Gene Profiling of Circulating Tumor Cells May Predict Treatment Response in Pancreatic Cancer

Kenneth Yu, MD

A simple blood draw could lead to life-prolonging adjustments in treatment for patients with pancreatic cancer, the authors of a study have found.

By profiling the gene expression of circulating tumor cells in blood, the investigators have demonstrated, it is possible to predict whether specific patients with pancreatic cancer will respond to standard chemotherapy combinations, are becoming resistant to such treatments, or are experiencing progression of their disease. The results of the ongoing prospective study, being conducted at Memorial Sloan-Kettering Cancer Center in New York, were presented at the 10th Annual Gastrointestinal Cancers Symposium held January 24-26 in San Francisco, California. The findings may help physicians choose the most effective therapies for individual patients and, when necessary, switch to different treatments. Picking effective treatments is crucial in pancreatic cancer because the disease is uniformly lethal, with survival <5% at 5 years and cytotoxic drugs the only current method for extending life, the authors wrote. The blood tests could also eliminate the need for repeated tumor biopsies, which can be difficult to obtain in patients with the disease, ASCO wrote in a statement released prior to the conference.

While there have been previous studies of pharmacogenomics modeling in other cancers, this is the first to examine the response to a cocktail of drugs, ASCO reported.

“We’re making some progress in developing more effective treatments; in the last 3 years, a drug cocktail, FOLFIRINOX, has brought dramatic improvement, and a combination of gemcitabine and nab-paclitaxel has been shown to improve survival, too,” said Kenneth Yu, MD, of Sloan-Kettering, an author of the study. “Now, the problem remains that we don’t have a validated biomarker to help us choose effective treatments for each individual patient we see. The purpose of this study was to see whether a pharmacogenomic analysis of circulating tumor cells could help.”

In the study, which has enrolled 50 of an anticipated 60 patients with unresectable pancreatic cancer, 10 mL of peripheral blood is collected prior to treatment with chemotherapy, and again at disease progression. Using a new model created by CellPath Therapeutics, in Baltimore, Maryland, investigators isolate tumor progenitor cells, extract total RNA, and perform gene-expression analysis to test how the variations in specific tumors affect response to 12 chemotherapy drug combinations.

Participating doctors are blinded to the results of the profiling, and are treating the patients as they normally would.

So far, investigators have analyzed expression differences between 30 patients with continued response to their treatments and 20 patients whose disease has progressed, Yu said. Analysis was performed at both the molecular pathway and individual gene levels.

While the results of the profiling were not used to guide treatment in the study, researchers found that 6 patients who received chemotherapies predicted to be effective had better outcomes than 8 patients who received treatments predicted to be ineffective (average time to disease progression 7.3 months vs 3.7 months, P =.017), Yu reported.

Among the 20 patients who experienced tumor progression, researchers found that chemotherapy sensitivity profiles had changed, suggesting the need for different treatment regimens. Specifically, disease progression and treatment resistance were accompanied by increased dysregulation in the Rb1 and PLC biological pathways and changes in the ErbB3, ARE/Nrf2, and insulin pathways, ASCO said in its statement. Among patients with stage IV cancers who experienced progression, major differences were seen in the E2F1 and NFκB pathways, ASCO stated.

“Hedgehog pathway overexpression was associated with resistance to gemcitabine, but clinical response to 5-FU-based treatment,” the authors noted.

The investigators concluded that this method of gene-expression profiling can be performed reliably in pancreatic cancer, and that it can predict treatment response and identify key pathways associated with treatment resistance.

CellPath is testing the model in other cancers, too, Yu said.

“Picking the right first regimen is really critical,” he said, “because if patients don’t respond, they may become too ill to get another line of chemotherapy.”

Yu added that the model could also help doctors identify patients unlikely to respond to any therapies.

“That would be important to know, as well,” he said, “so that doctors can work to improve the quality of life of those patients, instead of giving them drugs that are really toxic.”

Sangar V, Ricigliano M, O’Reilly EM, et al. Use of pharmacogenomic modeling in pancreatic cancer for prediction of chemotherapy response and resistance. Presented at: 10th Annual Gastrointestinal Cancers Symposium; January 24-26, 2013; San Francisco, CA. Abstract 142.

<<<

View coverage from the 2013 GI Cancers Symposium