Skip to main content
Premium Trial:

Request an Annual Quote

Cancer Consortia Members Using Deep Sequencing, Other Strategies to Address Pancreatic Cancer

By Andrea Anderson

MONTREAL (GenomeWeb News) – Research teams participating in The Cancer Genome Atlas and the International Cancer Genome Consortium are doing deep exome sequencing — and testing various enrichment strategies — to deal with some of the challenges associated with characterizing pancreatic cancer, attendees of the International Congress of Human Genetics heard at a session on cancer genomes here today.

Pancreatic cancer poses a particular problem for researchers involved in large-scale sequencing studies of cancer, in part because it is difficult to get a hold of enough samples that are the size and quality needed to do these studies, noted Broad Institute researcher Stacey Gabriel.

During her presentation, Gabriel described progress being made by members of TCGA overall and spoke in more detail about TCGA efforts that are underway at the Broad, including studies of pancreatic cancer.

Pancreatic cancer has a notoriously low survival rate, with only two percent or so of individuals surviving five years after being diagnosed with the disease. And some of the same features that may contribute to its deadliness — such as genetic heterogeneity and stromal cell contamination — have also complicated efforts to get at the genetic underpinnings of the disease.

In some cases, Gabriel noted that researchers have run into problems meeting the TCGA's quality control criteria, since many of the pancreatic cancer samples coming into the TCGA biospecimen centers don't meet the required tumor cellularity level of at least 60 percent due to stromal cell introgression into the tumor.

Moreover, the number of primary tumor samples available is generally low — in part because pancreatic cancer tends to be detected at later stages, often after it has metastasized to other parts of the body.

Treatment response also tends to be quite poor, Ontario Institute for Cancer Research's Genome Technologies Director John McPherson explained during his own presentation at the ICHG cancer genomes session, noting that surgical resections to remove primary tumors are typically done in only around 15 percent of pancreatic cancer cases.

McPherson's team is involved in one of the three ICGC projects that are aimed at characterizing pancreatic cancer. So far the researchers have collected 113 matched tumor-normal samples from individuals with pancreatic cancer, he said. Of these, around two-thirds of the samples have entered their analysis pipeline. Over the next few years, the researchers plan to assess 350 pancreatic cancers for ICGC.

An Australian team is sequencing another 350 pancreatic cancers through ICGC, McPherson noted, while a Baylor College of Medicine-led team is sequencing additional pancreatic cancers.

Given the limited number of pancreatic cancer samples available that meet the ICGC's prospective collection requirements, McPherson explained, investigators are working hard to come up with strategies to make use of samples that are available. To do this, they have not only been looking at the possibility of doing deep sequencing of the samples, but have also been testing several strategies to enrich for tumor DNA prior to sequencing.

Among these strategies are pancreatic ductal adenocarcinoma "coring" to separate tumor from stromal tissue using a biopsy needle, antibody-based enrichment, and xenograft experiments in mice to generate more and more homogenous tumor samples.

Though they are still working out some of the details and limitations of the approaches, the researchers have been able to find recurrent mutations in 35 pancreatic cancer exomes using a combination of primary and xenograft tumor data, for instance.

At the Broad Institute, meanwhile, researchers participating in TCGA have embarked on a pilot effort to do deep sequencing of 14 tumor-normal pairs using gross bulk tumor samples.

In their preliminary studies, the team found pancreatic tumor samples with negligible copy number changes or mutations, Gabriel said. But at a depth of 300 to 400 times exome coverage, they have started to pick up some potential mutations present at low frequency in the tumor material.