Chris Anderson

Dx companies are investing heavily in computational and informatics platforms in order for genetic sequencing to be of value to clinicians.

The cancer diagnostic landscape has evolved significantly since the first companion diagnostic was launched to detect overexpression of the HER2 gene in breast cancer and thus indicate whether a patient would be highly likely to respond to trastuzumab (Herceptin) treatment. Since then, and over the course of the last decade, companion diagnostics that focused on a single distinctive molecular marker for a specific cancer have been used to both inform the development of relevant therapeutic targets and determine whether individual patients would respond favorably to a therapeutic. The development of these diagnostics was the first step to unraveling the sticky question of why some patients showed great improvement when prescribed a cancer medication while others showed no response at all.

In the mid to late 2000s, cancer diagnostics moved to gene expression profiling. These tests don’t determine the likelihood of response to a particular drug; rather they stratify patients based on the likelihood their cancer will recur after initial treatment. The most successful diagnostics, Oncotype DX from Genomic Health and MammaPrint from Agendia, target breast cancer and today are used to help determine how aggressively to treat the cancer based on each patient’s risk.

“The tests are more prognostic than diagnostic, and the idea is you would refrain from treating those patients who are low risk,” said Roger D. Klein, M.D., J.D., professional relations chair for the Association for Molecular Pathology. “The problem they are trying to solve is overtreatment of patients who have low risk. I think that oncologists have been very concerned about giving chemotherapy to patients if they don’t need it and they haven’t had effective ways of sorting these patients out.”

Next-Generation Genetic Sequencing

In addition to information for how aggressively to treat specific cancers, oncologists are increasingly looking to provide targeted therapies based on specific identifiable molecular aberrations. Breast cancer alone has at least five subtypes and determining which subtype a patient has is vital to providing the most effective treatments. Recent diagnostic developments focus on discovering and detecting the specific genetic mutations in cancer subtypes, and for that the industry has largely turned to gene sequencing.

Using biopsy tissue, cancer diagnostic companies perform sequencing analysis on panels of 10, 20, or 40 or more genes at a time targeting the genes which, when altered, are most associated with that form or cancer. So for non-small-cell lung cancer, screening panels would include oncogenes associated with that cancer such as EGFR, KLAS, and others. Panels for other cancers vary based on those oncogenes that are most clinically relevant based on the research literature and the universe of existing drugs targeting those mutations.

“The idea of a single marker linked to a single drug in an era where an individual may need to be tested for two, three, five, 10, or 20 markers really doesn’t fit for delivering targeted therapy, given the explosion of genomic information available,” said Vincent Miller, chief medical officer with Foundation Medicine. Even better, as the price of genetic sequencing has come down, diagnostics companies can perform the sequencing on panels with dozens of genes for the same amount it cost to sequence one or two genes just a few years ago.

But in order for genetic sequencing to be of value to clinicians, diagnostics companies have also invested heavily in their computational and informatics platforms.

“The molecular technologies are data-generation technologies. Some people have described medicine as more of an art than a science. I think what you are seeing with the evolution of these technologies is they are driving medicine away from an art form and making it much more an information science,” said David Jackson, CSO, with MolecularHealth. That’s because companies like MolecularHealth and Foundation Medicine use sequencing data derived from the tumor cells and crunch it along with data from the research literature and other public sources to show not only which genes have aberrations, but also which of those genes are clinically relevant and boil it all down into a single report that doctors can use to help them make treatment decisions.

“In the absence of mutational information, targeted therapies tend to not be very useful,” Jackson added. “That broader view on the tumor is essential to increasing response rates, and it is that challenge that is being addressed today using more complex knowledge mining and data analysis methods.”

Less Invasive Sample Collection

While the advances made in cancer diagnostics employing genetic sequencing have been significant in the past few years, commercial applications still rely on tumor tissue. In many cases performing the biopsy to collect the tissue is a highly invasive procedure, and depending on the form of cancer or progression of the disease, may be contra-indicated. For this reason, researchers are now turning their attention to developing methods to capture and sequence DNA from blood, urine, and cerebrospinal fluid samples. Prime sources of genetic material in these fluids include circulating tumor cells, exosomes, and cell-free DNA, among others.

Being able to easily collect and analyze these kinds of samples could have significant therapeutic value. For instance, lung cancer patients typically don’t get a second biopsy should their cancer recur. Using urine or blood samples and sequencing DNA derived from these sources is potentially valuable to clinicians who could see the genetic signature of the new cancer and whether it had new or different mutations from when it was first diagnosed.

“It really is the Holy Grail to be able to do genetic profiling in the blood,” said Johan Skog, CSO, Exosome Diagnostics. “One of the big advantages is you can do this repetitively, so we can also do longitudinal studies which can help show treatment efficacy and better identify the different signature of the responders.”

Foundation Health’s Miller said these potential sources of genetic material are very encouraging, and the company is currently assessing which are the most promising. That said, he envisions each source having a role in the future. “There are exosomes, there is cell-free DNA, there are circulating tumor cells,” he said. “They are not mutually exclusive in terms of their value, and we are looking at all of those to potentially be an overlay for our Foundation One tests.”

Regardless of the source of genetic material for gene sequencing diagnostics, improving the treatment of cancer should only get more precise in the coming years. As Miller noted, if you charted a graph showing the number of clinically relevant biomarkers discovered over the past ten years, the curve would be very steep. “When one sees the value of a predictive biomarker that is genomically faced and the magnitude difference in outcomes that is possible when linking that biomarker to a therapy, it changes the fundamental equation around the value of diagnostics and treatment,” he concluded.

Chris Anderson ([email protected]) is the former chief editor of Drug Discovery News, which he helped launch in 2005. 

This article was originally published in the May 29 issue of Clinical OMICs. For more content like this and details on how to get a free subscription to this new digital publication, go to

Previous articleSmall Peptide Breaks Up Biofilms
Next articlePatient Group Issues Roadmap for Accelerating Cancer R&D