Organs of the human body shift in size, shape, and position from week to week as oncology patients react to radiation therapy or as the progression of tumors induces loss of weight. Cancer fighters have known this for some time. Now, however, thanks in part to research conducted at the Comprehensive Cancer Center at the University of California, San Francisco (UCSF), it is understood that organ movement does not always occur in predictable ways. This is a discovery of significant concern for the field of radiation oncology.
Jean Pouliot, PhD, professor in the UCSF department of radiation oncology, says, "Traditionally, radiation treatments have been delivered based on a dose that was planned 2 or 3 weeks earlier; relying on a number of quality-assurance procedures would, it was hoped, provide a dose that, at the time of actual delivery, would be the correct one." Pouliot is a Comprehensive Cancer Center imaging researcher who, for more than a decade, has been working to unlock the secrets that will allow radiation therapy to be administered with greater precision and efficacy. "Because of our new insights into organ movement, we are realizing, to an overwhelming extent, that you cannot obtain a CT study of the patient 2 weeks before treatment and assume that the anatomy that you acquired then will remain unchanged (so that you can position the patient exactly the same way, time after time)," he says. "Moreover, as we now see it, the real endpoint is not only having the patient correctly positioned, it is also ensuring the dose is being primarily delivered to the tumor. That can only be accomplished if you know the positioning information immediately before you treat."
Pouliot and his colleagues at the Comprehensive Cancer Center have devised a collection of innovative tools and techniques to produce that information in just such a timely manner. "The dream of having all the information at the same place, at the same time, for the same patient is now being realized," he says. "We are now able to verify the delivered dose. We are able to base clinical knowledge on the real dose delivered, not on what was planned weeks earlier." These abilities come under the heading of a class of interventions known as adaptive radiation therapy (ART).
"Adaptive radiation therapy is really more of a concept than it is a technology," Pouliot explains, adding that the ideas upon which adaptive radiation therapy is based were first arrived at more than a decade ago, when the requisite technology was still in a nascent stage. Then, for the first few years after adaptive radiation therapy moved from the drawing board to the imaging suite, he adds, "It was technology-driven, but not any more, because it is a concept that can be approached in a number of ways, depending on the technology you favor or have access to for this purpose. Today, what we find is that adaptive radiation therapy is increasingly a dose-driven proposition."
Even so, an imperative continues to be the correct positioning of the patient. The difference now is that the options for establishing positioning have been expanded. "One of the choices is to use the Siemens PRIMATOM, a combination linear accelerator and CT scanner on rails," Pouliot says. "You simply roll the patient into the CT, take a three-dimensional (3D) image, and then roll the patient back to the treatment position and begin administering the dose."
The choice embraced by Pouliot is megavoltage (MV) cone-beam CT, which is currently under clinical research evaluation. Here, a Siemens linear accelerator generating 6 MV beams is paired with an amorphous-silicon electronic portal imaging device.
"We are using the linear accelerator itself as the imaging tool," Pouliot says. "The linear accelerator is rotated around the patient and a portal image is acquired with very low exposure at each angle around a 180 degree arc. This arrangement permits creation of the CT image directly from the beam and, using classic Feldkamp algorithms, allows for volumetric reconstructions with a filtered back-projection. We get an image and verify that the patient is properly positioned and that the anatomy is correct, and then we begin treatment."
MV conebeam CT image of a head phantom with implanted radiopaque markers.
MV conebeam CT image of a head phantom with implanted radiopaque markers.
Patients with cancers of the head and neck are initially expected to see the most benefit from this approach. "That is primarily because, with cancer in the neck area, there is a significant degree of freedom in terms of rotation and translation," Pouliot notes. His version of adaptive radiation therapy also should prove invaluable as a way of minimizing organ risk. "In the head and neck area there is a great deal of organ risk, given that you have structures there that include the brain stem and the spine," he says. "Obviously, we do not want to deliver radiation to those organs. We want to be able to control radiation delivery meticulously there, and that is precisely what MV cone-beam CT is allowing us to do."
Since cancers of the head and neck typically respond well to radiation therapy, it is thought that the anticipated superior accuracy made possible by MV cone-beam CT will encourage practitioners to use higher doses in that region than are currently recognized as advisable. "If we could deliver higher doses, we would improve survival rates," Pouliot says, "but, at the moment, we cannot increase the dose because of the associated side effects and complications. If we could target the tumor better and avoid the organ risk, there is a real chance to improve survival rates."
Evaluation of the proper patient alignment immediately before radiation treatment: The MV conebeam image CT is superimposed on the image from the regular kV planning CT. The positions of the targets (tumors) and organs at risk contours drawn on the planning CT, as well as the treatment beams seen from different views, provide all the required information for the verification and the correction of the proper patient alignment.
Pouliot expects that MV cone-beam CT will one day be applied, as well, to the treatment of other types of cancer (especially prostate cancer), and will be found effective against them. This expectation may be met soon, as UCSF is poised to begin a small-scale clinical trial of MV cone-beam CT. "The first phase of the study will involve a cohort of just 10 patients," Pouliot says. He will serve as lead investigator of the single-site endeavor. "Each patient will probably be imaged several times to give us, perhaps, 50 initial data sets of complete two-dimensional (2D) and 3D images. Later, we intend to determine the protocols that will allow us to have a clinical impact."
With this study, Pouliot wants to move toward answering several relevant questions. "Today, we can image the complete anatomy in 3D, whereas, before, it was just 2D imaging," he begins. "That gives us a lot more information, but the real questions are how we can best use that information and what we gain from that. How is human perception improved? Can we improve the speed with which an error is detected? After studying these issues in the first phase of the investigation, our attention will turn to another round of questions. Specifically, now that you know exactly where to deliver your dose, do you want to increase the dose to the tumor, or do you want to reduce the dose to the surrounding tissues?"
PRECISION IN TREATMENT
In time, MV cone-beam CT and other forms of adaptive radiation therapy may permit highly individualized treatment of cancer patients through customized dose delivery. "Adaptive radiation therapy is not the end point in all this," Pouliot says. "The first phase was to develop a number of tools to improve precision in delivery. Intensity-modulated radiation therapy (IMRT) was the key player in that. We can now control the dose to a significant extent."
"The second phase is adaptive radiation therapy: what we are doing now, using imaging modalities to guide the delivery of the dose, and also using other imaging modalities at the diagnostic level to define the target better. The third phase, still in the future, for the most part, is functional or molecular imaging. This will allow us to differentiate which part of a tumor is the most aggressive, which part of the organ really has cancer, and which is healthy tissue. IMRT and adaptive radiation therapy will be used to target the tumor better, based on functional and molecular imaging using MRI and PET/CT imaging."
Once these modalities are in routine use, adaptive radiation therapy will provide the framework for using the collected information, Pouliot adds. "We will not be talking about adaptive radiation therapy as a stand-alone treatment. Right now, we administer the same total dose for all patients within the same group of cancers. As we learn more about dose response from functional imaging, for instance, we will be able to stop treatment earlier for patients who respond well to radiation. Conversely, we will be able to extend treatment for certain individuals beyond the recommended limit, if necessary, to ensure that all cancer cells have been killed." Pouliot concludes, "It will all be possible thanks to having the fully up-to-the-moment information needed so that tumors can be targeted better and doses can be delivered far more accurately."