NEW YORK – Using functional drug screening together with molecular testing significantly improved advanced leukemia patients' overall survival times in a pilot clinical trial presented at the American Society of Hematology's annual meeting.
Pamela Becker, a hematologist at the City of Hope, shared data showing that giving personalized treatments to patients using genomic analysis and functional testing reduced their risk of death by 83 percent, compared to those who weren't treated with a matched regimen.
"Genomic analysis and functional screening … can comprehensively predict in vivo efficacy and individualize optimal treatment for patients with relapsed or refractory acute leukemia," Becker said, acknowledging that the findings of this study need to be validated in clinical trials that involve more patients, healthcare centers, and drugs and drug combinations. For now though, the study presented at this meeting provides some insight into the value multiple predictive methods can potentially bring to matching leukemia patients to the best treatment.
"We absolutely believe that just mutational testing alone is not revealing who's likely to respond," Becker said, adding that there have been numerous basket trials demonstrating this failure and highlighting the need for better-matched treatments. For example, the SHIVA trial is often cited by experts who feel the value of molecularly driven therapies has been overhyped. The National Cancer Institute's MATCH trial has also been criticized since a lot of the arms in that genomically informed platform trial failed to accrue.
The clinical trial led by Becker was sponsored by the University of Washington, where she previously worked, and the NCI. The primary goal of the study was to assess the feasibility of testing patients' cells in a high-throughput assay against individual drugs and drug combinations and select optimal treatments for patients according to information gleaned from functional testing and molecular sequencing. Becker and colleagues pre-specified that the trial would be considered successful if they tested and initiated treatment within 21 days for at least nine out of 15 patients.
The feasibility question is one that's been top of mind for functional precision medicine researchers and biotech firms globally. A growing body of research is elucidating the benefits of testing patients' tumor models for drug responses ex vivo before treating the patients themselves, but it is less certain how such functional evaluations could be integrated into routine practice.
Becker and colleagues wanted to put functional drug screening logistics to the test. As secondary objectives, the investigators wanted to assess patients' outcomes on recommended therapies.
In the study, the therapy patients received was ultimately up to their treating physicians; they did not have to receive matched treatments. Although the study wasn't a randomized trial, the patients who received their physicians' choice of treatment became the cohort to which patients with matched treatment were compared.
"With the decision to go forward [with matched treatment], we could then monitor what happened," Becker said, explaining that all the patients had functional and molecular testing and received recommended matched therapy, even if not all patients took those treatments.
In total, 60 patients with advanced leukemia were enrolled onto the feasibility study, of whom 51 had acute myeloid leukemia, seven had acute lymphoblastic leukemia, and two had acute leukemias with ambiguous lineages. To be eligible for the study, patients had to have relapsed after two lines of therapy or after one line of therapy if they had high-risk disease and were ineligible for other trials. Among those enrolled, the median number of prior treatments was four and a quarter of patients had not responded to primary induction chemotherapy. Over one-third had relapsed after an allogeneic transplant.
Functional screening process
To conduct the functional drug screening, Becker and her team first collected samples from the patients, including blood and bone marrow or fluid and tissue with leukemic blasts. In the University of Washington's Quellos High Throughput Core facility, they used magnetic beads to enrich the cells, after which they performed functional drug sensitivity testing using the University of Washington's CLIA-certified Cancer Drug Sensitivity Test.
The test involves isolating and distributing the cells into thousands of tiny chambers in multi-well plates, then dosing the cells with different drugs or drug combinations and evaluating their cytotoxic activity. For this part, the researchers measured the percentage of surviving cells in each chamber after exposure to individual drugs or to drug combinations. The researchers included 170 drugs and drug combinations, including both US Food and Drug Administration-approved therapies and investigational agents. Tested therapies ranged from immunotherapies to MEK inhibitors and various chemotherapy combinations, among others.
For the molecular profiling part of the study, the researchers used InvivoScribe's MyAML next-generation sequencing test to gauge patients' samples for mutations in 194 genes. They also used Mission Bio's Tapestri myeloid gene panel to perform single-cell mutation analysis, in turn allowing them to view the samples' clonal compositions.
Based on functional drug testing and molecular analyses, the patients' oncologists received reports with ranked therapy suggestions.
Ultimately, 35 patients — at the discretion of their oncologists — received a drug or drug combination based on the predictive tests. The median turnaround time for the functional testing results was 5.6 days and 12.5 days for mutational analysis. Patients started treatment after a median of 7.8 days.
Among evaluable patients, the overall response rate was 33 percent, with two patients who achieved complete remissions, one patient who had a complete remission with incomplete cell recovery, and six patients who had partial responses. Of 22 patients who had circulating peripheral blasts when they began treatment, 95 percent had a reduction in these blasts and 32 percent saw the blasts completely eradicated.
After 100 days on treatment, the overall survival rate was 57 percent for those who received matched therapy versus 24 percent for those who did not. The median overall survival times were 135 days and 34 days for those who did and did not receive the matched therapy, respectively.
When Becker and colleagues adjusted the findings for patients' ages, sex, type of leukemia, cytogenic risk, and number of previous lines of therapy, the findings revealed that the assay-guided therapy could decrease the risk of death by 83 percent.
Future clinical trials, Becker said, will implement a new predictive model that's been built based on these patients' gene expression findings, co-occurring mutations, and clonal composition. "Every single patient exhibits a different pattern of drug sensitivity," Becker said, emphasizing that for this heavily refractory patient population, there's a great deal of value in incorporating functional drug screening into treatment selection to ensure the best outcomes.