Personalized Therapy Arrives for Patients with Colorectal Cancer
by David W. Sharp

Earlier this year the American Society of Clinical Oncology (ASCO) offered as a "provisional clinical opinion" the statement that "all patients with metastatic colon cancer should have their tumors tested for KRAS mutations" (J. Clin. Oncol., 27[12]: 2091-6, 2009). Others have said much the same thing—for example, it should be "standard practice" to restrict colon-cancer treatment targeted at the epidermal growth factor receptor (EGFR) to patients whose tumors test positive for wild-type (non-mutated) KRAS (D.Z. Chang, et al., J. Hematol. Oncol., 2:18, 2009).

The National Comprehensive Cancer Network clinical guideline on colon cancer reflects this change, too; in the U.K. the documentation for a clinical trial of chemotherapy with or without the EGFR-targeting drug cetuximab that is seeking to recruit patients with advanced bowel cancers is excluding those with KRAS mutations; official labelling for cetuximab and another drug that targets EGFR has lately been altered to reflect the change in recommended practice. Furthermore, KRAS tests are being widely marketed. Three papers in the Journal of Clinical Oncology that are in the Top Ten (#8) or hover outside that list (S. Khambata-Ford, et al., 25[22]: 3230-7, 2007; total cites 174, latest count 36, at #13; and A. Lievre, et al., 26[3]: 374-9, 2008; total cites 139, latest count 36, at #14) are being widely quoted because the data in them contribute significantly to accumulating evidence for this "step toward personalized medicine for patients with colorectal cancer" (Y. Jiang, et al., Cancer, 115[16]: 3609-17, 2009). The most recent of this trio appeared in April, 2008, but the evidence keeps rolling in—see, for example, C.S Karapetis, et al., from October of the same year (New Engl. J. Med., 359[17]: 1757-65, 2008).

KRAS is a gene, the human homolog of the Kirsten rat sarcoma-2 virus oncogene, that is linked with cellular signalling pathways, including those involving EGFR itself. The drugs active against EGFR include cetuximab and panitumumab. Dr. Rafael G. Amado and colleagues (#8) tested for KRAS mutations as part of a randomized trial comparing panitumumab and best supporting care in patients with metastatic colon cancer. Any effect of the drug was limited to those with wild-type KRAS. In #13 the drug was cetuximab and other markers (two EGFR ligands) were studied but the result for KRAS was similar to #8’s: patients whose tumors lacked KRAS mutations had significantly greater disease control from cetuximab. And a similar message emerged from the third paper (#14).

Might possession of a mutated KRAS by itself contribute to a poorer outlook for patients with advanced bowel cancer, irrespective of any treatment aimed at EGFR? Seemingly not. A presentation at this year’s ASCO meeting (C. Fuchs, et al., J. Clin. Oncol., 27[15s]: abstr 4037, 2009) drew on data from a chemotherapy trial starting 10 years ago and not involving EGFR-targeting agents. Possession of wild-type as opposed to mutated KRAS had no effect on disease-free or overall patient survival. A poster at a symposium on gastrointestinal cancers held during ASCO 2009 (V. Shankaran, et al. ) provided an answer to a different question: Since KRAS testing is not cheap and since drugs like cetuximab are expensive, does a policy of not giving such drugs to those whose tumors are wild-type KRAS make sense in terms of economics as well as clinically? It seems that the cost of KRAS testing incident cases (estimated at $13 million a year for the US) is very much less than the savings made by not giving drugs predicted to be of no benefit to the recipient. A health economic analysis of a large Canadian trial, noting the daunting added costs involved when cetuximab is added to best supportive care alone, found the impact much reduced for patients with wild-type KRAS (N. Mittmann, et al., J. Natl. Cancer Inst., 101[17]: 1182-92, 2009).

A former deputy editor of The Lancet, Mr. David W. Sharp, M.A. (Cambridge) is a freelance writer living in Minchinhampton, U.K.