Background: While EGFR mutant (EGFRm) non-small cell lung cancer (NSCLC) patients usually experience improved clinical benefit with EGFR TKIs, most eventually progress. Understanding mechanisms of resistance (MoR) may allow for more personalized treatment. Lazertinib is an irreversible third generation EGFR TKI for which MoR are unknown. Obtaining sufficient tumor tissue for genotyping at progression is often difficult. Therefore, we utilized plasma ctDNA from patients treated with lazertinib to explore MoR. Methods: Plasma samples from 47 NSCLC patients in the phase 2 trial of lazertinib (NCT03046992) were collected at screening and progressive disease (PD) and underwent ctDNA NGS of 74 genes using Guarant360. All patients were positive for an EGFR Ex19del or L858R (EGFRm) and T790M by tissue testing at screening. Acquired, nonsynonymous, characterized mutations detected in a PD sample but not in the screening sample from the respective patient were considered putative MoR, excluding aneuploidy. Patients with detectable plasma EGFRm and/or T790M at screening were evaluable. Results: ctDNA was detected in 47 (100%) screening samples and 43/45 (96%) PD samples (two failed sequencing). An EGFRm was detected in 85% of patients at screening (n = 40), 38 of which had PD ctDNA results and were included in analysis. T790M was detected in 30 patients at screening and subsequently not detected at PD in 21 of these patients, 55% of all 38 included patients. Among the ten patients with T790M detected at PD, on-target MoR were detected in 7 (18% of all included patients) including EGFR C797S (n = 3, 8%), EGFR amplification (n = 3, 8%), and EGFR T854A (n = 1, 3%). All C797S were in cis with T790M. No on-target MoR were detected in patients without T790M detected at PD. Off-target MoR were seen in 34% of patients (13/38) including mutations in PIK3CA (13%; 2 E545K, 2 E542K, 1 E81K), ERBB2 (5%; 1 D769H, 1 V777L), KRAS (3%; 1 G12C), and BRAF (3%; 1 G469A). Gene amplifications were detected in CCND1 (n = 1, 3%), CCNE1 (n = 2, 5%), ERBB2 (n = 1, 3%), FGFR1 (n = 1, 3%), MET (n = 4, 11%), and PIK3CA (n = 1, 3%), with some patients having multiple MoR. Conclusions: The spectrum of MoR identified in this cohort of patients treated with lazertinib is similar to that reported in other third generation EGFR TKIs, but with some differences in frequencies. The most common resistance mechanisms are T790M loss and PIK3CA alterations which may address the mechanism of action. Our findings suggest putative MoR of lazertinib and show that ctDNA NGS is an effective way to identify MoR in patients progressing on targeted therapy. Clinical trial information: NCT03046992.