Feasibility Trial of Individualized Therapy for Relapsed or Refractory Acute Myeloid Leukemia Based on a High Throughput in Vitro Drug Sensitivity Assay

Introduction. Resistance to therapy, rather than treatment-related mortality, is the usual cause of failure to cure AML. Typically all patients receive the same therapy despite great inter-patient variation in the mutations that underlie the disease. Thus an individualized approach to therapy might be more productive. To this end, we developed a high-throughput sensitivity assay for 160 drugs; 45 are FDA approved and 115 investigational, encompassing a wide range of targets and mechanisms of action. We previously validated the assay in 30 primary patient blast samples and 14 acute leukemia cell lines. Here we report a clinical trial (NCT01872819 at clinicaltrials.gov) utilizing this assay to select drugs for patients with refractory AML.

Method. The primary objectives were to obtain assay results within 10 days and initiate treatment within 21 days. The secondary objective was to achieve a response (cytoreduction or at least partial response) greater that that expected for comparable refractory populations with other therapies. Mononuclear cells from marrow or peripheral blood were obtained by density centrifugation and enriched for blasts using magnetic bead separation if the initial sample contained < 80% blasts. Cells were incubated in coated 384 well plates overnight, then drugs were added at 8 concentrations spanning 4 log orders of magnitude, in duplicate. After 4 days, live cells were detected with CellTiter-Glo® (Promega). XLfit (idbs) was used to plot survival curves (4 parameter logistic dose fit) and to calculate EC50s. Individual drugs were chosen on the basis of EC50 and drug availability, and patients received the single agents at the accepted maximal tolerated dose.

Results. Fifteen patients were enrolled. Ten had unfavorable cytogenetics, and 3 had the Flt3ITD and 1 the Flt3D835 mutation. Eight patients had antecedent hematologic disorder. They had received an average of 5 prior therapies (range 3-6). The average time from sample procurement to assay result was 5.1 days (range 4-8). Within an average of 11.6 (median 9, range 4-28) days, 13 patients received single drugs to which their cells appeared to be sensitive with an EC50 range of 0.026 - 0.175 μmol/L , including cladribine, mitoxantrone, bortezomib, or vinblastine. For the patient with the Flt3ITD mutation, the blasts exhibited sensitivity to 6 Flt3 inhibitors in the high throughput assay. Although only FDA approved drugs were able to be procured, as the pharmaceutical companies denied requests for individual patient use, most patients received a drug they had not previously received. All patients exhibited a decline in blast number after receipt of the indicated drug, on average, by 92.6% (range 80.5-99.8%). Toxicity was as expected if the patients had received standard investigational protocols for relapsed/refractory AML. Median overall survival was 88 (range 7-276) days from start of treatment.

For one patient without circulating blasts, the marrow blast percent declined from 27% by flow to 0% at day 15 and also 0% at day 51. 6 of 9 evaluable participants exhibited a reduction in bone marrow blasts by flow cytometry on a day 14-21 marrow. There were also 2 patients whose day 14-21 marrows were severely hypocellular. Moreover, 1 patient achieved CR, and 2 patients, CRp, that occurred after additional cycles of combination chemotherapy regimens for 2 of the 3 patients, that included drugs identified by the high throughput assay.

Conclusion. In vitro high throughput testing to guide individual treatment choice is feasible and warrants further evaluation in larger clinical trials, with panels that include investigational drugs.