Scientists and regulators collaborate to bring a new generation of powerful early diagnosis tools to the clinic.
Around 1600 BC, an anonymous Egyptian surgeon reported his clinical experiences with cancer. Eight of his patients presented with obvious lumps and lesions on their
breasts. Though the condition was incurable, the surgeon noted that cauterization with a “fire drill” provided temporary relief—or perhaps just discouraged the patients from complaining further.
While cancer treatments have evolved dramatically over the millennia, the quest for better diagnostic markers has been somewhat slower.
“Using tests based on histopathology or other devices . . . most of the tumors have been detected when they’re incurable,” says Sudhir Srivastava, PhD, director of the Cancer Biomarkers Research Group at the National Institute of Health’s (NIH) National Cancer Institute (NCI), Bethesda, Md.
Armed with the latest generation of genomic and proteomic tools, though, researchers like Srivastava are now scaling up the search for better biomarkers.
In addition to the availability of new technologies, the recent surge of interest in biomarkers is being driven by the increasing pressure to shorten drug development times. “In traditional trials, we use mortality as the endpoint, and that can take many years, but if you have a marker or a panel of biomarkers that is predictive of that mortality, then you don’t have to wait that long,” says Srivastava.
To help find those new biomarkers, researchers at the NCI and elsewhere have begun collaborating in larger groups. Besides its intramural biomarker work, the institute is now a major partner in efforts such as the Cancer Genome Atlas and the Early Detection Research Network. The Cancer Genome Atlas (TCGA) is applying large-scale genome sequencing technology to identify novel genes involved in cancer pathogenesis. Though it is still in a pilot phase, many experts are optimistic that the Atlas will eventually produce a plethora of genetic markers that can be used not only to diagnose cancer earlier, but to track its development more accurately.
The Early Detection Research Network (EDRN), as its name suggests, is also focused on early diagnosis, but its goals are much more varied than the Atlas’s. Srivastava, who runs the Network, adds that the EDRN, which coordinates biomarker research across the institutes at the NIH, is also working on everything from breast cancer biomarker validation to improved models of prostate cancer prediction.
Outside the NIH, biomarker researchers at far-flung academic research centers have also begun to coalesce into consortia. For example, the International Cancer Biomarkers Consortium (ICBC) helps researchers around the world coordinate their efforts and share data on a voluntary basis. “The ICBC is really somewhat of a loose coalition of the willing, who kind of would like to get together and exchange ideas about marker discovery,” explains Samir Hanash, PhD, program director for molecular diagnostics at the Fred Hutchinson Cancer Research Center, Seattle, Wash., where the ICBC is based.
Though ICBC membership does not help scientists get any new funding, it has drawn plenty of interest. Hanash explains that biomarker researchers in many countries have found it difficult to track the emerging field’s rapid progress. While scholarly societies such as the American Association for Cancer Research hold regular meetings, Hanash explains that “those types of meetings tend to be humongous, [with] 15,000 people that gather together once a year, so it’s hard to have effective exchange on a very focused topic.”
A match made in Maryland
Corporate biomarker scientists face fewer funding problems than academics, but often have trouble arranging collaborations.
“In the pharmaceutical and biotechnology industries . . . traditionally the mantra is avoid the competition at all costs, don’t talk to the enemy,” says Anthony Altar, PhD, director of the recently formed Biomarkers Consortium at the Foundation for the NIH, Bethesda, Md.
The new Consortium, a public-private effort, hopes to knock down the barriers to corporate collaboration on biomarkers, while simultaneously tapping new funding sources for academics.
After laying the groundwork for nearly a year, the Pharmaceutical Research and Manufacturers’ Association (PhRMA), the Foundation for the NIH, the Food and Drug Administration (FDA), several advocacy groups, and representatives from nearly every major pharmaceutical company officially launched the Consortium last October (see sidebar).
“We are able to combine multiple groups that hadn’t worked together before to produce an even more powerful project proposal, and then we fund those primarily through funds that are obtained from commercial entities,” says Altar. Indeed, the Consortium has already put its money where its mouth is, raising $7 million to fund one project looking for early cancer biomarkers on fluorodeoxyglucose-positron emission tomography (FDG-PET) scans. Contributors and collaborators on that project include pharmaceutical heavy-hitters such as Amgen, AstraZeneca, and Pfizer.
Consortium leaders say several other projects of a similar scale are also in the pipeline. “It’s part of a strategy to help make industry more effective or more productive in its research and development process, by identifying things that really don’t need to be done kind of redundantly and competitively across every single company,” explains Darrick Fu, associate vice president of scientific and regulatory affairs for PhRMA.
Though the Consortium is starting with cancer biomarkers, its overall focus is actually broader, encompassing biomarkers for any indication. “It’s kind of a big tent partnering venue for a variety of types of biomarker research and development projects, where really anyone can find sort of a space to collaborate,” says Fu.
The structure of the projects is also open-ended. “Some projects may be just among half a dozen or a dozen pharmaceutical companies. . .but other projects may be a dozen different academic groups with NIH and a very small industry component, so it’s really open to any kind of mixture,” says Fu.
In addition to saving money and pre-empting a biomarker land rush, companies joining the Biomarkers Consortium also hope to push promising new findings from the lab into the clinic. “It’s basic biomarker research, but it’s also trying to address another need, which is the translation of that research into things that are useful for drug developers, and generally we call that the qualification of biomarkers,” says Fu.
Others prefer to call it biomarker validation, but in either case, the question is the same: how does one decide when a change in some biomolecule constitutes a clinically meaningful test? So far, it’s been tough to answer.
If a biomarker seems to provide a reliable prediction of tumor growth in animal models and a small sample of patients, for example, researchers next need to scale the studies up. “The first step I would like to do is a big study where we have many control subjects without the disease and subjects with the disease, and see if the biomarkers can distinguish cases from controls,” says the NCI’s Srivastava.
Even if a biomarker distinguishes diseased from healthy patients in the initial trial, it needs to be tested across several laboratories. “Are the assays verifiable? Are they reproducible in several labs? Do they measure the analyte of interest with greater accuracy [than current tests]?” asks Srivastava.
Compounding the validation problem, some researchers in genomics and proteomics laboratories find potentially useful biomarkers by the truckload. “If you do a one-time experiment comparing ‘A’ with ‘B,’ finding hundreds of differences, and say ‘well, based on that I have a long list to validate,’ well, you have a problem,” says Hanash. As scientists have realized the futility of chasing down so many leads, the field has moved toward more selective experimental designs.
As an example, Hanash explains that a study comparing advanced-stage cancer patients to completely healthy controls will inevitably uncover more red herrings than useful leads. “When you do that, you’re going to find a lot of differences that may have nothing to do with cancer—you’re dealing with patients who are very sick, who have a lot of other types of problems,” says Hanash. Instead, he suggests comparing more similar groups, such as early-stage lung cancer patients and longtime smokers with inflamed lungs but no overt tumors.
Finding such ideal controls, however, can be extremely difficult, and making sure they’re all sampled the same way can be even harder.
Often, cancer patients’ blood samples are collected in a clinic after the patients have fasted overnight, while control samples may have been collected from well-fed volunteers in the middle of the day. And even relatively ill control subjects are unlikely to have experienced the weight loss, emotional stress, and chemotherapy drug intake of their cancerous peers—all factors that can fill sensitive genomic and proteomic screens with misleading hits.
Researchers can’t eliminate all of the false leads in their initial screens, of course, but they are making progress. “For example, we have joined with the women’s health initiative funded by NIH, and they have a number of samples, what we call preclinical samples, that could be very useful to validate our biomarkers,” says Srivastava. By comparing samples taken from patients before and after they received a cancer diagnosis, the researchers can use the same patient as case and control, eliminating many confounding factors.
About the Author
Originally trained as a microbiologist, Alan Dove has been writing about science and its interfaces with industry and government for more than a decade.
This article was published in Drug Discovery & Development magazine: Vol. 10, No. 8, August, 2007, pp. 16-20.
Filed Under: Drug Discovery