By: Bruce Quinn, MD, PhD

Several conferences held last summer provided a basis for the notion that a relative gold rush of companion diagnostics in oncology is nearing reality. However, cancer biomarkers also present unique complexities for the lab industry, regulators, and physicians.

The American Society for Clinical Oncology (ASCO) focused its 2009 meeting on the idea that a personalized medicine approach must become the standard for effective clinical oncology. Three years ago, Her-2/neu stood out as one of few widely used companion diagnostic tests in chemotherapy. The Her-2/neu test selects breast cancers more likely to respond to the monoclonal antibody trastuzumab (Herceptin). With a ten-year history, the Her-2/neu test served as the poster child of companion diagnostics. It was cleared by FDA, and it played a pivotal role in treating a common cancer. The test even made an appearance in a major television movie.1 Those who attended personalized medicine conferences in 2007 or 2008—when they were becoming increasingly frequent—heard the Her-2/neu story again and again. By 2007, Her-2/neu was known not only for its successes but for its problems with clinical accuracy.2

One also-ran for the role of next new thing was UGT1A1 testing for irinotecan (Camposar), which has some value for predicting a subpopulation of poor metabolizers who are prone to suffering excessive side effects from the drug. But some of the most toxic side effects of irinotecan were not among those predicted by UGT1A1 genetics, and clinical uptake was weak.

We now see increasing attention being paid to several genes that will make a big impact on clinical oncology care. Both the promises and difficulties for these new genes were highlighted at a major Institute of Medicine conference held in June and at other personalized medicine conferences this summer.3 One such gene, KRAS, lies in the cellular pathway that is affected by cetuximab (Erbitux) and panitumumab (Vectibix), such that these monoclonal therapies are ineffective in colon cancer patients with certain KRAS mutations. In July, FDA added a recommendation for KRAS testing to the labels of Erbitux and Vectibix. Just last December, FDA staff had been reluctant to recognize retrospective but convincing studies that showed KRAS played a major role in this treatment paradigm.4 The monoclonals getifinib (Iressa) and erlotinib (Tarceva) target the EGFR receptor in lung cancer patients, and EGFR mutations pick out the subgroup of patients in which these agents work.5 More genes in the colon and lung cancer signaling pathways are being detected. This news is good for patients and for the profile of the molecular diagnostics industry. So what are the challenges? The answer lies in the execution of the assay of these genes, which is complicated.

There are diverse methodologies and mutation sites for KRAS as well as for the EGFR genes. Issues for testing KRAS were described in CAP Today, published by the College of American Pathologists.6 FDA added KRAS testing to Tarceva and Vectibix labeling, but there is no FDA-cleared test at this point. Different aspects of the EGFR pathway need to be measured in lung cancer versus colon cancer. This places a heavy information burden on oncologists, and good education or information systems at the laboratory level will be critical to ensuring appropriate patient testing and care. For the EGFR gene test,7 sequencing analysis, hybridization screening analysis, and PCR analysis each have benefits and trade-offs.

Medicare recently announced it will hold a public advisory board on cancer companion diagnostics in January 2010.8 The molecular diagnostic industry has a high-profile opportunity to show it can meet these challenges effectively over the next two years.