Background: Currently approved epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) are ineffective in patients (pts) with EGFR exon 20 insertion NSCLC. TAK-788 is an EGFR TKI with potent and selective preclinical inhibitory activity against EGFR exon 20 insertions, and has demonstrated preliminary efficacy in a single-arm phase 1/2 clinical trial (NCT02716116). We performed an indirect comparison of real-world outcomes with clinical trial data for this subset of pts to determine whether TAK-788 provides superior efficacy over standard treatment options. Methods: We compared efficacy in pts with refractory NSCLC with EGFR exon 20 insertions treated with TAK-788 160 mg qd (1–7 prior lines) from the ongoing clinical trial (data cut Mar 1, 2019) vs real-world data (RWD) in the second-line setting from the US Flatiron Health electronic health record–derived database (Jan 2011‒Jun 2018). This analysis was conducted using an unadjusted data set, as well as by applying propensity score modeling with inverse probability of treatment weighting (IPTW) to adjust for group differences in key baseline characteristics. Progression-free survival (PFS) and objective response rate (ORR) were compared between groups. Results: A total of 99 pts were included, n=28 TAK-788 and n=71 RWD; mean age 62/65 y; male 25%/46%; Asian 18%/10%; former smoker 39%/45%; brain metastases 43%/34%. In the RWD, there was no consistent regimen for second-line treatment (including 29.6% immuno-oncologic agents, 25.4% EGFR TKI, 10% docetaxel). Baseline characteristics were comparable after weighting. PFS and ORR showed statistically significant improvements with TAK-788 vs RWD (Table). Specifically, after weighting, median PFS for TAK-788 vs RWD is 7.3 vs 3.5 mo, and ORR is 43% vs 13%. Conclusions: Despite a more heavily pretreated pt population, the efficacy of TAK-788 in pts with refractory NSCLC with EGFR exon 20 insertions appears better than other second-line treatment options used in the real-world setting. Clinical trial information: NCT02716116.

^aPer investigator RECIST 1.1 for TAK-788 and clinician-reported tumor growth for RWD. ^bCox regression model with IPTW. ^cLogistic regression model with IPTW. HR, hazard ratio; NE, not estimable; OR, odds ratio.