By Julia Karow
The Centre for Clinical Diagnostic Genomics is getting ready to deliver next-gen sequencing-based cancer diagnostics to patients in British Columbia, Canada, Clinical Sequencing News has learned.
This spring, the CDG, which is housed in the Michael Smith Genome Sciences Centre of the BC Cancer Agency in Vancouver, will start offering breast cancer susceptibility testing on the Illumina HiSeq platform. A 14-gene panel testing for predisposition to several cancers is scheduled to follow next year. In parallel, the center is exploring genome-wide sequencing for tumor diagnostics.
The CDG, which receives funding from Genome British Columbia, the BC Cancer Foundation, and the Provincial Health Services Authority, was set up a year and a half ago and aims to "implement emerging technologies to provide the most cost-effective clinical testing for genetic disease," according to its website.
So far, it has provided BRCA 1 and BRCA 2 testing by capillary sequencing, analyzing about 500 samples per year. But the intent has always been to move that test over to a next-gen sequencing platform and to add other tests in the future, according to Aly Karsan, who heads the center.
Following a validation study that is mostly complete, which compared the test on the Sanger and Illumina platforms, the CDG's lab will start running the BRCA 1/2 test on the Illumina HiSeq this spring, pooling 24 indexed samples per flow cell lane, including controls. The genes will be amplified by RainDance Technologies' microdroplet PCR platform, and all potentially pathogenic variants will be validated by targeted Sanger sequencing, as they are for the current, Sanger-based test.
"We're comfortable in being able to deliver all the variants that we need to with a high level of coverage and have an accurate result at the end of it," said Karsan.
Running the test on the HiSeq will decrease both its cost — by an estimated 30 percent to 50 percent — and increase the lab's capacity. The lab has seen a 10 percent increase in requests for the BRCA test every year, Karsan explained. It receives a set amount of funding to provide the test, irrespective of the method it uses, providing an incentive to reduce its cost.
The CDG's lab has been accredited by the College of American Pathologists, and has so far validated two HiSeq instruments for clinical sequencing, though the plan is to validate all of the dozen or so HiSeqs of the Genome Sciences Centre, so whichever machine is available could be used for clinical work.
Validating the instruments has taken some time. "We are basically trying to use [CAP] guidelines that have been set up for capillary sequencing and apply them to the next-gen pipeline," Karsan said. As of now, "there are no specific guidelines from any regulatory body as to what's required for using any type of next-generation sequencing technology," so he and his team are applying the existing ones.
CAP has formed a workgroup to create an accreditation checklist for next-gen sequencing in clinical labs that it hopes to have in place by the end of 2012 (CSN 9/14/2011).
Moving the BRCA test to next-gen sequencing will serve as a proof of concept for larger test panels. The center has already begun to develop a 14-gene panel for cancer predisposition genes, mostly for breast, ovarian, and colorectal cancer, which it hopes to test on 100 patients and start offering clinically sometime in 2013. It is evaluating the RainDance platform for that test as well, along with other gene selection methods. "We are trying to make sure that we are completely confident in anything we put out," Karsan said. Beyond that, the CDG is also looking into gene panels for other cancer types.
In designing the panel, the researchers took input from medical geneticists and genetic counselors and considered whether it would be cost effective to include rare genes. "We're trying to balance cost and turnaround time with accuracy and compliance with regulatory bodies," Karsan said.
In another project, the center is comparing three types of genome-wide testing — transcriptome, genome, and exome sequencing — in acute myeloid leukemia in order to see if such a test can identify variants that are currently used in clinical management, as well as new markers that are being evaluated in clinical trials. This project includes a retrospective study of 184 AML patients in which the outcome is already known, and a one-year prospective trial involving 80 patients that will start this summer.
For the prospective trial, the researchers plan to report clinically actionable variants back to physicians. However, "the trial is meant as a diagnostic trial. It won't impact clinical management," Karsan said. "It's meant to define what we can pick up and report that's clinically actionable and make sure that we are able to deliver, in a timely fashion and in a way that corresponds with standard testing, all the variants that are necessary."
The AML project will also be using the HiSeq platform because the MiSeq is currently "not tractable" for whole-transcriptome or whole-exome sequencing, though it might be in the future, he said.
One of the greatest challenges so far in developing next-gen-based clinical sequencing tests has been the bioinformatics, mostly because it changes so much. "There are multiple types of algorithms one could use, and they are constantly being tweaked, and in a clinical environment, one needs to lock down the process, [and] changes need to be documented," Karsan said. For that reason, the lab has started to be "a lot more stringent" about its use of bioinformatic tools.
Following the AML study, the center hopes to embark on a number of solid tumors, Karsan said. "The goal is that we are able to impact the majority of patients" by identifying variants in their tumors that will help stratify them to known therapies, by identifying patients who would qualify for newer experimental therapies, and, through informatics approaches, by predicting therapies that are currently not used for the patient's type of cancer. "Going a genome- or transcriptome-wide route would allow us to approach this in a multi-tiered way," Karsan said.
Have topics you'd like to see covered in In Sequence? Contact the editor at jkarow [at] genomeweb [.] com.