A Multi-Omic Precision Medicine Clinical Trial in Acute Leukemia

Background: Conventional precision medicine for cancer targets specific gene mutations, but single agent inhibitors rarely result in remission or improve survival in acute leukemia. Current functional screening strategies assay individual drugs, thus missing potential synergistic combinations. We therefore developed a multi-omic approach that integrates mutation data, gene expression data (transcriptome), and in vitro drug sensitivity (functional) data to select drugs including drug combinations for individual patients. Herein we report results of the clinical trial testing this concept [NCT02551718]. This study contributed to validation of 44 genes identified for which expression correlates with drug sensitivity (Lee S-I et al. Nat Commun 2018).

Patients and Methods: Eligible patients failed at least 2 prior regimens, or if adverse risk, one multi-drug intensive regimen, had ECOG PS 0-3 and at least 1 million blasts in blood, marrow, tissue or fluid for analysis. The original enrollment was 25 patients to establish feasibility, and the study was later expanded to provide an option for refractory patients. Samples were obtained from the 54 consented patients (44 AML, 8 ALL, 2 acute leukemias of ambiguous lineage). Median age is 58 (range 23-82) and 31.5% of patients had an antecedent hematologic disorder. Of the AML patients (n=44), 7 were favorable, 13 intermediate, and 24 adverse risk (ELN 2017). Patients had a median of 3 prior treatments (range 1-6); 39 patients (72%) had relapsed after prior complete remission (CR) with median duration of 6 months (m) (range: 0-52 m), and 20 had relapsed after allogeneic transplant, 8 within the first 100 days, and 3 never achieved a CR after transplant.

Enriched blasts were assayed in a custom CLIA-approved high throughput sensitivity (HTS) screen with 153 drugs and combinations, both conventional and targeted inhibitors, both FDA approved and investigational. Results were obtained within a mean of 5.2 (range 4-7) days. Mutation testing was performed by Invivoscribe using MyAML® targeted NGS panel. Clinical outcomes examined were peripheral blast reduction, response and survival.

Results: Twenty-nine patients (53.7%; 25 AML, 2 ALL, 2 acute leukemias of ambiguous lineage) received therapy based on the HTS and mutation analysis. The remaining patients did not receive protocol treatment for a variety of reasons, including insufficient marrow blasts for testing, opting for palliative care, returning to their local area, denial by insurance , lack of access to investigational drugs, or medical complications making them ineligible. Of the 22 patients who had circulating peripheral blasts, 21 (95%) had a reduction, including 7 eradication, of circulating peripheral blasts following therapy. The median number of assay-directed regimens patients received was 2 (range 1-10, mean 2.5). Therapies administered are shown in the Table. Mutation testing on 47 patients revealed that 32 samples (68%) harbored mutations for which an approved or investigational targeted inhibitor could be considered (Figure 1). Figure 2 shows the heatmaps for individual patient mutations and drug sensitivity.

Median survival following initiation of protocol therapy was 70 days (range 19-811). Of these often heavily pre-treated patients, 2 achieved CR, 1 CRi, and 6 had partial remissions (PRs), overall response rate (ORR) 31%. It was not always possible to obtain the top drugs if the use was off label or they were investigational, so drugs were selected that were lower in the ranked list of IC50s. When we observed 1-3 log reduction in IC50s for a combination compared to the single agents, the combination regimen would be recommended. For the 9 patients who received intensive combination regimens, there were 1 CR and 2 PRs, (ORR 33%), and for the 13 who received low intensity regimens, there was 1 CR, 1 CRi and 1 PR (ORR 23%). Enrolled allogeneic transplant recipients who received study treatment had a median survival of 476 days post-transplant compared to 340 days for those who did not receive assay guided therapy.

Conclusions: This study demonstrates the feasibility of simultaneous collection of genomics, gene expression, and functional drug sensitivity data with the intent to guide choice of therapy. Responses were observed after study guided treatment. Future trials will incorporate new algorithms based on the correlative analyses obtained by this study to optimize treatment choices.

View largeDownload slide

View largeDownload slide

Close modal