Molecular residual disease (MRD) analysis from the ADAURA trial of adjuvant (adj) osimertinib in patients (pts) with resected EGFR-mutated (EGFRm) stage IB–IIIA non-small cell lung cancer (NSCLC).

8005 Background: Osimertinib (osi) is a third-generation, central nervous system-active EGFR-TKI, that potently and selectively inhibits EGFR-TKI sensitizing and EGFR T790M resistance mutations. Adj osi (3 years [yrs]) is recommended for resected EGFRm stage IB–IIIA NSCLC, based on significant improvements in disease-free survival (DFS) and overall survival (OS) in the Phase III ADAURA study (NCT02511106). A trend towards an increased DFS event rate beyond 3 yrs suggests some pts may benefit from longer adj osi treatment (tx). We explored if plasma ctDNA-based, tumor-informed MRD could predict disease recurrence. Methods: Eligible pts (≥18 yrs [≥20 in Japan/Taiwan], WHO PS 0/1 with completely resected EGFRm [Ex19del/L858R] stage IB, II or IIIA [AJCC 7th edition] NSCLC) were randomized 1:1 to osi 80 mg once daily or placebo (pbo) until disease recurrence, tx completion (up to 3 yrs), or a discontinuation criterion was met. Personalized MRD panels (RaDaR, NeoGenomics) were used, comprised of ≤50 tumor-specific variants, based on whole exome sequencing of resected tumor tissue (variants identified in germline DNA removed). Plasma sample collection: baseline (BL; at randomization surgery and adj Ctx, if received), on tx (every 12 weeks [wks]), tx discontinuation, and post-tx completion (wk 12, wk 24, then every 24 wks until 5 yrs). Plasma samples were analyzed for detection of ctDNA (MRD+). Molecular recurrence or DFS (MRD/DFS) was defined as time from randomization to post-BL MRD+, disease recurrence, or death and was compared across arms. DFS was investigator-assessed. Results: Of 682 pts randomized, 245 (36%) had samples required to produce MRD panels, which were evaluable for 220 (32%) pts across both arms. In the osi vs pbo arms, 5/112 (4%) vs 13/108 (12%) pts were MRD+ at BL; 4/5 pts became MRD– during osi tx vs 0/13 pts on pbo. On tx, MRD detection had clinical sensitivity of 65% (62 MRD+/96 DFS+), specificity of 95% (118 MRD–/124 DFS–) and preceded a DFS event by a median (95% CI) of 4.7 (2.2, 5.6) months (mos) across both arms. Median follow-up time from randomization was 44.2 (95% CI 42.4, 49.1) and 19.1 (95% CI 11.1, 28.3) mos for osi and pbo arms, respectively. Overall, 86% (95% CI 78, 92) vs 36% (95% CI 27, 45) pts in the osi vs pbo arms were MRD/DFS event free at 36 mos (HR: 0.23; 95% CI 0.15, 0.36). MRD/DFS events in the osi arm were detected at any time post-BL in 28 (25%) pts, of whom 19/28 (68%) had events post-adj osi. Most (11/19 [58%]) MRD/DFS events detected occurred within 12 mos of osi tx completion. Conclusions: MRD+ preceded DFS events in most pts with a median lead time of 4.7 mos across both arms. MRD– was maintained for most pts during adj osi tx with the majority of MRD/DFS events occurring after osi tx completion. MRD detection could potentially identify a subset of pts likely to benefit from longer adj osi. Clinical trial information: NCT02511106 .