In the field of cancer, for example, molecular imaging is playing an increasing role, in particular for drug discovery and assessment of therapeutic response. It thus appears intuitive that molecular imaging will become established as a biomarker by which to characterize disease status and demonstrate the effectiveness of a given therapeutic intervention. Indeed, trials using a variety of molecular imaging techniques to validate this function within cancer are underway.

Methods of particular relevance within clinical practice include MR imaging and MR spectroscopy, as well as nuclear medicine techniques such as PET and SPECT. Their utility and promise are evident in the many clinical trials that have explored sophisticated new molecular imaging agents and technologies. Not one of these new agents, however, has been officially approved or reimbursed for monitoring therapy.

Optical imaging - fluorescence or luminescence - is another technique that's widely used in preclinical studies to evaluate the effects of intervention. Its application to date has been largely in animal and in vitro models, although clinical use is increasing, especially in gynaecologic malignancies and situations where endoscopic evaluation is possible. While the technique's sensitivity and resolution are exquisite, the current signal degradation at a small distance from the detector precludes accurate quantification and adequate resolution in animal/clinical models.

MR matters
MR imaging and spectroscopy are increasingly employed to characterize the various physical qualities of cancer. For this purpose, the most widely used technique is dynamic contrast-enhanced (DCE) MRI, a computer-enhanced modality that seeks to measure tissue (tumour) vascularity - information of particular relevance for the evaluation of anti-angiogenic agents in cancer. The technique has been evaluated using surface coils to examine disease close to the surface of the body.

Current MR techniques boast a high resolution (sub-millimetre). However, their broad utility is limited by factors that include low sensitivity (usually millimolar), the inability to reproducibly obtain dynamic parameters deep inside the body, and the lack of standardized protocols that can be used in multicentre studies. While there has been considerable progress made in these techniques, these factors (other than the inherent lack of sensitivity of proton imaging/spectroscopy) still need to be addressed.

It should be noted that most such MR techniques utilize the imaging of an agent that remains inside the vasculature, and thus largely measure vascular perturbation. Purists may, therefore, be forgiven for believing this to be a non-molecular, albeit functional, imaging modality.

MR spectroscopy, on the other hand, does involve the direct assessment of molecular profiles within tissue. The most widely used such parameter is probably choline level, with several studies demonstrating a link between elevated choline content and tumour aggressiveness. Again, MR's lack of sensitivity results in the limited utility of this technique. The necessity for specialized instrumentation such as high-field-strength magnets has additionally hampered the clinical study of other molecules of interest - particularly sodium (Na-23) and phosphorus (P-31), with studies thus being limited to preclinical in vitro and animal models.

Track and trace
The use of radioactive tracers in molecular imaging is widespread, both in preclinical and clinical use. This is primarily because tracers enable researchers to study nanomolar quantities of agents that target molecules of relevance in cancer pathophysiology. Tracers can also track other functional features of cancer, such as vascularity and oxygenation. Indeed, perhaps the earliest example of molecular imaging was an image of the distribution of iodide within a thyroid gland.

The evolution of PET, now commonly combined with CT for anatomical imaging, has led to an explosion in the study of radiotracers - usually labelled with short-lived positron emitters such as carbon-11 or fluorine-18 - for evaluating metabolic characteristics of cancer. The study of glucose metabolism using 18F-fluorodeoxyglucose (FDG), for example, led to an increased understanding of carbohydrate metabolism and served as the paradigm for the development of PET imaging of cancer metabolism.

FDG-PET/CT has now become an integral tool for evaluating the extent of disease in a large number of haematologic and solid neoplasms, as well as for assessing response to therapy. An increasing number of other radiotracers, such as radiolabelled choline and radiolabelled steroid hormones, are also being studied for their utility in evaluating functional characteristics of the cancer phenotype.

Imaging with FDG-PET/CT is also becoming increasingly employed to define tumour volumes for radiation therapy planning. While this has been of considerable value in the delineation of viable tumour, its utility in treatment follow-up is less than clear. Moreover, there is no consensus on methodology for definition of viable tumour volumes to be selected for therapy.

Finally, the utility of FDG-PET/CT to define tumour volumes during therapy is also not well understood. Several groups have demonstrated that such PET-delineated volumes can be variable and may not always represent viable tumour: increased glucose uptake is a component of the Warburg effect (the enhanced conversion of glucose to lactate by tumour cells) and is also a feature of increased glucose metabolism by other tissues, for example, peritumoural inflammatory cells.

More specific methods for evaluation of cancer phenotype are now being actively investigated. Parameters such as amino-acid or nucleoside turnover in cancer cells may be more specific to viable tumour, and radiolabelled tyrosine and thymidine radiotracers are among the agents being studied. In addition, other features of cancer, particularly hypoxia, are being evaluated.

It is becoming increasingly clear that molecular imaging will bring a new perspective to our understanding of cancer biology and its relevance in the planning of radiation treatments for cancer. FDG-PET/CT is becoming an integral part of the initial work-up in patients scheduled to undergo radiation therapy, while other radiotracers - particularly those that evaluate hypoxia and other key features of the tumour micro-environment - are being increasingly studied.

Most imaging modalities are now being evaluated for use both at the beginning and the end of therapy. Molecular imaging for image-guided radiation therapy is starting to be employed for optimizing dose distribution and delivery, and methods for adaptive radiation therapy are also being utilized. All of these applications mandate a prospective approach with clearly defined end points that will eventually assure the proper utilization of these exciting new ways to image the characteristics of cancer.