The efficacy of trastuzumab deruxtecan after HER2-TKI exposure in HER2 exon 20 insertion–positive non–small cell lung cancer: Results from a large-scale nationwide genomic screening (LC-SCRUM-Asia).

8531 Background: Although reduced efficacy of HER2-TKIs after trastuzumab deruxtecan (T-DXd) has been reported in HER2 exon 20 insertion–positive non–small cell lung cancer (NSCLC), the efficacy of T-DXd following prior HER2-TKI exposure remains unclear. Methods: We identified patients with HER2 exon 20 insertion–positive NSCLC treated with T-DXd from two cohorts: the National Cancer Center Hospital East (N = 26,971) and the LC-SCRUM-Asia (N = 20,902). Clinical outcomes (ORR, DCR, and PFS) were evaluated across three groups based on prior HER2-TKI therapy: (1) HER2-TKI-naïve patients (naïve cohort), (2) those treated with selective HER2-TKIs (zongertinib or severtinib) (selective TKI cohort), and (3) those treated with non-selective pan-HER2-TKIs (such as poziotinib, afatinib, and others) (non-selective TKI cohort). Results: Of 77 patients (0.2%) included, 60 in the naïve cohort, 7 in the selective TKI cohort, and 10 in the non-selective TKI cohort. Overall, the median age was 60 years (range, 44–76), and 56% of patients were female. By NGS analysis, all tumors harbored ERBB2 exon 20 insertions within the tyrosine kinase domain; the YVMA subtype was the most common (57%). T-DXd was administered as a median fourth-line therapy (range, 2–8). Baseline clinical and genomic characteristics were generally comparable across the three groups. According to prior HER2-TKI exposure, the ORR was 60% (36/60) in the naïve cohort, 29% (2/7) in the selective TKI cohort, and 60% (6/10) in the non-selective TKI cohort. The DCRs were 83% (50/60), 71% (5/7), and 80% (8/10), respectively. Median PFS with T-DXd was 9.9 months in the naïve cohort, compared with 6.5 months in the selective TKI cohort and 5.3 months in the non-selective TKI cohort. The 6-month PFS rate was 72%, 71%, and 50%, respectively. Among 10 patients who did not achieve an objective response to prior HER2-TKIs (best response of stable or progressive disease), 6 achieved a partial response with subsequent T-DXd. Conclusions: Although prior HER2-TKI exposure, particularly selective HER2-TKIs, may attenuate the efficacy of T-DXd in HER2 exon 20 insertion–positive NSCLC compared with HER2-TKI–naïve patients, T-DXd retained clinically meaningful activity in a subset of patients, including those who did not achieve objective responses to prior HER2-TKIs.