By Monica Heger
While clinical sequencing made great strides in 2011, it's arguable that it made the biggest impact in the field of cancer. Multiple groups are aiming to launch or have already launched targeted sequencing-based cancer tests, while others, such as the University of Michigan, have implemented comprehensive sequencing pipelines for cancer patients. Additionally, a number of research studies have found that sequencing can help guide treatment for cancer patients.
Some molecular diagnostic companies entering the next-gen sequencing market for the first time are starting in the cancer space, including Asuragen, which launched several targeted sequencing-based cancer panels late last year on both the Illumina Genome Analyzer and the Ion Torrent PGM.
Vendors, too, are seeing an opportunity in the cancer field. Service provider Complete Genomics now offers a cancer sequencing service, charging $12,000 to sequence a tumor/normal pair to 80-fold coverage and $18,000 for a trio, such as tumor, metastasis, and normal samples from the same individual.
Illumina, meantime, offers sequencing under its Illumina Genome Network for $10,000 per tumor/normal sample, while Life Technologies' Ion Torrent business last year launched its AmpliSeq Cancer Panel, which targets 190 amplicons from 46 cancer genes.
Even the UK government has gotten on board, and is funding an initiative to bring sequencing-based cancer tests to the market (CSN 6/21/2011).
As these efforts have proliferated over the last year, however, a debate has emerged between two camps: those who argue for a comprehensive sequencing strategy and those who think that a targeted approach makes more sense.
In late 2011, the first CLIA-certified sequencing-based cancer test was launched by Washington University's School of Medicine through its new Genomics and Pathology Services Laboratory, a collaboration between the department of genetics and the department of pathology and immunology (CSN 11/30/2011).
Close on its heels was Baylor's targeted test using Ion's AmpliSeq Cancer Panel, which it planned to release in December from its CLIA-certified, CAP-accredited laboratory for under $2,000 (CSN 11/30/2011).
The WashU lab offers a 28-gene cancer panel, as well as separate panels that will sequence a subset of genes for specific cancer types including brain, colon, lung, ovarian, and endometrial cancers.
The 28 genes were chosen because they are already well known to be clinically actionable, and include all genes for which there are drugs available to target.
According to Shashikant Kulkarni, medical director of GPS, this approach makes the most sense for getting actionable results to physicians and impacting as many patients as possible.
More comprehensive approaches, like whole-genome sequencing or exome sequencing, while useful for research, are not yet practical for the clinic because they are too expensive, take too long, and most of the results are not medically actionable, he argued.
Other researchers agree, even some who were initially sold on a more comprehensive sequencing strategy. The Fox Chase Cancer Center, for instance, was scheduled to begin offering exome and transcriptome sequencing last fall for cancer patients on Life Technologies' SOLiD instrument.
But, it recently decided to switch to a targeted sequencing approach on the Ion Torrent PGM because from a "patient and clinical perspective" that was the way to go, Jeff Boyd, executive director of Fox Chase's Cancer Genome Institute, told Clinical Sequencing News at the time (CSN 12/21/2011).
According to Matthew Ferber, assistant professor of laboratory medicine and pathology at the Mayo Clinic who is developing an 18-gene panel to diagnose subtypes of hereditary colon cancer, creating a diagnostic test with a clear focus was critical for getting physician support. Also attractive to physicians was the panel's ability to test for all subtypes in one single test, rather than an array of tests that might still not yield a clear diagnosis.
Even with the 18-gene panel, some physicians were still "apprehensive" about how they would deal with and interpret all that data, Ferber said (CSN 5/10/2011).
Researchers at the University of Washington last year launched a similar test for hereditary colon cancer. The UW test, ColoSeq, is even smaller, targeting just seven genes that have been routinely tested in single-gene tests (CSN 12/7/2011).
The broadest of these cancer panels will likely be from Foundation Medicine, which is planning to launch a test for a range of cancers, based on a 200-gene panel, in the second quarter of this year (CSN 6/8/2011).
Researchers doing comprehensive sequencing of cancer patients, which often includes a combination of whole-genome, exome, and transcriptome sequencing, argue that because the cancer genome is so mutated and often includes structural variants, a gene panel will inevitably miss important mutations.
Additionally, because cancer patients often spend tens of thousands of dollars on treatment, a sequencing test that costs several thousand dollars is still reasonable if it can guide treatment and potentially save money in the long term, or at least ensure that patients receive an effective treatment from the outset.
While most groups have implemented such programs on a very small scale, the University of Michigan this year launched a program dubbed Mi-OncoSeq, and expects to enroll around 100 patients over the coming year (CSN 12/7/2011).
The Mi-OncoSeq protocol originally involved low-pass whole-genome sequencing of tumor DNA, exome sequencing of tumor and matched normal samples, and transcriptome sequencing of tumor DNA. However, because most of the relevant mutations were found in the exome and transcriptome data, the team may drop the low-pass whole-genome sequencing in order to reduce the cost.
The Mi-OncoSeq protocol is currently paid for through university funding, and the reagent cost is estimated to be $3,600 per patient.
According to Arul Chinnaiyan, who heads the program, because of the comprehensive sequencing approach the team has identified clinically relevant mutations they likely would not have identified using a targeted approach.
For instance, in a patient with colorectal cancer, the team found a relevant mutation in a gene not commonly associated with colon cancer. Additionally, the transcriptome data has been particularly useful for identifying druggable gene fusions and rearrangements that would likely be missed with gene panels.
Other groups have attested to transcriptome sequencing's value. For example, Elaine Mardis and her team at the Genome Institute of Washington University found that transcriptome sequencing can home in on only mutations that are expressed in the tumor. At last year's Personal Genomes meeting at Cold Spring Harbor Laboratory, she presented data from a patient where only 40 percent of the identified mutations were actually expressed (CSN 10/5/2011).
Like Chinnaiyan, Mardis said that transcriptome sequencing has been especially useful for identifying drug targets. Aside from ruling out potentially actionable mutations because RNA-seq finds that they are not actually expressed in the tumor, the method has identified targets that were not found by whole-genome sequencing.
For instance, whole-genome sequencing of a patient with acute lymphocytic leukemia identified two major subpopulations of tumor cells, but with no obvious drug target. After looking at the transcriptome data, however, the Wash U researchers identified overexpression of a gene that could be targeted and subsequently started the patient on the drug. Mardis said at the time that the patient showed significant improvement just three days into treatment.
Whole-genome sequencing has also shown that it can identify potentially novel indications for existing drugs. A study using next-gen sequencing in triple-negative breast cancer patients has led to the initiation of a phase I clinical trial of combination MEK/AKT inhibitors, which had not previously been tested for this indication (CSN 12/14/2011).
Perhaps another reason some groups opt for targeted sequencing over whole-genome, exome, or transcriptome sequencing is that it reduces the likelihood of running into some thorny ethical issues, such as what to do when variants are found that may be relevant not only for the patient, but the patient's family.
In one of the first papers demonstrating that whole-genome sequencing of cancer patients to guide treatment is feasible, Wash U researchers uncovered de novo mutations to cancer genes in a patient's germline DNA, indicating she had hereditary cancer. While the patient had passed away years before, she had children for whom that information would be relevant.
Wash U has in place a "moveable firewall" that allows clinically relevant variants uncovered in a research setting to be conveyed to a patient's physician while still preserving the patient's anonymity. Nevertheless, speaking about the incident at the CSHL Personal Genomes meeting, Wash U's Jennifer Ivanovich said that specific case was not a success in terms of demonstrating how sequencing results should be returned (CSN 10/12/2011).
The UMich group is dealing with the issue by giving patients a choice about what data is returned, so patients can choose not to receive information about germline variants or variants that are predictive of other diseases.
2012 and Beyond
Looking ahead, it seems that there is no consensus in sight as to which approach is ultimately the best. While falling sequencing costs have been predicted to push all sequencing toward whole-genome, for cancer, both the targeted and comprehensive approaches seem poised to grow.
Just last week, for example, Myriad Genetics said it plans to develop targeted next-gen sequencing tests but does not think that whole-genome sequencing is ready for the clinic (CSN 1/4/2012).
Additionally, researchers at the University of California, San Diego, are moving forward with a clinical trial using a targeted sequencing approach dubbed UDT-Seq (CSN 12/21/2011).
Other groups remain committed to the comprehensive approach, however. Researchers at the Translational Genomics Research Institute are in the midst of a three-year pilot study using whole-genome and transcriptome sequencing of patients with rare or unusual cancer to test the technology's ability to guide treatment. Last year, the team reported results from one patient, which identified a second-line treatment option (CSN 7/13/2011).
And the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University is slated to open a personalized cancer center this year (CSN 9/7/2011).
Regardless of the strategy, it is clear that sequencing for cancer patients is quickly moving into the clinic. As Illumina CEO Jay Flatley noted this week at the JP Morgan Healthcare Conference in San Francisco, in the next few years it will become "almost irresponsible" for a physician to not have a patient's tumor sequenced.
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.