Abstract 5141: Using single cell gene expression to develop treatment response predictions in multiple myeloma

Abstract The promise of precision medicine in oncology is to identify the ideal treatment for each patient, eliminating failed treatment cycles and reducing treatment burdens on patients and payors. While gene sequencing can predict effective treatment options in some cases, studies have estimated that only 10-15% of cancer patients are treated with genotype matched drugs. Additional treatment selection tools for predicting patient response thus remains a critical clinical need. While gene sequencing has clear utility in predicting therapeutic response, gene expression is less commonly used in clinical tests to guide treatment selection. Gene expression is theoretically attractive to predicting therapeutic response because, unlike mutational analysis, it provides a direct readout of the dysregulated biochemical activities in the cells which may be targeted by particular therapies. One of the major problems with conventional approaches to using gene expression to predict therapeutic response is tumor heterogeneity; that is, a biopsied tumor contains multiple cells types and even different tumor clones that can confound a conventional bulk gene expression analysis. More recently, single-cell RNA expression analysis technologies have grown in popularity. These technologies enable the fine characterization of tumors and their cellular makeup and hold the promise of increasing the power of gene expression measurements for identifying therapy response prediction tools. Here we show the application of single cell gene expression analysis with Proteona’s MapResponse™ machine-learning algorithms to developing a predictive classifier for patient response to Daratumumab treatment in multiple myeloma. Starting with publicly available data (Cohen et al. Nat. Medicine 2021), we developed a classifier that accurately predicted the response of 94% of subjects in the published study. This classifier was able to accurately predict response in an independent cohort of Daratumumab-treated patients. In contrast, the classifier performed poorly in a cohort of patients who did not receive Daratumumab. Together these data suggest the classifier has specificity to predicting response to Daratumumab and is not a general prognostic signature. These early findings point to the potential value of combining single-cell RNA sequencing with machine-learning methods such as MapResponse™ to bring precision medicine to more patients. Citation Format: Jonathan Scolnick, Stacy Xu, Sanjay de Mel, Cinnie Yentia Soekojo, Wee Joo Chng. Using single cell gene expression to develop treatment response predictions in multiple myeloma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5141.