Predictive Simulation-Driven Personalization Methodology For Refractory Multiple Myeloma

Background: Given the genomic variability observed across multiple myeloma patients, treatment personalization becomes imperative. Personalization (N = 1) driven by actionable mutations alone is a step forward but has limitations. Holistically incorporating genomic profiling information to create a patient simulation avatar provides a personalization paradigm beyond the one-drug one-mutation method of individualization. Methods: A bone marrow sample from the patient was analyzed for chromosome alterations and molecular aberrations by NYU and Gen Path Diagnostics. Using this genomic profiling information, a predictive simulation avatar of the patient was created that comprehensively models signal transduction, epigenetic regulation, protein homeostasis, autophagy, and metabolic pathways. The simulation highlights the patient9s dominant driver characteristics and enables personalized therapy design. Here, a library of FDA-approved, molecularly targeted drugs from across indications were simulated individually and in combination on the patient simulation avatar. Results: The patient had high risk IgA multiple myeloma and refractory to following treatment regimens: Thalidomide/Dexamethasone (DEX), Velcade/DEX, Revlimid (REV)/DEX, REV/DEX/Elotuzumab, stem cell collection with high dose Cytoxan, stem cell transplant with high dose Melphalan, Velcade/Bendamustine, Pomalidomide, Carfilzomib (CAR), Doxil, CAR/Cytoxan, CAR/Pomalidomide. The patient9s cytogenetic information revealed tetrasomy of chromosomes (Ch) 9 and 15; trisomy of Ch 2, 3, 5, 7, 11, 14, 19, 20, and 22; monosomy of Ch 16. Also, additional material was attached to the short (p) arms of Ch 1 and 21 and the long (q) arms of Ch 6, 14 (involving the IGH locus), and 16. The amplified and deleted genes were used to create the patient simulation avatar definition. The resulting patient avatar demonstrated increased expression of the following key growth factor receptors and signaling intermediates: EGFR, HGF, MET, IGF1R, SHC, and SRC. Additionally, increased PI3K and AKT copy number was noted, resulting in a dominant AKT/PI3K signaling axis. IL6, IL6R, and JAK2 amplification was observed, resulting in increased JAK/STAT signaling. Nelfinavir and Ruxolitinib was administered to impact the PI3K/AKT and JAK/STAT signaling axis predicted by simulation. This treatment exhibited a clinical response in the highly refractory patient indicated by a significant drop in the IgA from 3610 to 2650 mg/dl, the M spike from 3.9 to 2.7 g/dl and the total protein from 10.6 to 8.2 gm/dl. The response lasted 2.5 months with minimal toxicity. Conclusions: This study demonstrates and provides validation of technology to holistically incorporate big data from genomics for actionable insights for treatment personalization. Citation Format: Nicole A. Doudican, Amitabha Mazumder, Shireen Vali, Kabya Basu, Ansu Kumar, Neeraj Kumar Singh, Zeba Sultana, Taher Abbasi. Predictive simulation-driven personalization methodology for refractory multiple myeloma. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5443. doi:10.1158/1538-7445.AM2015-5443