Innovative Approach to Precision Trial Design: NCI-MATCH and Beat AML

WHILE ON THE ONE HAND

, researchers and drug developers are identifying molecular targets in specific cancer subtypes to improve outcomes, they have also been innovating on the clinical trial design front. At a late session during the 58th American Society of Hematology Annual Meeting & Exposition, held December 3-6 in San Diego, CA, representatives from 2 national clinical trials, Beat AML (acute myeloid leukemia) and NCI-MATCH (National Cancer Institute-Molecular Analysis for Therapy Choice), detailed how they were incorporating genomic profiling to assign patients to different treatment arms.

Providing an update on the Beat AML trial was Brian Druker, MD, director of the Knight Cancer Institute at Oregon Health & Science University, which is collaborating with the Leukemia & Lymphoma Society on the trial. Providing a background on the disease, Druker said that AML is the most common leukemia in adults, with a median age at diagnosis of 67 years, and although the dozens of different molecular subtypes make it a very complicated disease to treat, a few targeted agents have seen some progress in treating AML, including spleen tyrosine kinase inhibitors, IDH1/2 agents, kinases (FLT3, KIT), and the more recent immune checkpoint inhibitors.

However, despite all the progress with characterizing the molecular abnormalities associated with the disease, progress on the treatment front has been dismal. “Treatment evolution for AML has lacked significantly,” Druker said, adding that disease outcomes have remained poor over the past decade and there have been very few approvals. “The fact that AML is a very heterogeneous disease could also have a role to play,” he added.

In addition, conducting a clinical trial for AML remains a hurdle, Druker noted. “Challenges include the fact that the standard of care remains beneficial, single-agent treatment will not be beneficial, genomic assays take long to deliver, and trials are hard to recruit for.” With all these challenges, the Beat AML trial has been designed with the following objectives:

1. Perform genomic screening of patients at clinical trial entry
2. Assess the feasibility of waiting 7 days for the genomic test results
3. Assign therapy based on genomic screening
4. Incorporate a marker-negative arm so all patients have a treatment option
5. Provide a network for junior clinical investigators

The trial has a multi-arm protocol, with:

1. Each arm independent from the other, with consistent eligibility
2. Window design ensuring documentation of all large effects in treatment-naïve patients
3. Initial focus on those 60 and older

TABLE

Trial eligibility criteria are straightforward: patients 60 years and older who have previously untreated AML can participate. Following genomic analysis of their tissue, patients will be assigned to independent treatment arms in the protocol. “The primary objective of the Beat AML trial is to assess the feasibility of trial design,” Druker said. “Secondary objectives are to determine how many patients can be successfully enrolled, determine if patients can reach allogenic stem cell transplant, and assess impact on outcomes.” He listed the treatment substudies and their start date, as shown in the .

For biomarker assessment, Druker said that cytogenetics assays will be local. Meanwhile, biopsy samples will be sent to Foundation Medicine to conduct a more long-term 300-gene panel assay. “However, critical genes will be assayed by the company in 7 days, including NPM1, IDH1/2, and FLT3.”

The order of patient assignment to a treatment arm will be based on:

1. Chemotherapy response
2. Molecular marker with high variant allele frequency
3. Higher-risk group that may confound efficacy
4. Marker negative

Trial endpoints are standard, Druker told the audience, and include primary endpoints of complete response and response duration. Secondary endpoints include event-free survival, progression-free survival (PFS), overall survival, and minimal residual disease.

Still in its early stages of conception, Beat AML has “enrolled 4 patients to date,” Druker said. “The goal is to allow patients to be enrolled in active treatment arms, and the master protocol allows switching between the arms.” He added that in the future, the trial would like to include additional arms on the protocol and test novel combinations.

The second presentation of the session, by Keith Flaherty, MD, provided an update on the NCI-MATCH trial. Flaherty, director of the Henri and Belinda Termeer Center for Targeted Therapies at Massachusetts General Hospital and associate professor of medicine at Harvard Medical Center, also chairs ECOG-ACRIN, which is collaborating with NCI on this trial. He was very excited to share with the audience that the trial was expected to hit it’s 6000 patient enrollment target in the next 6 months.

“We are currently enrolling 120 to 150 patients being each week,” Flaherty said. “The objective of this phase 2 precision-med trial is to match genetic abnormalities of tumors with a suitable targeted drug, regardless of cancer type,” he explained. “It’s a signal-finding trial, meaning promising treatments can be expanded to a more definitive trial in the future.”

Eligibility criteria for enrollment in NCI-MATCH include adults over 18 years, those who lack or have exhausted standard treatment, patients who have developed either solid or liquid tumors, patients with a good ECOG performance status and adequate organ function, and patients who can tolerate being off treatment for 6 weeks. Flaherty listed the following criteria for source material for genetic and immunohistochemistry analysis:

1. The trial mandates a fresh tumor biopsy to identify gene abnormalities
2. Patients can be screened with local next-generation sequencing, but results have to be confirmed on an NCI-MATCH assay
3. Biopsy and sequencing on progression for responders
4. Planned assays for research purposes: Whole-exome DNA sequencing RNA analysis by whole transcriptome analysis microRNA assay

The following Levels-of-Evidence strategy is being implemented by NCI-MATCH:

Level 1: gene variant credentialed for selection of an approved drug

Level 2a: variant eligible for an ongoing clinical trial

Level 2b: variant identified in an N of 1 response Level 3: preclinical inferential data

Levels of Evidence for drugs in NCI-MATCH include:

Level 1: FDA-approved for any indication for that target

Level 2: agent met a clinical endpoint, with evidence of target inhibition

Level 3: agent demonstrated evidence of clinical activity, with evidence of target inhibition at some level

Among the 6000 patients that will be the final enrollment, 929 treatment enrollments are anticipated across 24 gene abnormalities that are currently being evaluated as part of this trial. The primary trial endpoint is overall response rate, with secondary endpoints of PFS, time to progression, toxicity, and biomarker expression.

Flaherty explained that the trial demands 4 core biopsies at initial entry, which are shipped to the central lab at MD Anderson. H&E sections are assayed by a pathologist for tumor type, content, percent necrosis, and inflammation, and scanned into a high-resolution image database. RNA and DNA are then extracted and distributed to a network of laboratories.

Currently, immunohistochemistry analysis is being conducted for PTEN, MLH1, MSH2, and Rb. “We have also added mismatch repair genes and are evaluating PD-1 expression,” he added. The trial has incorporated a customized Oncomine assay, which has been developed by Thermo Fischer. The panel includes 143 genes, 2530 amplicons in the DNA panel, and 207 amplicons in the RNA panel. Flaherty provided a very uplifting picture on patient wait times:

• Sample submission from sites to central lab at MD Anderson: 7 days
• Completion of tumor testing by lab network and return of results to site: 15 days
• Secondary screening for patients assigned to a treatment arm: 14 days.

“As of November 27, we have 3149 patients with tumor samples, of whom 2589 have received their test results; 468 had a genetic abnormality matching an available treatment,” Flaherty told the audience, “and 22% of currently enrolled patients have a gene abnormality that matches one being studied in the trial.” Although the trial currently has 24 arms, this number is expected to increase.

Download PDF: Innovative Approach to Precision Trial Design: NCI-MATCH and Beat AML