Superior therapy response predictions for patients with myelodysplastic syndrome (MDS) using Cellworks Singula: MyCare-009-02.

Abstract

e19528 Background: Despite using cytogenetic and molecular-risk stratification and precision medicine, the current overall outcome of MDS patients remains relatively poor. Therapy selection is often based on information considering only cytogenetics and single molecular aberrations and ignoring other patient-specific omics data that could potentially enable more effective treatments. The Cellworks Singula™ report predicts response for physician prescribed therapies (PPT) using the novel Cellworks Omics Biology Model (CBM) to simulate downstream molecular effects of cell signaling, drugs, and radiation on patient-specific in silico diseased cells. We test the hypothesis that Singula is a more accurate predictor of patient-specific therapy response than PPT. Methods: Singula’s ability to predict response was evaluated in an independent, randomly selected, retrospective cohort of 146 MDS patients aged 28 to 89 years (median 69) treated with PPT. Patient omics data was available from PubMed and TCGA. The accuracy of Singula was compared to that of PPT using McNemar’s test to account for the correlation between Singula and PPT. Multivariate logistic regression modeled complete response (CR) as a function of patient age, PPT, and Singula against any non-response (NR). Likelihood ratio tests were performed to further validate if Singula provides predictive information beyond PPT or patient age. Similar analyses were performed for overall survival (OS) using proportional hazards regression. Results: Singula was a better predictor for CR than PPT (McNemar’s χ2 = 42.0, p-value < 0.0001), with an overall accuracy of 73.3% (Exact 95% CI: 65.3%, 80.2%) compared to 37.7% (95% CI: 30.0%, 46.1%) for PPT. Singula exhibited a sensitivity and specificity of 90.9% (95% CI: 80.0%, 97.0%) and 62.6% (95% CI: 51.8%, 72.6%), respectively. In multivariate regression analysis, Singula (p < 0.0001) remained an independent predictor for CR after adjusting for patient age (p = 0.0759) and PPT (p = 0.0496). Singula provided alternative therapy selections for 17 of 53 true negative detected by Cellworks. Conclusions: Singula is a superior independent predictor for CR compared to PPT in MDS patients. The Singula report can also validate therapy selection, correctly identify non-responders to PPT and further provide alternative therapy selections.