Abstract PO1-01-05: Spatial PK/PD Model with HER2 Expression for Predicting Individual Tumor Response to T-DM1

Abstract Introduction: As new antibody drug conjugates are developed, the question of which specific patients will respond to these therapies remains. This work develops a spatial PK/PD model for the antibody drug conjugate trastuzumab emtansine, T-DM1, and explores the relationship between HER2 (target) expression and other individual features of the tumor microenvironment using a computational approach. Methods: We developed a spatial PK/PD model for T-DM1 based upon an allometrically-scaled mouse model (Jumbe, Nelson L., et al., 2010) and the reported outcomes from the KRISTINE trial, which investigated T-DM1 plus pertuzumab in the neoadjuvant early-stage breast cancer setting (Hurvitz, Sara A., et al., 2018). The T-DM1 spatial PK/PD model was then integrated into our previously described 4D biophysical simulation model (Howard, Frederick M., et al., 2022). The biophysical model leverages baseline breast MRI and clinicopathologic features and predicts the individual response to a treatment regimen. Using digital twin tumors from our TumorBank to simulate the KRISTINE trial cohort (n=40), the PD model was parameterized to achieve the trial response rate. To orthogonally validate the spatial PK/PD model, volumetric response and pCR were predicted in an independent cohort of ISPY2 trial patients (n=44) who received T-DM1 followed by dose-dense doxorubicin and cyclophosphamide. We then assessed the added value of the target expression data by developing logistic regression models to predict which patients would achieve pCR based upon HER2 expression alone or HER2 expression combined with the final tumor volume simulated using the spatial PK/PD model. Results: A logistic regression model based upon HER2 expression from transcriptomic data is a strong predictor of response to T-DM1 with an odds ratio of 9.6 and an area under the ROC curve (AUC) of 0.81. The tumorHER2 model, which combines the HER2 expression with the spatial PK/PD model, improves the predictive value of the model with an odds ratio of 34.0 and an AUC of 0.88. Conclusion: The spatial PK/PD improves prognostic predictions compared to HER2 gene expression data alone. In leveraging the SBS biophysical simulation platform, the spatial PK/PD model captures the individual variability in tumor growth rate, tumor perfusion, and drug disposition, all of which influence the treatment efficacy. The rate of target expression will highly influence response to an ADC, but the biophysical model demonstrates that features of the spatial tumor microenvironment also influence response. Table 1. Performance comparison between the SBS tumorHER2model and SBSHER2 Signature. Citation Format: Nicole Hobbs, Bradley Feiger, John Cole, Tricia Carrigan, Joseph Peterson, John Pfeiffer, Daniel Cook. Spatial PK/PD Model with HER2 Expression for Predicting Individual Tumor Response to T-DM1 abstract. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-01-05.