CHICAGO (GenomeWeb News) — Duke University’s Institute for Genome Sciences and Policy this fall will launch a prospective Phase IIb trial to learn whether a pharmacogenomic-based prognostic model can be used to change the way oncologists treat an early form of non-small cell lung cancer.
Joe Nevins,director of the Institute’s Center for Applied Genomics & Technology, said that his lab has created a gene-expression model using Stage 1a NSCLC tissue samples that can distinguish patients in that population for whom chemotherapy would be more beneficial from those in whom the risks of the drugs outweigh their benefits.
He chose NSCLC because patients diagnosed with the disease in that early stage typically are treated with surgery or radiation. For these patients, a gene expression-based predictive model that can help guide treatment “sooner rather than later is important,” Nevins said. “We’re talking about refined prognosis: Who to treat, who doesn’t need to be treated in a particular context, and better utilization of existing drugs.”
Nevins, who is also professor of breast cancer genomics at Duke’s Department of Molecular Genetics and Microbiology, spoke at the American Society of Clinical Oncology’s annual meeting here, which began on Friday.
His research, a follow-up to a study published in 2006 in The New England Journal of Medicine, is in-line with a trend among drug and diagnostic makers to use genomic biomarkers to isolate patients with different genetic subtypes of a heterogeneous cancer. The ultimate aim is to match late-stage drug candidates with patients genetically more likely to respond to them — whether chemotherapy or targeted therapy — rather than killing the candidate when its efficacy wanes in the broader patient population.
For now, Nevins’ research focuses on existing chemotherapeutic agents. Not finding a use for a drug that may have a narrow clinical application “is an opportunity that is a significant one, and [one] that is often missed,” he said. “There’s a plethora of cytotoxins and targeted agents that work — [though] in many cases they don’t work well. The issue is, is there an opportunity to make better use of existing drugs?”
Invoking Herceptin and its companion diagnostic, Nevins said “we need a strategy like HER2-Herceptin for every single drug coming along.”
To be sure, there are manifold examples of pharmacogenomic-based prognostic models, but according to Nevins NSCLC presents a novel opportunity. Clinical trials have shown that patients with stage IB, II, or IIIA NSCLC benefit from chemotherapy following surgery. But in Stage 1A NSCLC, “there is no evidence for benefit of adjuvant chemotherapy,” he said.
However, he said, patients in this stage have a roughly 30 percent risk of recurrence, “so it’s clear that there is a population of patients that are at risk.” What this suggests is “the ability to tease out a population of patients whose clinical prognosis is essentially misclassified.”
Nevins said that the upcoming randomized Phase IIb trial, which will use study patients treated with adjuvant chemotherapy and those who are not, will seek to validate the prognostic model. It also will evaluate whether it is possible to identify individuals who would likely benefit from a batch of six commonly prescribed adjuvant chemotherapeutic agents “as if [the patients] were diagnosed at Stage II or Stage III.”
Nevins said that the next step would be to devise methods to more accurately predict which patient population would likely respond to these drugs.