p<0.0001), with better baseline functional status (CCI 8.6 vs 9, p<0.0001; ECOG 0-1 77.5% vs 57.9%, p<0.0001) and earlier stage disease (Stage I-III 10.6% vs 0.1%, p<0.0001). A total of 82 LB subjects had an AGA identified either through liquid or tissue NGS. Amongst tested subjects, 44 had matching AGAs on liquid and tissue sequencing, while 18 had non-matching AGAs. LB failed to detect an AGA in 32.4% of cases, while LB detected AGAs that were missed on TGT 11.1% of the time; most commonly EGFR 14.7% (5/34) and KRAS G12C 23.8% (5/21) mutants. Compared to TGT of PS controls, median TAT for LB was shorter (9 vs. 16 days, p <0.001) as was TTI (23 vs. 36 days, p <0.001). OS was significantly higher in the LB group amongst all subjects (LB group: 18.8 months, 95% CI 15.7-24.5 vs control: 6.9 months, 95% CI 5.7-8.4), and showed a trend towards higher median OS in the PS matched group (LB group: 18.5 months, 95% CI 14.5-24.5 vs PS control: 12.9 months, 95% CI 7.9-16.2). Conclusions: LB significantly improved TAT by 7 days and TTI by 13 days. Amongst patients with AGAs, liquid biopsy served to identify AGAs not seen on tissue in 11.1% of cases, most commonly EGFR and KRAS. When compared to PS controls there was a trend toward higher OS for patients who underwent LB. The use of LB paired with tissue testing in NSCLC work-up shows potential as a practice-changing tool to maximize identification of additional therapeutic options.