Tailoring Therapy to Hidden Primary Tumors

A tumor's location usually guides what treatment a patient receives, but as many as 9% of cancer patients are diagnosed with carcinoma of unknown primary site (CUP). In these cases, the primary site of origin for the tumor remains uncertain despite extensive—and often expensive—searches using multiple types of imaging, invasive procedures, immunohistochemistry, mRNA expression profiling, and other techniques.

Because molecular profiling and gene sequencing of CUPs has only begun to emerge, oncologists usually lack the information necessary to tailor treatments to the patient, so the standard therapy for CUP involves cytotoxic drugs such as paclitaxel and carboplatin. The 5-year survival for this nonspecific approach is only 11%.

However, oncologists may be able to personalize therapy without locating the primary tumor, according to a team led by Jeffrey Ross, MD, of Albany Medical College in New York and Foundation Medicine of Cambridge, MA. To determine whether genomic profiling of CUP samples could reveal mutations that would guide treatment, the researchers sequenced 255 genes in tumor samples from 200 patients with CUP. Sequencing detected base substitutions, insertions, and deletions, amplifications, and some fusions and rearrangements.

Of the 200 tumor samples, 85% had a genomic alteration “that could potentially guide decisions for targeted treatment,” the researchers reported last month in JAMA Oncology. Potential treatments include already-approved drugs as well as those available through a clinical trial. For example, 3% of the tumors contained either of two mutations that activate EGFR and that predict a positive response to erlotinib (Tarceva; Genentech) and afatinib (Gilotrif; Boehringer Ingelheim Pharmaceuticals) in non–small cell lung cancer patients. In addition, 5.5% of the samples showed mutations in BRAF; patients with melanoma who have these mutations receive drugs such as vemurafenib (Zelboraf; Roche) and dabrafenib (Tafinlar; GlaxoSmithKline).

“Comprehensive genome profiling of CUP leads to significant opportunities for patients to be treated with targeted therapies,” says Ross. The next step, he says, is to launch clinical trials to determine which treatments work for CUP patients with particular mutations. In the meantime, he says, doctors should consider prescribing targeted therapies off-label for these patients.

“This is the first published, peer-reviewed study of a large group of CUP patients profiled like this,” says F. Anthony Greco, MD, of the Sarah Cannon Research Institute in Nashville, TN.

However, Greco and his Sarah Cannon Research Institute colleague John Hainsworth, MD, say that the clinical relevance of the analysis remains unclear. Because there are no approved treatments for most of the mutations the study identified, Hainsworth estimates that just 15% to 20% of patients might benefit from the approach. “Even so,” he says, “15% to 20% is important.”

Clinical trials to determine if treatments suggested by genomic profiling help CUP patients will be difficult to undertake because a large number of subjects would be required for an adequately powered trial, says Gauri Varadhachary, MD, of The University of Texas MD Anderson Cancer Center in Houston. As an alternative, she says, researchers could study CUP patients who are taking part in basket trials such as M-PACT and MATCH, in which participants receive treatments tailored to their genomic alterations, not their cancer type.