Carcinoma of unknown primary (CUP) is a rare disease—representing less than five percent of all cancers—in which malignant cancer cells are found in the body, but the place where the cancer began is not known. In an effort to improve treatment options, Mayo Clinic and the North Central Cancer Treatment Group (now the Alliance for Clinical Trials in Oncology) conducted a prospective, phase II trial, combining everolimus, an inhibitor of the mTOR protein, with carboplatin and paclitaxel for CUP patients. This trial was conducted across 24 sites in the Alliance for Clinical Trials in Oncology (NCT00936702). Sites obtained institutional review board approval, and patients provided written informed consent.
The trial evaluated whether a molecular test that identifies tissue of origin (TOO) could be helpful to identify responsive patients. The results, published in Annals of Oncology, demonstrated promising antitumor activity of this combination and, additionally, showed that a biomarker associated with TOO could identify chemotherapy-responsive patients.
“CUP tumors can be difficult to treat because they tend to lack identifying characteristics and appear undifferentiated under the microscope,” says Matthew Goetz, M.D., Professor of Oncology and Pharmacology at Mayo and senior author of the article. “So there is great interest in not only identifying the tissue of origin but using that information to select the best systemic therapy.”
While the combination of carboplatin and paclitaxel is considered a standard treatment for CUP, some tumor types that encompass CUP (e.g., pancreatic or biliary tract cancer) are poorly responsive to this regimen. Therefore, the prevailing dogma is that if the tissue of origin could be correctly identified, chemotherapy treatments tailored according to the tissue of origin would result in better response rates and survival for CUP patients.
“If you don’t know where the cancer came from, it’s unlikely that you’re going to know how to treat it, and so you’re going to use nonspecific, standard chemotherapy such as carboplatin and paclitaxel,” says Dr. Goetz. “Ideally, if you can identify where this cancer came from, you can provide, or bring forward, the best treatment. So that’s really the underlying premise of trying to find where these tissues arise from.”
In this study, Dr. Goetz and colleagues evaluated whether a gene expression profiling (GEP) test that predicted TOO could identify chemotherapy-responsive patients, using tumor tissue obtained prior to drug treatment. After central review by a Mayo pathologist to confirm a diagnosis of CUP, the TOO test was performed and categorized patients into two groups: those in which the TOO test indicated a type of tumor known to respond to carboplatin and paclitaxel (e.g., breast and ovarian cancer) and those tumor types in which carboplatin and paclitaxel were known to not be efficacious (e.g., pancreatic and colon cancer).
Results demonstrated that response rates, progression-free survival, and overall survival were substantially better in patients with the predicted tumor types that were known to respond well to carboplatin and paclitaxel. The value of this assay is that it accurately identifies patients into two groups of responsive and unresponsive. At this time, though, survival for patients whose tumors will not likely respond to carboplatin and paclitaxel still remains low; scientists continue to research new treatments for patients with these types of tumors.
“What’s still missing is a prospective randomized trial to determine whether chemotherapy or a targeted treatment, when delivered according to the tissue of origin, leads to better outcomes compared to an ‘unselected’ treatment,” says Dr. Goetz.
However, he believes that the best way forward will be to enroll CUP patients into clinical trials such as the National Cancer Institute MATCH (Molecular Analysis for Therapy Choice) Trial, in which comprehensive genomic sequencing is performed, and patients are then treated with novel drugs and/or drug combinations based on specific molecular alterations identified from tumor sequencing.