GI Specialist Discusses Hurdles in Targeting Gastric Cancers

Alok A. Khorana, MD

Alok A. Khorana, MD, specializes in predictive factors in gastrointestinal oncology and in cancer-associated thrombosis at the Cleveland Clinic, which he joined in 2013. Khorana is the Sondra and Stephen Hardis Chair in Oncology Research and vice chair (Clinical Services) and director of the GI Malignancies Program at the Taussig Cancer Institute, as well as a professor at the Cleveland Clinic Lerner College of Medicine. Previously, he served as an associate professor of Medicine at the James P. Wilmot Cancer Center of the University of Rochester. He provided insight on gastric cancer research in an interview with OncLive.

OncLive: What are the most promising targeted therapies for the treatment of gastric cancer?

Dr Khorana: Ramucirumab and trastuzumab are, of course, approved and are the only currently available targeted agents in this setting. Probably the most impressive recent data are for AMG 337 [a MET inhibitor] which, in early-phase studies, has shown excellent responses in MET-amplified patients. Note that these data represent only a subgroup of an early-phase study and further validation is necessary. An alternative MET inhibition approach using a monoclonal antibody failed in more advanced-phase trials.

In your opinion, why did ramucirumab succeed where bevacizumab failed and are there any other antiangiogenic therapies that hold promise in gastric cancer?

Although both drugs target similar pathways, they may interact differently with angiogenic receptors. In addition, the study populations were somewhat different. Indeed, in the bevacizumab study, if only the subgroup of North American patients was analyzed, bevacizumab did appear to provide benefit. So there may be subtle issues related to ethnicity and staging that differ by geographic location and are incompletely understood.

What are the reasons behind the failure of EGFR-targeting therapy?

It is unclear; EGFR-activated pathways may not be as important as we had previously thought. Or they may be redundant with other pathways.

What insights do recent molecular characterization studies offer about how best to treat gastric cancer?

We have been treating many cancers, including gastric, by categorizing patients (and treatments) based on anatomic location. This may, however, not turn out to be the best way to categorize.

In gastric cancer, for instance, patients with lower esophageal cancer invading into the gastroesophageal junction may have quite different pathophysiology than patients with body tumors—the former are likely caused by reflux/ Barrett esophagus and the latter related to Helicobacter pylori (among other influences).

So really what we think of as “gastric cancer” is a conglomerate of several different cancers entirely, yet due to a lack of better knowledge we have been grouping these together and treating them similarly. The Cancer Genome Atlas project suggests just as much, dividing gastric cancer into four distinct subtypes. This molecular classification provides clues to pathophysiology (important for prevention in the future) as well as differing approaches to treatment based on which pathways are upregulated.

What are the most significant hurdles to the development of targeted therapy in gastric cancer?

The lack of appropriate biomarkers to identify patients who benefit from targeted agents is a major challenge. There are no biomarkers to predict therapeutic benefit from ramucirumab. Similarly, issues with MET amplification versus immunohistochemistry (competing assays)^a may have been one of the potential causes for the failure of the MET-directed antibody mentioned previously.

^aNote: Several studies have suggested that the negative findings observed in trials of MET-targeted therapies may be related to difficulties in testing for MET aberrations. A significant number of patients who test positive for MET expression on immunohistochemistry (IHC) do not have MET-amplified disease. IHC or MET gene copy number may not be optimal for detecting patients potentially sensitive to MET-targeting strategies.