Last Modified: August 21, 2002
Authors: Jason Lee MD,1 Indra J. Das PhD,1 Shiao Y. Woo MD,2 Walter Grant PhD,2 Bin S. Teh MD,2 J. Kam Chiu MD,2 E. Brian Butler MD2
Affiliations: Abramson Cancer Center of the University of Pennsylvania1 Department of Radiation Oncology, The Methodist Hospital, Baylor College of Medicine2
Intensity-modulated radiation therapy (IMRT) is a rapidly emerging technique for delivering conformal treatment plans. IMRT relies on inverse treatment planning based on clinical target and avoidance parameters, rather than a forward planning process of various beam, weight, and energy configurations. Computer optimization can generate treatment plans consisting of spatially nonuniform radiation beams that produce a relatively uniform dose distribution within a target volume. The implementation of IMRT has been made possible by high-speed computer equipment and the introduction of multileaf collimators on most linear accelerators. IMRT has a wide range of potential clinical applications, particularly in central nervous system, head and neck, breast, and prostate cancer patients. A number of clinical studies have demonstrated superior dose distributions with IMRT compared to conventional radiation therapy, reduced acute radiation toxicity as a result of normal tissue avoidance, and ability to perform dose-escalation. In addition to the clinical uses of IMRT, the process of IMRT treatment planning and issues related to implementation of IMRT are also reviewed in this two-part article.
Radiation therapy (RT) plays an important role in the curative and palliative management of cancer patients. Nearly two-thirds of patients will receive radiation therapy during the course of their cancer. Advances in radiation oncology are taking place at both at the physical and biologic level. With the advent of superior imaging technology and faster computing equipment, there has been considerable progress in the ability to physically deliver radiation dose to regions containing tumor and limit dose to normal tissues. One of the recent advances in radiation oncology is the introduction of "intensity-modulated radiation therapy," or IMRT. This article is divided into two parts: the first part will discuss the concept and planning of IMRT, and the second part with discuss the delivery of IMRT, its uses in various clinical settings, and a comparision of IMRT to conventionally delivered RT.
After reading this article, patients, clinicians and therapists should be able to:
Historical Perspective of Conventional Radiation Therapy
Radiation therapy is a common cancer treatment used in half or more of all patients with malignancies. Although radiation therapy has been effective for many years in curing many cancers, the potential acute and late-term damage to normal tissues is the major limiting factor. The ability to deliver radiation has improved with (1) better understanding of radiation reactions, (2) better radiation equipment, (3) better diagnostic imaging technology, (4) better treatment planning, and (5) better delivery systems.
Treatment planning in radiation oncology has undergone major evolution in the past several decades (Figure 2-1). Early radiation treatment delivery was designed to aim at clinically visible or palpable disease. Later, radiation fields were crudely shaped with standard lead blocks around external landmarks. The introduction of the simulator in the 1960's allowed greater flexibility in beam arrangement and custom shaping of portals. The advances in simulators with fluoroscopy gained wide spread application in treatment. While clinical setups and conventional simulation are still in wide use today, CT-based treatment planning allows precise target definition and more focused (conformal) delivery of radiation dose. "Three-dimensional conformal radiation therapy" (3D-CRT) systems proliferated as a method of permitting higher dose to tumor, limiting dose to normal tissue, and ultimately improving local control and patient outcomes. (1)
Figure 2-1: Historical perspective in treatment planning
Treatment planning with 3D-CRT added flexibility in optimizing delivery of radiation dose to tumor while limiting normal tissue reactions. 3D-CRT allowed radiation oncologists to achieve this goal by the use of multiple portals (often 5 or 6), non-coplanar beam arrangements, proper weighting of beams, and tissue compensation for variations in tissue contour or density. Several clinical trials have demonstrated decreased normal tissue toxicity, particularly with prostate radiotherapy,(2) and the ability to escalate dose for prostate, lung, and brain tumors.(3-5) The treatment planning process also became more complex, as depicted in Figure 2-2. Following patient immobilization and multiplanar imaging, traditional "forward-planning" 3D-CRT required clinicians and dosimetrists to develop beam parameters and evaluate dose distribution more or less by trial and error. For fairly routine cases such as prostate radiotherapy, various sets of standardized beam arrangements have been developed, saving themselves and others the exercise of iterative treatment planning. For more complex cases with irregular tumor contours or varying critical surrounding normal tissues, custom beam parameters need to be developed that satisfactorally deliver adequate dose to tumor while keeping surrounding doses within normal tissue tolerance. With the addition of variables such as oblique beam angles, couch rotation, and 3D dose computation, this process can often be time-consuming and sometimes may not lead to the development of an optimal plan.
Figure 2-2: Forward 3D Treatment Planning Model
Inverse Treatment Planning
In contrast to forward planning, inverse treatment planning (ITP) produces highly conformal dose distributions while minimizing the trial-and-error process. The basic planning process is outlined in Figure 2-3. While the acquisition of patient data is no different, the method of generating beam parameters is revolutionized. In ITP, the radiation oncologist specifies target and normal structure volumes as well as dose-restrictions on these volumes, and optimization of beam parameters is performed by computer.(6) Although there are multiple optimization algorithms, the basic goal is to "backproject" the desired target volume isodose through the patient tissues to determine the most favorable portal geometry and radiation intensity. If the calculated isodose distribution and associated dose-volume histogram is not satisfactory, the optimization is repeated with modifications to clinical parameters, until an acceptable solution is reached. Often these iterations are used for "fine-tuning" of dose-volume constraints. The mathematics of backprojection have been described over 20 years ago, but it is now available for clinical use. Another term used is simulated annealing which is basically an inverse planning process. With the introduction of less expensive and more powerful computing equipment, this optimization process has become practical and cost-effective for many radiation oncology facilities. Automated multileaf collimation available on modern linear accelerators has allowed efficient delivery of multiple and complex portal geometries. A step-by-step outline of the planning process is further discussed below.
Figure 2-3: Inverse Treatment Planning Model
Radiation portals in conventional radiation therapy and 3D-CRT can be shaped by custom blocks or multileaf collimators, but the intensity or fluence of the radiation beam has traditionally been uniform within the treated region. By adjusting the beam-intensity across the field, another dimension to treatment delivery could be provided to permit even more conformal dose distributions. This treatment delivery is known as intensity-modulated radiation therapy, or IMRT.(7) In IMRT, the intensity of the radiation exposure in one portion of the field is modified depending on whether tumor or critical normal structures are present in the beam pathway. Hence, the beam is divided into multiple beamlets, which are the essence of IMRT. When the beamlet hits sensitive normal tissues such as the spinal cord or parotid glands, the intensity is lowered, and when the beamlet hits tumor, the intensity is higher. By dividing the radiation beam into multiple slices, the beam-intensity in any slice can be varied by computer-controlled multileaf collimation during the radiation exposure. A highly schematized model of this process is shown in Figure 2-4, where the treated tumor wraps around a critical normal organ at risk for radiation injury. While the radiation source, or linear accelerator gantry head, moves around the patient, both the shape of the beam portal and the intensity of the beam slices (lighter shades representing higher radiation intensity) are varied to avoid the organ at risk while treating tumor. This results in highly conformal distributions of radiation dose, even around concave tumor volumes.
Figure 2-4: IMRT model
Many vendors have embraced the concept of IMRT, and a number of commercial systems are available for the delivery of IMRT. While there are many potential advantages to IMRT, a number of disadvantages and limitations also exist and will be further discussed later in this article (Table 2-1). The remainder of this two part article will expand upon the treatment planning and delivery of IMRT as well as the present clinical uses of IMRT.
|Conformal dose distribution around tumor
Avoidance of critical structures and less local toxicity
| Equipment costs higher
Special immobilization and gating required
Treatment time often longer
Additional quality assurance necessary
Learning curve can be steep
|Table 2-1: Potential Advantages and Disadvantages of IMRT Compared with Conventional RadiationTherapy.|
ImmobilizationBecause IMRT delivers radiation in a precise fashion, patient positioning and immobilization are crucial in successful treatment of patients. Radiation oncologists pursue IMRT as a method of minimizing dose to normal tissues by tightening margins on tumor volumes. This requires daily reproducible setup accuracy within a few millimeters or less, depending on the type of cancer and the size and location of the target. Certain tumor sites are well-suited for IMRT, such as head and neck cancer, brain cancer, and prostate cancer, because of their relatively stable position in relation to bony anatomy and close proximity to critical normal structures. New immobilization devices have been developed to improve patient immobilization and localization of the target volumes in 3D space.
For head and neck or intracranial tumors, an immobilization device which consists of two screws inserted into the vertex of the skull and fixated to the treatment table had been first used for a variety of intracranial tumors. A noninvasive method using special reinforced thermoplastic masks is more convenient, and results in similar precision in patient setup.(8) Unlike a standard themoplastic mask, these are further stabilized by a custom mold for the occiput and additional straps over the mask which further limit head motion (Figure 3-1). For prostate cancer, many centers are performing body vacuum cast immobilization to limit body motion (Figure 3-2). Because the prostate itself may move within the body due to the contents of the bladder or rectum, additional localization methods have been developed or tested to improve targeting. These methods include transabdominal ultrasound localization,(9) rectal balloon insufflation, or even real-time micro-adjustment based on intraprostatic markers.(10) Each of these techniques has its advantages and disadvantages in terms of accuracy, setup time, and patient inconvenience.
Figure 3-1: Mask immobilization used for cancers of the brain and head and neck
Figure 3-2: Body cradle for prostate treatment
In some forms of IMRT delivery, the patient position is adjusted in relation to the radiation source by movement of the treatment couch. This is true with treatment using the Peacock (NOMOS Corporation) system, or any system which delivers sequential arcs along the length of the patient. Even a 1 mm error in the couch position between arcs can result in a 10% deviation in dosimetry.(11) Therefore, one solution has been the commercial development of a highly accurate table positioning device known as the Crane (NOMOS Corporation), shown in Figure 3-3. The Crane measures the "table index," or the linear-position of the couch, within hundreths of a millimeter in order to minimize dose-inhomogeneity between arcs. A newer table indexing device attaches directly to the treatment couch and is less cumbersome (Figure 3-4).
Figure 3-3: Floor-based Crane table indexing device
Figure 3-4: Crane II table indexing device
Acquisition of Patient Data
Treatment planning for 3D-CRT and IMRT both rely on patient images based on CT scans most commonly, but also MRI and PET scans. Image fusion has been available for better delineation of the target volume. These CT scans are performed with the patient in the treatment position and with necessary immobilization devices in place. Some facilities may rely on a CT scanner in a radiology department, and have the images imported into planning computers. Many radiation oncology departments now have a dedicated CT scanner for image acquisition or a combination CT scanner and "virtual simulator" such as the AcQSim?lt;/sup> (Picker International, Cleveland, OH). Instead of conventional fluoroscopy simulation, a virtual simulator uses the CT images to generate digitally reconstructed radiographic images to design radiation portals. During the acquisition of the CT scan, a number of special markers are placed on the patient or the patient's immobilization device which allow planning computers to localize any point within the patient in 3D space. These markers are known as "fiducial marks," and may consist of small BBs, cross-hatches, rods, etc. Target volumes and normal tissues are then outlined on the CT images for both forward planning and inverse treatment planning systems.
Figure 3-5: CT simulator by Picker
Target and Normal Tissue Definition
Once the CT images are obtained in the treatment position, the radiation oncologist and dosimetrist outline tumor volumes and normal tissue structures. If gross tumor is present, a target corresponding to this gross tumor volume (GTV) is outlined. In addition, a target corresponding to microscopic or subclinical disease may be outlined, and referred to as the clinical target volume (CTV). If there is no gross tumor present, only a CTV may be relevant for a particular case. In IMRT, multiple gross and subclinical target volumes can be outlined, as one may elect to treat each volume with a different daily dose. For treatment, an additional margin may be added to the GTV and CTV to allow for small variations in patient positioning and setup. A detailed description of this terminology can be found in ICRU 50 and ICRU 62 (12,13).
One of the greatests strengths of IMRT is its ability to limit dose to normal structures. The normal organs at risk (OAR) for potential radiation injury must be contoured. For head and neck or intracranial tumors, the OARs generally include the brainstem, lens, optic nerve and chiasm, parotid glands, spinal cord, and portions of normal brain, such as the temporal lobe. For intrathoracic tumors, the OARs may include the spinal cord, heart, lungs, and possibly esophagus. For abdominal tumors, the liver, kidneys, and spinal cord are the structures of greatest concern. In the example shown in Figure 3-6, the primary pancreatic tumor is contoured in red, the nodal regions at risk in green, and the right and left kidneys in blue and purple, respectively. For prostate cancer, the morbidity arises from rectal or bladder injury or weakening of the femoral heads. In forward planning systems, the dose per volume (ie, dose-volume histogram) can be obtained for the critical organs. In ITP, the treatment can be tailored to certain dose-volume constraints defined for these structures.
Figure 3-6: Outlining volumes of interest
Target and Normal Tissue Dose CriteriaOnce the target volumes and normal tissue avoidance structures have been contoured, the desired dose to targets and dose-limits to normal tissues must be specified. The first available inverse planning system allowed the physician to specify a dose-limit or goal to various structures, and assigned each structure a weighting. The differential weighting of tumor and normal tissues tells the computer the relative importance of constraining to the specified dose. In most cases, the tumor receives the highest weighting; however, critical structures can be assigned higher weightings if injury would result in unacceptable consequences, such as blindness or myelopathy. The typical dose-limits and weightings to various normal structures for prostate cancer IMRT used at The Methodist Hospital (Houston, TX; Baylor College of Medicine) are shown in Table 3-1. These dose-limits and weightings are continually refined as experience is gained to adjust for the observed degree of toxicity (or lack thereof) and cure rates in treated patients.
|Normal and Target Structures||Dose Threshold or Goal (cGy)||Weighting|
|Soft Tissues of Pelvis||7,000||0.02|
|Prostate + 4 mm Margin||7,000||2.0|
The target dose goals and normal tissue dose-constraints must be realistic. A set of criteria specifying 100% of dose to the target and 0% of dose to normal tissues is not likely to result in an identifiable treatment plan. On the other hand, the dose-constraints should also represent clinically ideal but achievable intentions. In the case of odd-shaped targets and multiple adjacent normal structures (common with brain tumors), dose-constraints sometimes need to be relaxed; dose-homogeneity within the target may need to be sacrificed in order to reduce dose to a critical structure. A "hot spot" in tumor may be preferable to exceeding tolerance of a critical organ. Inverse treatment planning does not replace clinical judgment, but helps provide solutions to difficult radiotherapy problems.
Following the specification of optimization criteria, the computer performs an iterative search to determine the optimal nonuniform beam fluence pattern. A number of optimization algorithms have been developed.(14) In general, these algorithms express the optimization criteria mathetically as an objective function. The goodness of a proposed treatment plan is assigned a value or score based on the objective function; the optimization algorithm searches for a plan to minimize this score. A simple example is an algorithm which minimizes the squares of the differences between the desired dose and calculated dose to the target, as well as the penalties for exceeding dose-limits to normal structures is shown below:
A patient's CT image is divided into small volume units, known as voxels, and the dose to each voxel is calculated for an initial set of beam weighting. The beam is also divided into small rays, also known as beamlets or pencil beams, and the contribution of each beamlet to the normal tissues and target is determined. The size of the beamlets vary based on the delivery system, but typically involve a voxel size of 1 ? 1 cm3. The weighting of each beamlet is slightly adjusted to produce favorable changes to the overall treatment plan. This process is repeated until the objective function can no longer be improved. The result of the optimization is a planned dose distribution with a relatively uniform dose to the target volume with doses to critical organs at risk, or a portion thereof, held below a tolerance level specified by the radiation oncologist. If the dose distribution is acceptable, the beamlet pattern is converted into instructions for control of the MLC and gantry for patient treatment.
Michalski JM, Purdy JA, Winter K, et al. Preliminary report of toxicity following 3D radiation therapy for prostate cancer on 3DOG/RTOG 9406. Int J Radiat Oncol Biol Phys 2000;46(2):391-402.
Zelefsky MJ, Leibel SA, Gaudin PB, et al. Dose escalation with three-dimensional conformal radiation therapy affects the outcome in prostate cancer. Int J Radiat Oncol Biol Phys 1998;41(3):491-500.
Robertson JM, Ten Haken RK, Hazuka MB, et al. Dose escalation for non-small cell lung cancer using conformal radiation therapy. Int J Radiat Oncol Biol Phys 1997;37(5):1079-85.
Lee SW, Fraass BA, Marsh LH, et al. Patterns of failure following high-dose 3-D conformal radiotherapy for high-grade astrocytomas: a quantitative dosimetric study. Int J Radiat Oncol Biol Phys 1999;43(1):79-88.
Bortfeld T. Optimizing planning using physical objectives and contraints. Sem Radiat Oncol 1999;9(1):20-34.
Verley LJ. Comparison of three-dimensional conformal radiation therapy and intensity-modulated radiation therapy systems. Sem Radiat Oncol 1999;9(1):78-98.
Tsai JS, Engler MJ, Ling MN, et al. A non-invasive immobilization system and related quality assurance for dynamic intensity modulated radiation therapy of intracranial and head and neck disease. Int J Radiat Oncol Biol Phys 1999;43(2):455-67.
Lattanzi J, McNeeley S, Hanlon A, et al. Ultrasound-based stereotactic guidance of precision conformal external beam radiation therapy in clinically localized prostate cancer. Urology 2000;55(1):73-8.
Shimizu S, Shirato H, Kitamura K, et al. Three-dimensional (3D) localization and verification of pelvic tumors using internal markers and a double-axial x-ray television system [Abstract]. Int J Radiat Oncol Biol Phys 1999;45(3 Suppl.):409.
Low DA, Chao KS, Mutic S, et al. Quality assurance of serial tomotherapy for head and neck patient treatments. Int J Radiat Oncol Biol Phys 1998;42(3):681-92.
International Commission on Radiation Units and Measurements. ICRU 50: In Prescribing, Recording, and Reporting Photon Beam Therapy. Bethesda, MD, 1993.
International Commission on Radiation Units and Measurements. ICRU 62: In Prescribing, Recording, and Reporting Photon Beam Therapy (Supplement to ICRU Report 50). Bethesda, MD, 1999.
Bortfeld T. Optimized planning using physical objectives and constraints. Sem Radiat Oncol 1999;9(1):20-34.
Boyer AL, Ochran TG, Nyerick CE, et al. Clinical dosimetry for implementation of a multileaf collimator. Med Phys 1992;19:1255-61.
Jordan TJ, Williams PC. The design and performance characteristics of a multileaf collimator. Phys Med Biol 1994;39:231-251.
Das IJ, Desobry GE, McNeeley SW, et al. Beam characteristics of a retrofitted double-focused multileaf collimator. Med Phys 1998;25:1676-84.
Burman C, Chui CS, Kutcher G, et al. Planning, delivery, and quality assurance of intensity-modulated radiotherapy using dynamic multileaf collimator: a strategy for large-scale implementation for the treatment of carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1997;39(4):863-73.
Eisbruch A, Marsh LH, Martel MK, et al. Comprehensive irradiation of head and neck cancer using conformal multisegmental fields: assessment of target coverage and noninvolved tissue sparing. Int J Radiat Oncol Biol Phys 1998;41(3):559-568.