Background
Myelodysplastic syndrome (MDS) patients who are refractory to hypomethylating agents (HMAs) have a poor prognosis with median survival <6 months and few treatment options. A precision medicine approach is appealing in MDS given the biologic heterogeneity associated with the large variety of cytogenetic abnormalities and somatic mutations. We sought to determine whether a precision medicine approach combining molecular testing, ex vivo drug sensitivity screening (DSS), and in silico computational biology modeling (CBM) could be performed within an actionable timeframe (≤30 days) to allow for personalized treatment recommendations for patients with HMA-refractory MDS.
Methods
Study design: We performed a prospective feasibility study in 21 patients with HMA-refractory MDS enrolled at Stanford University from April 2018 through March 2019. All patients had a baseline bone marrow (BM) biopsy with BM aspirate and peripheral blood (PB) samples sent for mutation testing (596-gene panel, Tempus, Chicago, IL) and ex vivo DSS (Notable Labs, Foster City, CA).
Ex vivo DSS: BM aspirate and PB specimens were RBC-lysed and re-suspended in serum-free media with cytokines. Samples were plated in 384-well microtiter plates and screened against FDA-approved and investigational drugs (up to 76) and drug combinations in triplicate. Specimens were treated for 72 hours and assayed using high-throughput, multi-parametic flow cytometry for cytotoxicity and differentiation (Blood 2016;128:5206).
In silico CBM: Genomic data were input into a computational biology model (Cell Works Group, San Jose, CA) to generate protein network maps for each patient. Mathematical modeling of MDS cell proliferation or inhibition was simulated for each patient and used to calculate drug efficacy scores for numerous agents (Leuk Res 2017;52:1-7).
Study endpoints: Once the gene panel, ex vivo DSS, and in silico CBM results were available, we (M.A.S., A.A., J.Z., P.L.G.) met for a molecular tumor board (MTB) to review the data and provide personalized treatment recommendations for each patient. The primary endpoint was the feasibility of providing personalized recommendations within an actionable timeframe (≤30 days). Secondary endpoints included concordance between the ex vivo and in silico assays and the accuracy of our MTB recommendations in predicting clinical responses in vivo.
Results
The median age of the patients was 76 years (range 55-87) and 71% were male. Seventeen patients had MDS, 3 had an MDS/MPN disorder, and 1 patient had progressed to AML. 76% had higher risk disease by IPSS-R, 57% had excess blasts, and 52% had adverse cytogenetics or mutations. Patients had a median of 2 pathogenic mutations (range 0-6) with the most common including TET2, ASXL1, STAG2, DNMT3A, RUNX1, and SRSF2.
The median turnaround time for results of the gene panel, ex vivo DSS, and in silico CBM were 14.5, 15, and 20 days, respectively. The median turnaround time to our MTB was 27 days (range 20-32 days). MTB recommendations varied widely among patients and encompassed various drug classes including targeted therapies (venetoclax, sorafenib, lenalidomide, ruxolitinib, midostaurin, everolimus), cytotoxic agents (cytarabine, fludarabine), differentiative agents (calcitriol, ATRA), HMAs, and androgens (danazol) as single agents or in combination. The ex vivo and in silico assays were highly concordant, particularly in predicting sensitivity to HMAs and venetoclax.
Eight patients received treatment per our MTB recommendations. Of these 8 patients, 6 (75%) responded to the recommended therapy and 2 (25%) had stable disease. Two responding patients were bridged to allogeneic hematopoietic cell transplantation (HCT). The remaining patients elected for best supportive care (N=5), hospice (N=3), other approved therapies (N=3), a clinical trial (N=1), or allogeneic HCT without bridging therapy (N=1).
Conclusions
We demonstrate the feasibility of a novel precision medicine approach for HMA-refractory MDS patients combining mutation data, ex vivo DSS, and in silico CBM to guide clinical therapeutic decisions within an actionable timeframe. Personalized treatment recommendations accurately predicted clinical responses in vivo and enabled some patients to be bridged to allogeneic HCT. Randomized prospective trials are needed to determine whether this approach may improve outcomes for patients with HMA-refractory MDS.
Disclosures
Aleshin: Notable Labs: Consultancy. Santaguida:Notable Labs: Employment. Schaffert:Notable Labs: Employment. Abbasi:Cell Works Group, Inc.: Employment. Patterson:Notable Labs: Employment. Heiser:Notable Labs: Employment. Greenberg:Notable Labs: Research Funding; Celgene: Research Funding; Genentech: Research Funding; H3 Biotech: Research Funding; Aprea: Research Funding; Novartis: Membership on an entity's Board of Directors or advisory committees.
RE: “TOOLS FOR THE PRECISION MEDICINE ERA: HOW TO DEVELOP HIGHLY PERSONALIZED TREATMENT RECOMMENDATIONS FROM COHORT AND REGISTRY DATA USING Q-LEARNING”