Increasingly, genomic alterations are being recognized and targeted for clinical trials. Improving clinical trial accrual in this environment may require multiple histology basket trials that increase the percentage of patients with “actionable” alterations, said Funda Meric-Bernstam, MD, at PMO Live 2015.
The implications for the use of predictive markers for designing and executing clinical trials was reviewed by Meric-Bernstam, Chair, Department of Investigational Cancer Therapeutics, MD Anderson Cancer Center, Houston, TX.
Much discussion still revolves around the best approach for molecular characterization of cancer, she said. Genetic profiling has gained momentum over the past few years, especially for clinical trials, but transcriptional profiling had already been implemented in the Clinical Laboratory Improvement Amendments environment for some clinical trials and had clear value demonstrated in more limited panels. One example is breast cancer, in which treatment of estrogen receptor–positive disease has been transformed. Proteomic profiling, methylomics, and metabolomics are also of great interest.
The rapid evolution of genomic testing has been the most striking in the clinical environment. Sanger sequencing of a single gene has given way to 400-gene targeted open reading frame sequencing and whole exome/whole genome sequencing. The result is that the extent of information available about an individual patient enrolled in a clinical trial has skyrocketed.
“Many of the clinical trials that are leveraging genomic profiling pick 1 of 2 options,” said Meric-Bernstam. One is the biomarker-selected single-arm trial in which patients are treated with a drug that targets a specific alteration or downstream from the alteration. The second choice is a biomarker-selected randomized trial, usually comparing a targeted therapy versus standard of practice. Multiplex testing using protein-based assays or genomic-based assays to study several actionable alterations speed up the process.
Examples of umbrella trials allocating patients to phase 2 trials based on molecular profiling are BATTLE (lung cancer), BEAT-IT (metastatic breast cancer), ATTACC (metastatic colorectal cancer), and MelTTT (metastatic melanoma).
Proof-of-principle clinical trials in which molecular profiling is used to select patient therapy in phase 1 trials outperform trials using nonmatched therapy in terms of progression-free survival (PFS) as an outcome (Clin Cancer Res. 2012;18:6373-6383).
An important effort at MD Anderson Cancer Center is the IMPACT 2 trial, a randomized study evaluating molecular profiling and targeted agents in metastatic cancer. The objective is to determine whether patients treated with a targeted therapy selected on the basis of mutational analysis of the tumor have longer PFS from the time of randomization than those whose treatment is not selected based on alteration analysis.
The NCI-MATCH trial is enrolling patients with solid tumors or lymphomas who will have mutations/amplifications/translocations identified in the tumor sample and will have treatment selected based on actionable mutations detected. In addition, tumor biopsies and sequencing will also be performed at disease progression to illuminate resistance mechanisms.
At MD Anderson, “we wanted to roll out genomic testing broadly, and it quickly became apparent that there are only a handful of diseases where genomic testing is routinely accepted as a standard of care,” said Meric-Bernstam. In its Clearinghouse Protocol (J Clin Oncol. 2015;33:2753-2762), the Institute for Personalized Cancer Therapy at MD Anderson is attempting to determine the feasibility of implementing standardized genomic testing, to determine the frequency of mutations and comutations, to establish a database, and to determine predictors of response.
“Almost 40% of patients had a mutation in an actionable gene,” she said. Eighty-three patients (11%) had a potentially actionable alteration and have been enrolled in genotype-matched trials targeting these alterations. Patients with a variety of tumor types have been enrolled, enriched for breast cancer and colon cancer.
In 31% of patients, test results were not mentioned, and in one-third of patients, genotype-matched trials were not discussed. Based on this knowledge, opportunities exist for increasing awareness of results, available clinical trials, and relevant trials. Additional obstacles to enrollment included patients not being eligible for a trial based on their specific alteration, performance status–related issues, or patient’s desire to be treated locally.
When the study was started in 2012, using the 50-gene assay, only a handful of mutations were considered actionable, but many more targets are covered with ongoing clinical trials in 2015.
The number of histology-agnostic or multihistology basket trials was increased to amplify the percentage of patients with actionable alterations with the use of larger panels containing more genes as well as genes that would return copy number data.
