by Gina Shaw

Quick, what do you think of when you hear the word "biomarkers?" For most people, the first thought is probably of markers of disease (or therapeutic response or toxicity) found in enzymes, proteins, hormones, and other substances detectable in blood, urine, or cerebrospinal fluids. But one of the biggest growth areas in biomarker research does not involve drawing something from the body and studying it on a slide or chip: it is the study of biological activity in vivo.

Imaging biomarkers have one important advantage over other biomarkers: they're site-specific. In other words, they show you precisely where the disease is. And they offer one of the best opportunities to improve the drug development process in the short to near term, according to a report from Health Industry Insights, an IDC company.

"Building on accumulated biological knowledge and innovations in molecular imaging contrast reagents, the ability to directly and non-invasively visualize important disease interactions within the human body can provide critical data and information in hours to days, where previously months to years were required to infer drug effects (both desired and unintended)," writes Alan Louie, Ph.D., an HII research director.

Hundreds of imaging biomarkers have already been identified. Some are so long-established and standardized that they are virtually second nature. These include the measurement of tumor size using MRI, or the assessment of left ventricular ejection fraction (LVEF) after cardiotoxic chemotherapy using tools like Multi-Gated Access (MUGA) scans, echocardiography, CT or MRI. But there are dozens, if not hundreds, of other biomarkers in the pipeline.

"The challenge for the use of imaging biomarkers is to characterize them and validate them enough so that they can be used as surrogate endpoints in clinical trials. Things such as measuring the structure and density of bone, that's pretty well accepted. Measuring the size of tumors and the response to therapy is very well accepted. Measuring the number of plaques in patients with myelodegenerative disease, including MS, is moderately well established and accepted," says James Thrall, M.D., radiologist-in-chief at Massachusetts General Hospital and cofounder with Dr. Gregory A. Sorensen of MGH's Center for Biomarkers in Imaging.

"But there are literally dozens, if not hundreds, of potential biomarkers that are incompletely characterized and that have not been independently validated in clinical testing. If we can unlock the potential implicit in that rich array of imaging biomarkers, I think we could have an enormous impact on how clinical trials are conducted, beyond even what we're having today."

PET projects

because that's what we have to fix. Can you identify and separate patients by whether they have the protein targeted by the drug or not? Can you image where this drug goes, and determine if it hits the target in a patient to induce the desired pharmacological effect?"

The foundation of PET's success, of course, is glucose the building block of FDG-PET, now fully covered by Medicare for all cancer imaging and rapidly expanding into neurological indications such as Alzheimer's disease. "If you could pick one metabolic pathway only one to give you the greatest coverage of the normal functions of the body and identification of disease, you'd have to pick glycolysis," says Phelps. "Nothing is more important in a metabolic pathway throughout the body, nothing is a better marker of transformation to disease. It changes the way we treat in 30% to 60% of all cancer patients, it can detect Alzheimer's disease with 93% accuracy three years before clinical diagnosis of probable Alzheimer's, and can detect Huntington's seven to eight years before the onset of symptoms."

But FDG-PET is only the beginning. There are virtually no limits on the possibilities for new PET tracers, says Louie. "PET can look at a lot of things, anywhere in the body. It can look at blood flow, the binding of various antibody-antigen relationships, the metabolites associated with enzyme digestion… it's very targeted research giving you insights into specific areas within the body."

Some PET tracers now being validated and undergoing head-to-head comparisons with others in clinical study include F-18 fluoromisonidazole (FMISO), which detects hypoxia and predicts its effect on radio- and chemotherapy (surprisingly, it appears to make the tumor more resistant to treatment); F-18 labeled choline analogs, markers for increased lipid synthesis rates in cancer cells; and Pittsburgh Compound B (PIB), a hydroxylated benzothiozole that binds to the amyloid- protein.

One intriguing new prospect for PET: using imaging to infer the activity of a gene by imaging the consequences of gene expression. Thrall and his colleagues at Mass General, as well as other imaging researchers, are doing this by detecting cellular transfection with the KOS virus. "A cell that is successfully transfected with the KOS virus will express viral thymidine kinase (HSV1-TK). When an animal with transfected cells is exposed to F18-labeled FHGB, an acyclovir analog, the FHGB is taken up and phosphorylated through the action of viral thymidine kinase," Thrall says. "Once it's phosphorylated, it's inhibited from leaving the cell. So a tumor exposed to the KOS virus which is in itself a therapy for tumors and then later imaged with [F18]FHGB, will demonstrate positive uptake of the PET tracer, allowing us to infer the successful application of the KOS virus to the tumor."

Fluorescence in action

Another growth area for imaging biomarkers potentially rivaling PET, says Thrall is optical imaging, which appears to be particularly valuable in targeting enzyme expression. "Optical imaging is also turning out to be particularly valuable in targeting enzyme expression. Optical imaging agents and nuclear medicine pharmaceuticals including PET pharmaceuticals are actually architected in exactly the same way," he says. "Every one of these targeted imaging agents has two moieties: a molecular component that confers targeting or localization, and a second component that confers external detectability."

For radiopharmaceuticals, that's a radionuclide component whose decay and proton emission localizes the agent's biodistribution. Optical agents, on the other hand, are imaged by exposing tissues to a light source that stimulates fluorescence from the fluorophore. "Just as you use a gamma camera or PET tomography to image in nuclear medicine, you use an optical imaging device which consists of a light source to stimulate the fluorescence, typically in the near-infrared spectrum, and then a set of optics, usually including a CCD camera, to record the emission."

Image courtesy Anna-Liisa Brownell, Ph.D., Director of the Experimental PET Laboratory at Massachusetts General Hospital. Click here to enlarge.

One of the advantages of optical agents, says Thrall, is that they are extremely flexible to produce. "There are some extremely clever ways of producing them, such as the so-called 'smart' optical contrast agents." These agents can't be seen in their natural state, but became highly fluorescent when they interacted with a target. Sub-millimeter-sized tumors have been visualized using these agents. "The advantage is that there's no background signal the only signal detected is from the target incident."

There are ways to create smart agents for MR and ultrasound as well, but it's more complicated. "MR chemistry is not as flexible and new agents are not emerging as fast," Thrall says. "The great flexibility of PET, of course, is that the labels are intrinsic to organic molecules carbon, oxygen, nitrogen, and then fluorine, which is not intrinsic to many organic molecules but is often exchangeable with an atom from an organic molecule."

Within the next three to five years, Louie predicts, the increasing use of imaging biomarkers in clinical development and drug submissions will lead to established standards, processes, and reviews and subsequently, to routine use of imaging biomarkers in future drug development. "With a continually improving and innovating technology foundation, imaging biomarkers can be expected to expand into new applications as opportunities arise, with eventual expansion into the clinic."

But to take full advantage of the potential of imaging biomarkers in high-throughput, cutting-edge research applications, more high-performance computing capabilities are essential, Louie says. "High-resolution research tomographs, for example, are generating huge quantities of information, in part because the number of slices they're doing per scan is very fine. They're at the sub-millimeter slice level now. Low degrees of granularity are fine at the individual patient level, but if you're trying to look at very precise sub-segments of a particular organ like the brain, at times you need very high-resolution capability."