Redesigning Clinical Trials Necessary to Enhance Discovery of Effective Targeted Drugs in the Genomic Era

Wayne Kuznar

October 2013, Vol 4, No 8 - Personalized Medicine


Boston, MA—Defining optimal therapeutic efficacy in the genomic era will require that clinical trial design in oncology move from a drug-centric approach to a patient-centric one. Retrofitting current knowledge into old paradigms will slow the progress in discovering effective targeted agents, said Razelle Kurzrock, MD, Director, Center for Personalized Therapy and Clinical Trials, Moores Cancer Center, University of California, San Diego, at the Global Biomarkers Consortium Second Annual Conference.

Common cancers are comprised of multiple subgroups in which different pathways may be activated. Patients with cancer may have hundreds of different genetic mutations, especially in the metastatic setting. “The individual landscape of a patient is unique,” said Dr Kurzrock.

The uniqueness of each patient with cancer means that each patient requires a specially tailored treatment regimen. In a randomized trial of unselected patients, however, a drug that hits the more common pathway will be declared the superior drug to one that hits a less common pathway. This unselected approach runs the risk of abandoning an effective therapy that works on the less common pathway, Dr Kurzrock said.

Tailored treatment regimens may include drug combinations to hit the multiple targets involved in a patient’s cancer. Conducting clinical trials the traditional way would take a millennium to decipher the optimal drug combination for a cancer, because the number of 2-drug combinations would approach 45,000, assuming 300 effective oncology drugs, and would exceed 4 million 3-drug combinations.

Genomic Profiling: Impact on Outcomes
Enriching clinical trial populations for molecular targets is more efficient at finding successful treatments. Genomic profiling has resulted in superior response rates compared with the traditional clinical trial paradigm in unselected patients.

Before the genomic era, complete response rates obtained with therapies approved for solid tumors based on clinical trial results in unselected patients were typically near 0%, and the survival gains were in the range of 1 to 3 months. In contrast, when phase 2 clinical trials of single agents in the treatment of lung cancer were enriched for patients with putative molecular drug targets, the median response rate improved to 49% versus 9.7% in clinical trials of unselected patients, and overall survival reached 11.3 months versus 7.5 months in trials of unselected patients.

The P13K/AKT/mTOR pathway is activated in a subset of a large number of cancers. In some patients, other pathways, such as the MAP kinase, may also be activated at the same time and may form a resistance pathway. Many drugs target the P13K/AKT/mTOR pathway, and others target the resistance pathway. “There are trials in which both pathways are targeted simultaneously by 2 drugs that are matched to target part of each pathway,” said Dr Kurzrock. “Pathways are a different way of looking at cancer rather than organ of origin. We are looking at it by the molecular driver.”

A PIK3CA mutation is evident in 10% of advanced cancers originating in various organs. Approximately 30% of heavily pretreated patients with PIK3CA-mutant gynecologic and breast metastatic malignancies exhibit a partial response when treated with inhibitors of P13K/AKT/mTOR. Patients with the PIK3CA H1047R mutation are more amenable to treatment than patients with other mutations, with a 38% response rate. “We can’t be simplistic enough to just say there’s a mutation in a gene,” Dr Kurzrock said. “We have to know precisely where that mutation occurs. It may make a big difference in how the disease behaves, and what therapies it will respond to.”

Mutations in PTEN can also activate the PIK3CA pathway. Patients having either of these mutations are much more likely to have the resistance pathway (ie, KRAS, BRAF) also activated, adding another layer of complexity to treatment. “If we target the PIK3CA pathway,…we have to look for mutations in the resistance pathways and give combination therapy to those patients,” Dr Kurzrock said.

The complexity of cancer means that clinical trials of new targeted agents will also have to be smaller to improve the chance of finding effective agents in patients with particular subsets of cancer. Phase 1 cancer trials should be redesigned to include patients with the relevant mutations or genetic defects, believes Dr Kurzrock.

Example in Colorectal Cancer
FOCUS4 is an integrated clinical trial program of parallel, molecularly stratified randomized comparisons for patients with advanced or metastatic colorectal cancer who are fit for first-line chemotherapy. The trial design exploits a “window of opportunity” to test the clinical efficacy of targeted agents after first-line chemotherapy but before resistance to standard agents occurs in prespecified biomarker-defined subgroups, explained Rob Coleman, MBBS, MD, Director, Sheffield Cancer Research Centre, United Kingdom.

The program is designed to be “adaptable to new biomarker and clinical data as they proceed,” Dr Coleman said. “We want to evaluate multiple treatments and biomarkers in the same protocol. We don’t want to have to write a new protocol every time there’s a new drug or biomarker, because of the delays that brings in.” Each biomarker and treatment has its own control group.

The goal is to look for early signals of efficacy and also early stopping rules for lack of efficacy. “We’re not looking for small differences with targeted therapy…we’re looking for hazard ratio of 0.4 to 0.6,” said Dr Coleman.

The adaptive design allows for interim analyses to test new hypotheses, biomarker cohorts, and agents after patients are randomized. Select arms of the study can be closed when a conclusion is reached.