Evaluating antitumor activity and response determinants to trastuzumab deruxtecan in pediatric solid tumors.

10049 Background: Trastuzumab deruxtecan (T-DXd) is an anti-HER2 antibody-drug conjugate (ADC) linked to a topoisomerase inhibitor FDA-approved for several indications, including the treatment of HER2+ breast and gastric cancer. Although HER2 amplification is uncommon in pediatric cancers, recent demonstration of T-DXd efficacy in HER2-low breast cancer patients prompted us to examine the potential clinical relevance of T-DXd in pediatrics by evaluating HER2 expression and activity of T-DXd in preclinical pediatric solid tumor models. Methods: Cell viability was assessed in a panel of histologically diverse pediatric solid tumor cell lines and 2 HER2-amplified adult cancer cell lines treated with T-DXd, payload (DXd), and a control IgG-ADC. HER2 protein expression in patient-derived xenograft (PDX) tumors was evaluated by immunohistochemistry (IHC) using 5 clinically validated anti-HER2 antibodies. RNAseq was performed on clinical (n=290) and PDX (n=184) tumors to determine relative expression of ERBB2. In vivo activity of T-DXd was evaluated in osteosarcoma (OS, n=10), Wilms tumor (WT, n=1) and malignant rhabdoid tumor (MRT, n=1) PDX models and a desmoplastic small round cell tumor (DSRCT) cell line xenograft model. Differences in tumor volume and time to disease progression was assessed and compared across treatments and models. Responses were correlated with HER2 expression by IHC. Results: In vitro, HER2-amplified control cell lines demonstrated a >30-fold reduction of IC50 when comparing T-DXd to ADC control. In contrast, the ADC control IC50 was nearly identical to T-DXd across all pediatric cancer cell lines, consistent with the absence of HER2 amplification in these models. We observed focal and membranous staining of HER2 by IHC in PDXs but was quite variable across antibodies tested: WT 0-17% HER2+ cases (n=6), MRT 0-40% (n=5), DSRCT 0-33% (n=24), OS 0-7% (n=30). ERBB2 gene expression was highest in DSRCT followed by WT, OS and MRT. In vivo efficacy studies demonstrated complete and partial responses in OS, WT, and DSRCT and improved disease control rates (OS: 67%, p=0.006, Mann-Whitney; WT: 100%; DSRCT: 100%). However, T-DXd and ADC control demonstrated similar activity in all tumor types with no consistent correlation between in vivo response and HER2 expression. Consistent with these preclinical studies, 4 patients with progressive DSRCT were treated with T-DXd via compassionate/off-label access with signs of clinical and radiographic responses. Conclusions: T-DXd shows significant preclinical antitumor activity across multiple pediatric solid tumors but little correlation with HER2 expression suggesting a mechanism of action similar to the clinical activity observed in HER2-low breast cancer. Xenograft efficacy studies and preliminary clinical experience with T-DXd in DSRCT patients warrant formal clinical trial investigation in this largely incurable disease.