Trastuzumab deruxtecan (T-DXd), an antibody-drug conjugate (ADC), demonstrated efficacy in various cancers, including gastric cancer. The DESTINY-Gastric01 trial showed T-DXd's superiority over chemotherapy in HER2-positive gastric cancer, leading to its regulatory approval. However, HER2 expression heterogeneity in gastric cancer poses challenges in patient selection. This study aimed to identify biomarkers predictive of T-DXd response using circulating tumor DNA (ctDNA) and tissue samples from the DESTINY-Gastric01 trial, exploring potential mechanisms of resistance. The spatial and temporal heterogeneity of HER2 expression in gastric cancer, unlike the more consistent expression seen in HER2-positive breast cancer, complicates the identification of patients likely to benefit from T-DXd. This heterogeneity, along with the potential for decreased HER2 expression after prior HER2-targeted therapies (observed in 16–32% of patients), necessitates exploring non-invasive methods like ctDNA analysis to identify suitable candidates for T-DXd treatment. The relationship between ctDNA alterations and the efficacy of ADCs targeting specific oncoproteins remains understudied.

Previous research has established the efficacy of T-DXd in various cancers, including gastric cancer, as shown in clinical trials like DESTINY-Gastric01. These studies demonstrated a significant improvement in objective response rate (ORR) and overall survival (OS) in patients with HER2-positive gastric cancer treated with T-DXd compared to standard chemotherapy. However, the inherent heterogeneity of HER2 expression in gastric cancer and the potential for acquired resistance have prompted investigations into potential predictive biomarkers. Studies exploring ctDNA analysis for identifying patients likely to respond to HER2-targeted therapies in gastric cancer are still emerging, highlighting the need for further research in this area. The use of ctDNA analysis provides a minimally invasive approach to monitor and assess treatment efficacy and identify potential resistance mechanisms in real-time. Prior research has established the role of HER2 IHC/ISH, but there remains a need to explore other factors that may influence a patient’s response to T-DXd.

This exploratory biomarker analysis involved patients from the DESTINY-Gastric01 trial's primary (HER2-positive) and exploratory (HER2-low) cohorts. Baseline HER2 status was assessed via immunohistochemistry (IHC) and in situ hybridization (ISH) in tumor tissue samples. ctDNA analysis using the Guardant Health OMNI panel was performed on plasma samples from most patients, examining HER2 amplification, copy number, gene alterations, and blood tumor mutation burden (bTMB). Serum HER2 extracellular domain (HER2ECD) levels were measured via ELISA. RNA sequencing (RNA-seq) was used to analyze HER2 gene expression in a subset of tumor tissue samples. Univariate analyses correlated baseline HER2 levels (IHC, RNA-seq, ctDNA amplification, copy number, serum HER2ECD) with ORR. Genomic alterations in ctDNA were analyzed for associations with ORR, including amplifications of MET, EGFR, and FGFR2, as well as mutations in PIK3CA and KRAS/NRAS. Changes in ctDNA alterations from baseline to end-of-treatment (EoT) were also investigated. Finally, gene expression differences between responders and non-responders were examined using RNA-seq data, along with assessing the impact of the timing of tissue collection (before or after/during first trastuzumab treatment) on T-DXd efficacy.

In the primary HER2-positive cohort, higher HER2 levels (IHC, RNA-seq, ctDNA amplification, copy number, serum HER2ECD) were associated with higher ORR. For example, patients with plasma HER2 amplification had a 61% ORR versus 34% in those without amplification. Amplifications of MET, EGFR, and FGFR2 in ctDNA were linked to numerically lower ORRs. HER2 gain-of-function (GoF) mutations were found in 11% of patients, with an ORR of 58.3%. The timing of IHC sample collection did not significantly affect ORR. In the HER2-low exploratory cohorts, higher baseline HER2ECD levels were associated with a higher ORR (36.7% vs 0%). There was high concordance (64%) between plasma HER2 amplification and HER2 tissue expression. Patients with high HER2 apCN showed longer median OS (mOS). High prevalence of TP53, HER2, CCNE1, and EGFR alterations were observed in ctDNA. No common trends were observed in ctDNA SNVs/Indels and amplifications at EoT; however, acquired TOP1 variants were identified in some patients. Gene expression analysis revealed that several genes, including those in the HER2 amplicon on chromosome 17q12-21, were highly expressed in responders. No clear association was found between PD-L1 RNA level and T-DXd efficacy.

This exploratory analysis identified several promising biomarkers predictive of T-DXd response in HER2-positive and HER2-low gastric cancer. Higher HER2 levels across various platforms (tissue IHC, ctDNA, serum) were consistently associated with improved ORR, suggesting that T-DXd remains effective even in the presence of some heterogeneity. The findings regarding other genomic alterations (MET, EGFR, FGFR2) warrant further investigation to determine whether these represent resistance mechanisms or simply prognostic factors. The detection of acquired TOP1 mutations suggests a potential resistance mechanism, highlighting the need for exploring alternative ADC payloads. The lack of a control arm and the relatively small sample size, especially in the HER2-low cohorts, are limitations. However, the data from a well-designed clinical trial strengthens these findings. Future research should focus on validating these biomarkers in larger, independent cohorts and explore the clinical implications of these findings.

This study provides valuable insights into potential biomarkers for predicting T-DXd response in gastric cancer. Higher HER2 levels across different platforms were associated with improved outcomes, while certain other genomic alterations may indicate potential resistance mechanisms. While limitations exist due to the exploratory nature and sample size, these findings warrant further investigation and validation in larger studies to optimize patient selection and treatment strategies for T-DXd in gastric cancer. Further research could focus on developing combination therapies that target both HER2 and other relevant pathways, such as those identified in this analysis.

This analysis is exploratory and post hoc, with a small sample size, particularly in the HER2-low cohorts. The absence of a control arm limits the interpretation of some findings, specifically those related to other genomic alterations. Furthermore, the lack of external validation restricts the generalizability of the results. The study also did not fully investigate the potential impact of the location of metastasis on ctDNA levels and its correlation with response to treatment.