By Julia Karow
A year after closing its first venture capital financing round, personalized cancer genomics startup Foundation Medicine has developed a clinical sequencing-based test to analyze several hundred cancer genes in tumors for mutations. It plans to commercialize the test to help doctors make treatment decisions.
The company is currently using the test in a pilot collaboration with Novartis and plans other partnerships with pharmaceutical firms and academic medical centers to demonstrate its analytical validity and clinical utility. After that, it wants to make the test widely available and expand its content over time, though it has not yet provided timelines for these goals.
Foundation Medicine is based on the belief that an increasing number of cancer patients and their doctors "would benefit from a deep understanding of the genomic and other molecular aberrations that are driving a patient's specific tumor," said Alexis Borisy, the company's founding CEO.
Just like a CAT scan or MRI can help locate a tumor for radiation treatment, "the vision here is that it is going to become standard of care over the next 5 to 10 years to have a 'molecular MRI' of the cancer where you can see the cancer genome and what has gone wrong with the programming code for that patient's specific tumor," Borisy explained. This knowledge, he said, could help doctors optimize treatment.
The company's mission is to create the tools for this molecular analysis in a clinical environment. "We're taking what has been so far done in cancer genome centers in a research-grade manner, and we are making this something that can happen in real-world day-in-and-day-out clinical samples in a robust manner," he said.
Last April, the Cambridge, Mass.-based company completed a $25 million Series A funding round that was led by Third Rock Ventures, where Borisy is also a partner. This month, the firm, which has grown to about 30 full-time employees, is moving from temporary space into larger digs at One Kendall Square, the former labs of the Whitehead-MIT Center for Genome Research.
The company boasts a roaster of well-known founding academic advisors from the Broad Institute, Dana Farber Cancer Institute, Harvard Medical School, and MIT: Eric Lander, Todd Golub, Levi Garraway, and Matthew Meyerson. In addition, Rick Wilson, director of the Genome Institute at Washington University, which has been at the forefront of cancer genome sequencing, is among its scientific advisors.
Targeted … For Now
Foundation Medicine's test sequences several hundred cancer genes in tumor samples at very high coverage — "everything that could be plausibly clinically actionable today," Borisy told CSN. It analyzes these genes for many types of mutations, including point mutations, insertions and deletions, inversions, translocations, and copy number variants.
The test only requires small amounts of material, such as a few slices from formalin-fixed paraffin-embedded samples, core needle biopsies, or fine needle aspirates. The company has developed "extremely robust protocols for next-generation sequencing of FFPE samples," according to Phil Stephens, Foundation Medicine's executive director of research and product development, who spoke at the American Association for Cancer Research annual meeting earlier this month. He mentioned that the firm has "put a tremendous amount of effort, time, and resources into solving this problem."
Also, samples do not need to be pure tumor. "Unlike research studies, we are not asking for 80 percent tumor," Borisy said. "It could be 25 percent tumor. It's the real-world samples that are already being collected." To make up for low tumor content, the targets are sequenced at several hundred-fold coverage.
The company works with unamplified DNA and enriches the target genes by an undisclosed hybrid selection method prior to sequencing. Foundation Medicine has several Illumina HiSeq 2000 and Ion Torrent sequencers in its laboratory, for which it plans to obtain CLIA certification before the end of the year. It currently uses the HiSeqs for most of its work, "but technology platforms are continuously moving at a very rapid pace, so that can always change," Borisy said, adding that the firm continues to evaluate new technologies.
[ pagebreak ]
The test, which has a turnaround time of two weeks, is "extraordinarily" sensitive and specific, according to Borisy, including for "very rare alleles, even in a heterogeneous background."
Sensitivity and specificity are also the greatest current challenge, according to Stephens, who said that "you can always do more to identify different classes of mutations."
Stephens said that in order to reach a high level of sensitivity, the firm initially casts its net wide and then hones in on the true mutations. "We try to identify every single variant within that sample, and then we use intelligent post-processing and formatting to reduce the number to the true number of mutations in that sample."
He noted that detecting small insertions and deletions from short-read data is difficult, but said the company can now detect these "with high confidence" down to a threshold of 10 percent, and maybe at even lower levels in the future.
While it is "not that difficult" to detect copy number changes in the target genes, he said, it is trickier to find homozygous deletions and to discriminate them from hemizygous deletions in the presence of normal tissue.
He also said that not all rearrangements and gene fusions can be found with the approach, "but you can find the important ones."
In addition to developing the laboratory test, the company is working on a "knowledge base" for physicians to help them interpret the results and put them in the context of the scientific and medical literature. This will also help them make treatment decisions on the basis of the results.
"We don't presume to know the answer of what to do — the physician does, because he knows so much more about the patient and the patient's history — but we're trying to present what can be a very complicated set of information in the most relevant context that enables the physician," Borisy said. This informatics part of the test is still "under active development" and will be ready at commercial launch, the timing of which the firm has not yet announced.
Foundation Medicine has developed intellectual property in the areas of DNA isolation, library construction, hybrid capture, sequencing, and computational biology but has not yet disclosed any details of these technologies, according to Borisy. Over the next couple of years, "you will see a wave of presentations and publications from us, where we take this and demonstrate not only its analytical validity but its power in clinical decision making and utility," he said.
The company also plans to expand the test over the next several years, from selected genes to the whole exome and eventually the whole genome. Stephens said it would take about a year and a half to move to exome sequencing. Later on, the firm plans to add gene expression, methylation, metabolomic, and proteomic analyses. "The aim is to create the comprehensive universal cancer genomic-based test," Borisy said.
Such a comprehensive test will be necessary, he said, because many cancer mutations only occur at a low frequency and individual tests won't catch them. Also, there is often not enough tumor material available to run more than a couple of separate tests.
According to Thomas Lynch, director of the Yale Cancer Center and a professor of internal medicine at Yale University, the company's assay has the "potential to revolutionize the way we care for patients with cancer. It is a first step toward making genomic sequencing routine for practicing doctors."
Lynch, who said he has no financial interest in Foundation Medicine, has not yet seen any data from the company, but he hopes that his center will become one of the first sites to test the assay in patients. Five years from now, he predicted, "every cancer patient will want to have their tumor sequenced."
[ pagebreak ]
Earlier this year, Foundation Medicine said it had started a pilot collaboration with Novartis — its first partnership with a pharmaceutical firm — to "develop, enhance and optimize" its test "for Novartis' needs."
At the time, the firm said that if the pilot phase was successful, the two partners "will evaluate opportunities to collaborate on the production and commercialization" of the test.
The initial project was to apply the test to a number of FFPE samples and compare the results to other tests that Novartis had already performed on these samples, Borisy said.
According to Borisy, Foundation Medicine has signed and will shortly announce a partnership with another pharma company, and "there will be multiple more coming."
In addition, the firm is embarking on collaborations with several academic medical centers that it plans to announce over the next six months or so in order to demonstrate the test's reliability, sensitivity, and specificity, as well as "how much more information we find" compared to other tests, he said.
These studies should also help the company obtain reimbursement for the commercial test. "It takes time and a lot of evidence to secure physician reimbursement," he said. "Integrating our product in that whole system is something we take very seriously."
While the company has not yet disclosed pricing for the test, he said the price will be "very compelling in terms of the value it is going to bring."
He said the company does not have any direct competitors at the moment, although there is "a lot of activity" from companies — for example Complete Genomics and Illumina — as well as academic cancer centers to develop sequencing-based cancer tests.
Complete Genomics, for example, said last month that it plans to launch a new "cancer product" this year and that it sees cancer pathology as an important market (CSN 3/16/2011). Also, Illumina has sequenced several cancers in its CLIA lab as part of its individual genome sequencing service and offers discounts for customers with severe medical needs.
But Borisy said that these services currently don't "work in a clinical real-world setting." For example, Complete's service currently requires micrograms of DNA and has a turnaround time of more than 60 days.
Borisy declined to mention how long Foundation Medicine's current funding is expected to last and when it might need to raise additional capital. "Third Rock Ventures has very deep pockets, and we are very committed to making Foundation Medicine into a success," he said.
Have topics you'd like to see covered in Clinical Sequencing News? Contact the editor at jkarow [at] genomeweb [.] com.