Features

Seeing the Multiple Dimensions of Cancer:

How Targeted Imaging Technologies Are Bringing New Clarity to Cancer Care

A surgeon, a radiologist, and an oncologist sit in a dimly lit room, banks of monitors in front of them. Their attention is focused on a collection of pictures: black-and-white, color, human outlines, brightly lit spots in some places, dark in others. At the press of a button, the radiologist sends a command to a group of computers. Data are exchanged, and the images merge together effortlessly into a single picture of a human form, superimposing physiology on anatomy. The bright spots fuse, revealing the location, viability, and vulnerabilities of a tumor.

Left to right: Sankaran Subramanian, Ph.D., Staff Scientist; Mr. Frank Harrington, NIH Machinist; Murali Krishna, Ph.D.; and Jim Mitchell, Ph.D., show their original self-built magnet and field gradient assembly, which they used for electron paramagnetic resonance imaging (EPRI). This magnet was used to first demonstrate the feasibility of in vivo oxygen imaging.

This seamless scenario does not yet represent standard clinical practice. But it represents the ideal treatment planning or drug assessment scenario, one in which clinicians from different fields of oncology are able to share and integrate the data generated by a host of molecularly targeted imaging technologies—such as targeted optical fluorescent tagging, magnetic resonance imaging (MRI), and an emerging technology, electron paramagnetic resonance imaging (EPRI)—into single, holistic images that provide researchers and clinicians with a complete representation of the patient's tumor, including its location, its size, and its physiology.

Together, these technologies are fueling a new understanding of how tumor physiology and structure affect drug action while also bringing new precision to clinical treatment planning. The physician-scientists of CCR's Molecular Imaging Program (MIP) and Radiation Biology Branch (RBB) are leading the charge to refine these technologies and translate them into clinical practice, making the above scenario a reality.

The New Way: Seeing Is Believing

The traditional way of drug development, while effective and straightforward, is time-consuming and cumbersome. Researchers give the trial cohort a drug or treatment of interest, follow them for f e a t u r e months or years by MRI or computed tomography (CT) scanning, and look for changes in tumor size.

"Each technology has its strengths and weaknesses, and if we think broadly about how to leverage those strengths to answer specific problems, we can diagnose, track, and by extension treat cancers with greater specificity than is currently possible."

Traditional methods of treatment planning, particularly for radiation therapy or surgery, have similar limitations. Radiologists image the tumor using the same or similar techniques, with the goal of creating detailed three-dimensional representations of tumor size and location. At the same time, functional imaging—technologies like positron emission technology (PET)—have rapidly advanced the ability of doctors and scientists to see the activity within a tumor, as represented in the case of PET by the relatively insatiable appetite of cancer cells for glucose.

But the current imaging modalities have limitations. PET can tell radiologists how much glucose a tumor is using but cannot shed light on other aspects of tumor physiology or anatomy. MRI and CT can help provide unsurpassed anatomical detail but have difficulty defining metabolic dimensions.

CCR's MIP is stepping in to bridge the functional and the structural. "The MIP was established four years ago to try to find new points of view and new solutions to challenges in cancer imaging," said MIP Head and Senior Clinician Peter Choyke, M.D. "Each technology has its strengths and weaknesses, and if we think broadly about how to leverage those strengths to answer specific problems, we can diagnose, track, and by extension treat cancers with greater specificity than is currently possible."

Imaging has always been a component of the translational research conducted at CCR, but resources dedicated to nonclinical work were often limited. The MIP is changing that, but it is doing so in a way that complements the long-standing efforts of the RBB. "We now have a strong, integrated, cancer-focused, in vivo imaging but critical range of expertise," said Choyke. "With this disciplinary breadth, we can investigate the whole spectrum of imaging technologies and probes to create new families of clinically relevant image based biomarkers."

The availability of unique resources like the MIP's new dedicated clinical drug development imaging facility allows the program to serve as a focal point for research that is both high-risk and high-reward, like exploratory studies of new therapeutic agents and technology development (see "To Systematically Look Within").

Bringing Micrometastases into the Light

While radiology-based treatment planning methods provide anatomic information of unprecedented detail, once in the operating room, the most effective surgeries are those in which the surgeon can remove as much tumor as possible, including any metastatic colonies that may be present near the original malignancy. Currently, surgeons remove a margin, a buffer zone of apparently healthy tissue around the tumor, in the hopes of removing any micrometastases that may have spread, unseen, from the original cancer.

Hisataka Kobayashi, M.D., Ph.D., a Staff Scientist in the MIP, understands the challenges and importance of eliminating micrometastases early and efficiently, particularly for ovarian cancer patients. "Ovarian cancer is not a very aggressive cancer, but it is dangerous because it spreads silently. For this reason, gynecologic surgeons try to pick up as many metastatic nodules as they can." Also, because surgery, even when done endoscopically, is invasive, surgeons want to do as much as they can during a single operation.

But how can a surgeon know where the micrometastases are? Visual inspections by endoscope cannot reliably detect tiny tumors without some kind of guide or aid that makes the tumor stand out from the surrounding tissues. To address these visual limitations, Kobayashi and his colleagues—including MIP Visiting Postdoctoral Fellows Mikako Ogawa, Ph.D., and Nobuyuki Kosaka, M.D., Ph.D., as well as former MIP Clinical Fellow Yukihiro Hama, M.D., Ph.D., now an Assistant Professor at Japan's National Defense Medical College—have developed a system that literally makes ovarian micrometastases light up.

The system makes use of a natural response to antibody or receptor-ligand binding, namely that once an antibody is bound to a cell, it will be taken up by the cell and then sent to the lysosome, a cellular compartment or organelle that uses low pH to digest internalized proteins. In Kobayashi's system, exposure to the acidity of the lysosome triggers the fluorescent tag attached to the antibody, making the cell glow (Figure 1). "Because we use only a cancer-specific antibody, we only highlight cancer cells, not normal cells," Kobayashi said. The system also takes advantage of a second aspect of cellular physiology. Only viable, healthy cells are able to maintain a low lysosomal pH; if a cell is damaged, its lysosomes become alkaline. Thus, if a cancer cell that has internalized Kobayashi's tagged antibodies is damaged—by chemotherapeutics, for instance—the lysosomal pH rises, and the tag's signal fades.

"If we give this tagged antibody to an ovarian cancer patient before surgery," according to Kobayashi, "the surgeon can look for glowing areas and know that they represent micrometastases. At that point, the surgeon can remove them or paint them with a chemotherapeutic agent and observe, in real time, whether the drug has any effect."

Though the system is only in the preclinical stage, it already shows promise. In the December 2008 issue of Nature Medicine, Kobayashi and his team reported on the system's specificity at highlighting lung metastases as peritoneal metastases of ovarian cancer in mouse models.

Kobayashi believes the fluorescent system could have widespread applications. "Endoscopic surgery lends itself well to image guidance, which is effectively what we are developing with this technology. It can be applied to any cancer for which there is an appropriate antibody or ligand. We could adapt this method as a way for surgeons to better determine the edges of a tumor while conducting resections. It could be used as a way of guiding robotic surgery, an area NCI is interested in pursuing. We could even use multiple fluorescent tags responsive to different aspects of physiology to increase the scope of visual information we can gain in real time."

Figure 1: By using a cancer-specific ligand, like an antibody, conjugated to a fluorescent probe that glows only at low pH (green), researchers can see metastatic ovarian cancer cells (right, in a mouse model) and determine whether the cells respond to therapy.

Revealing Vascularity

Before the surgery can even take place, though, a surgeon needs to gather as much information as possible about the tumor's shape, location, and activity. Similarly, while deciding whether to employ chemotherapy, targeted therapy, or other treatment strategies, a medical oncologist should have as much information on tumor structure and physiology as possible.

To add physiological sensitivity to the anatomic detail provided by MRI, Choyke and his colleagues have turned to an imaging technique called dynamic contrast enhanced-MRI (DCE-MRI). "DCEMRI falls somewhere between molecular imaging and anatomic imaging," said Choyke. The true difference between the two is reflected in time. Standard MRI takes a snapshot of a tumor's anatomy and location. By comparison, DCE-MRI is, as the name implies, dynamic, producing a representation of a tumor's blood flow over time.

At the heart of DCE-MRI is a running series of MRI snapshots taken at very short intervals using a contrast agent called gadolinium. This rare earth element interacts with the protons in water molecules, making them stand out more clearly on an MRI scan than they normally would. "Gadolinium actually changes the properties of the water in the body," Choyke explained. "When we run a DCEMRI scan, the movie we produce actually captures the effects of gadolinium on the surrounding water, giving us a dynamic view of the agent's uptake into and clearance from the tumor."