Background: T-DXd is a HER2-targeted antibody drug conjugate approved for treatment of advanced HER2-positive breast and gastroesophageal cancers and HER2-low breast cancers. The prevalence of HER2 loss after exposure to T-DXd is unknown and has implications for treatment strategies in refractory patients. Methods: We investigated clinically reported HER2 immunohistochemistry (IHC) scoring on post-treatment tissue biopsies from patients who received at least 2 cycles of T-DXd as of 2/2023. IHC was performed using mAB clone 4B5 (Ventana) and scored by ASCO/CAP guidelines on a scale of 0 to 3+. MSK-IMPACT next generation sequencing (NGS) was performed on paired pre- and post-treatment samples when available. Statistics are descriptive. Results: A total of 62 patients with breast, gastroesophageal, or colon cancer had available post-treatment biopsies. The majority (n = 51) had breast cancer, including 32 with HER2-positive and 19 with HER2-low disease. Median time on therapy was 30 weeks in HER2-positive and 21 weeks in HER2-low breast cancer. All 32 patients with HER2-positive breast cancer had detectable HER2 expression by IHC on post-treatment biopsies (median IHC score 2+; range 1+ to 3+). Of those with HER2-low breast cancer, 12 (63.1%) patients had detectible IHC after treatment. Of the 7 with IHC scores of zero, 3 also had scores of zero on the most proximal pre-treatment biopsies. Seven patients with gastroesophageal and 3 with colorectal cancer were included, with a median time on therapy of 12 and 9 weeks respectively, of which 1 (9%) exhibited HER2-loss after treatment. Among all patients with HER2-positive cancers, the rate of complete HER2-loss by IHC after exposure to T-DXd was 2.3%. Thirty-two patients had paired genomic analysis, including 25 breast and 7 gastrointestinal cancers. No change was observed in the fraction of genome altered (p = 0.0736, q = .327) or tumor mutational burden (p = 0.139, q = .487) with T-DXd exposure. Changes in ERBB2 copy number did not show clear directionality, with 15 patients (46.7%) exhibiting ERBB2 amplification pre-treatment and 12 (37.5%) post-treatment, the sum of 3 temporal gains and 6 losses of ERBB2 amplification across the threshold of 1.8. Conclusions: HER2 remained detectable by IHC in the majority of patients treated with T-DXd in this cohort, especially those with HER2-positive cancers. These findings suggest that resistance to T-DXd may occur via target-independent factors, and that HER2 could still be exploited therapeutically in these populations. Further prospective studies using quantitative assays are needed to confirm these hypotheses.

*3 of these 7 patients had IHC scores of 0 pre-treatment.