(PA-205) CoMMpass Explorer: An Interactive Platform to Explore Clinico-Genomic Data from Newly Diagnosed Multiple Myeloma Patients from the CoMMpass Observational Trial

The CoMMpass study (NCT01454297) was a prospective, longitudinal trial of treatment naïve, newly diagnosed multiple myeloma (NDMM) patients (n = 1,141). Tumor samples were characterized using whole-exome sequencing and bulk RNA-seq at diagnosis and progression. Additionally, immune microenvironment of a subset of patients was assessed at single cell resolution (3’ scRNA-seq). To facilitate data accessibility and exploration of this multi-omic dataset, we developed CoMMpass Explorer (CE), a user-friendly platform that enables real-time exploration of clinico-genomic data from patient samples enrolled in the study.

Methods: CE provides four functional views – a) Overall Summary displays clinical feature distribution, and Kaplan–Meier survival curves b) Mutational Profile provides somatic mutation visualization and comparison, c) Tumor Profile shows expression-level analysis including differential expression and gene set enrichment and d) Immune Microenvironment compares cell type abundance and cell cycle distributions between cohorts using scRNA-seq data.

Results:

CE centers around the idea of cohort building by allowing users to filter on the rich set of clinical, survival, and genomic data elements from CoMMpass dataset to create custom cohorts. As proof of concept, we demonstrate the validity of CE by reproducing results from two previously published studies that used CoMMpass datasets.

In our first use case, we reproduced results from Simhal et al by building cohorts based on WEE1 gene expression as described by the authors. We were able to reproduce the results from the study by showing patient stratification on the basis of WEE1 gene expression and associate it with progression-free survival (PFS) using Kaplan–Meier survival curves. In our second use case, we reproduced results from Manojlovic et al by building cohorts based on self-reported race, one cohort has only African Americans and the second cohort with only Caucasians.

 
We also demonstrate the value of CE by comparing two cohorts of patients at baseline – NDMM patients that received triplet therapy up front (Cohort 1; N=241) vs that did not (Cohort 2; N=139). Additionally, the patients in these two cohorts did not receive any transplant. While there is no statistical difference in the PFS or overall survival between these two groups, CE identified distinct mutational patterns in TENT5C in Cohort 1, which is less differentially expressed compared to patients in Cohort 2. Cohort 1 is also characterized by a high EMT, PI3K/AKT/MTOR signal along with a lower abundance of CD8+ T (p < 0.05) and NK (p=0.01) cells.

Conclusions: CE democratizes access to clinico-genomic data, empowering the myeloma research community with an intuitive tool for data exploration. Using an example, CE demonstrates significant differences in gene expression, mutational landscape and cell type abundance between patients that received triplet up front vs patients that did not within a non-programmatic framework.