Recent Advances in Radiation Oncology

The field of radiation oncology has experienced dramatic progress in recent years. Advances in areas of tumor delineation, treatment planning, delivery, and verification allow modern radiotherapy to deliver high doses with great accuracy, less patient morbidity, and in a highly individualized manner. A good understanding of what can be achieved with modern radiotherapy is important in ensuring an effective multidisciplinary approach to the management of cancer and other benign, yet rapidly proliferating, lesions.

The past 10 years in radiation oncology have been exciting and full of advances. The large body of high-level evidence supporting the use of radiotherapy as an integral component of cancer care continues to grow. As technology progresses, radiation oncology has evolved into a highly sophisticated and technically demanding specialty.

Radiotherapy involves the use of high-energy radiation, primarily photons in the form of x-rays, to kill proliferating cells in cancers or benign tumors. In addition to photons, other subatomic particles such as electrons, protons, and neutrons can be used. Because most radiotherapy is accomplished using linear accelerators that generate and deliver high-energy x-rays, this article will focus primarily on photon-based radiotherapy.

Radiation causes damage to cells by interfering with the cell's ability to grow and reproduce. Cells that are growing and multiplying are especially sensitive to the effects of radiation. Neoplastic cells reproduce more frequently than normal cells, so they are more susceptible to damage from radiation. Normal cells can also be affected by radiation, but they are better able to recover from radiation damage. The damage to normal cells and the process of repair after radiation treatments underlie the side effects of radiation. These side effects are related to the dose of radiation and the volume of normal tissue that receives radiation. The difference between the responses of neoplastic and normal tissue is critical to radiation's therapeutic ratio. Recent advances have been aimed at amplifying this difference.

As tumors are often immediately adjacent to critical normal structures, the challenge in radiation delivery is primarily a geometric one. One advantage of radiation is that some tissues can be treated to a modest dose and some to a high dose within the same field arrangements. The greater the number of tumor cells, the greater the dose required to destroy them. This has been the rationale for the shrinking-field techniques and the current dose-sculpting techniques of intensity-modulated radiotherapy (IMRT).

Radiation can also be delivered by many different schedules or fractionation schemes, ranging from a single treatment to more than 40 treatments delivered daily over a period of weeks. The radiobiologic benefit of fractionated treatment is that it allows normal tissue to repair and therefore tolerate a higher total dose of treatment. The decision between using a single fraction versus multiple fractions is based on the type of disease, its geometry relative to normal tissues, and the radiation oncologist's experience.

Modern imaging techniques, extending beyond computed tomography (CT) to magnetic resonance imaging (MRI) and positron emission tomography, allow better delineation of targets and adjacent critical normal structures. The incorporation of information from these modalities is crucial to the future of radiation oncology.

CORNERSTONES TO MODERN RADIATION THERAPY

Dose Calculations

Modern high-speed computers have, together with CT imaging, improved the accuracy of radiation dose calculations. Previous estimates of doses were based on one measured point dose and simple rules calculated by hand or on a calculator. Modern dose calculation algorithms calculate dose with vastly improved accuracy using complex algorithms (¹) such as Monte Carlo simulation, an algorithm that simulates the interactions of billions of photons with the patient during a treatment.

Three-Dimensional Computed Tomography-Based Virtual Simulation

CT guidance in defining solid tumor targets and surrounding critical normal tissue structures has allowed radiation therapy to evolve from two-dimensional (2-D) to three-dimensional (3-D) treatment methods. In the past, when 2-D conventional x-ray simulators and planar radiography were used to image the treatment portals, bony and surface landmarks were frequently the only markers the radiation oncologist had to establish the external beam path and field to encompass the target. Radiation therapy beam direction choices were restricted to the anterior-posterior or lateral planes, where one could easily recognize and reproduce bony anatomy position for treatments.

In virtual CT simulation, a complete CT image data set of the relevant region is used to contour target volumes and normal tissue structures. Contours and CT slices are reformatted into 3-D representations, or volumes, which can be visualized from any angle. The CT data used for target volume generation are much more precise than the inferred target volumes used in the 2-D era.

In 1978, Reinstein et al (²) and McShan et al (³) took the first steps toward clinically usable 3-D radiation treatment planning with the development of the beam's eye-view display. The beam's eye-view display provides the radiation oncologist a view from the perspective of the head of the treatment machine, looking down the rays of the divergent beam. This display allows a view of the anatomy similar to that of a 2-D simulator radiograph. However, the 3-D volumes can also be projected onto the beam's eye-view display, giving the radiation oncologist a look at how the treatment beam relates to the tumor volume and the critical normal structures. Previously unusable nonstandard beam angles, such as superior-inferior oblique or vertex projections, can now be planned, projected, and visualized. Figure 1 shows several images from a virtual simulation, including a 3-D volume rendering of a pituitary adenoma patient, a sagittal reconstruction image, a vertex field projection without any target or normal tissue structures (as would have been seen on a 2-D simulator), and the same vertex field projection with the target and critical structures visible.

Fig. 1

Multi-leaf Collimators

To shape radiation fields, early linear accelerators used primary collimators in the x-axis and y-axis to create the smallest necessary rectangular field. Lead alloy, custom-shaped blocks were then placed on mounts in the head of the machine to further refine the shape of the fields. The physical process of creating the blocks and changing them for each patient and for each field involved substantial time and effort. Manufacturers have now developed multi-leaf collimators (MLCs), in which the solid tungsten collimators are split into multiple, individual leaves that can slide against one another and be set individually to create the shape of any of the external blocks used previously.

Figure 2 is a view into the head of an MLC of a linear accelerator, its leaves in position to treat a particular target. Each leaf position is computer-controlled and does not require human intervention, allowing a greater throughput of patients undergoing radiotherapy. Automated MLCs are an essential component of the implementation of IMRT.

Fig. 2

Intensity-Modulated Radiation Therapy

In IMRT, a technique that has risen to the forefront of radiotherapy during the past decade, the intensity of the radiation varies across the treatment field (hence intensity-modulated), in contrast to the homogeneous fields previously used (^4-8).

Conventional 3-D planning computer systems use trial and error to develop a treatment plan. A medical physicist or dosimetrist creates a plan with specific beam geometries and shapes, after which dosages are calculated and evaluated. Each plan is then modified iteratively until an acceptable plan is reached.

By contrast, IMRT uses an inverse planning algorithm. With inverse planning, the goals of the treatment plan (desired dosages to targets and dose limits for critical normal structures) are specified beforehand and the algorithm determines the photon intensity pattern necessary to achieve these goals. Whereas previous treatment fields used homogenous fields (the same photon intensity throughout the entire field), the intensity-modulated fields in IMRT use heterogeneous or variable photon intensities within the same field. The technique used in the delivery of these intensity patterns varies by treatment system.

The two most common IMRT delivery methods currently in use are: (¹) the static, segmental field step-and-shoot MLC technique (⁹); and (²) the dynamic, sliding window MLC technique (¹⁰). Figure 3 demonstrates one example of a static, segmental field technique and the beamlet fluence pattern that can be achieved. Other radiation delivery techniques are described in the next section.

Fig. 3

Inverse planning and IMRT allow for the conformation of the radiation dose around critical structures (optic chiasm, optic nerves, parotid glands), a capability not previously possible with conventional 3-D planning and delivery systems. This conformation has allowed additional dose escalation to the target and increased sparing of normal tissue. Clinicians now can devote more effort toward reducing the side effects of radiotherapy.

An example of an IMRT case is shown in Figure 4, a head and neck IMRT parotid-sparing plan. In patients with head and neck tumors, several normal tissue structures can be delineated, each with its own specific tolerance dose. Sparing the parotid gland, long been recognized as a critical structure (¹¹), has resulted in improved salivary flow rates and avoidance of long-term xerostomia (¹²). IMRT is now being applied to almost all clinical sites currently treated by radiation oncologists.

Fig. 4

RADIATION DELIVERY TECHNIQUES AND SYSTEMS

Radiation delivery techniques and systems are moving toward the ultimate goal of delivering a maximal dose to the target volume and a minimal dose to the surrounding normal tissue. Three components are necessary to achieve this goal: (¹) rapid dose fall-off outside the target volume; (²) conformality of the prescribed dose to the target volume; and (³) impeccable repositioning accuracy. Each of the techniques and systems described excels in the three components to varying degrees, but with different costs in terms of time and complexity of treatment.

Stereotactic Radiosurgery and Stereotactic Radiotherapy

When the total radiation dose is delivered in a single session, the technique is termed stereotactic radiosurgery (SRS). The term radiosurgery is actually a misnomer because no surgical procedure occurs. The name comes from the fact that its creator, Lars Leksell of the Karolinska Institute in Sweden, was a neurosurgeon. Stereotactic refers to the precise 3-D positioning of the patient and targeting of the lesion.

In the setting of intracranial lesions, positioning accuracy is obtained by attaching a stereotactic head ring to the patient's skull with surgical pins. Because the skull surrounds the brain, rigid immobilization of the skull translates into rigid immobilization of its contents. SRS allows for the delivery of a biologically high radiation dose in an effort to maximize local control of malignant tumors or to achieve obliteration of benign lesions such as arteriovenous malformations (AVMs).

When the total dose is delivered in more than one fraction, it is known as fractionated stereotactic radiotherapy (SRT). Immobilization devices less invasive than those used in SRS are used for SRT, such as the rigidly immobilized bite block. Fractionated SRT permits a high degree of conformality to neoplastic lesions with minimum acute and long-term toxicities. However, because of the delivery systems described, the volume of nontarget tissue that receives a significant dose is strongly dependent on the size of the target and the conformality of dose to the target. Although the use of multiple isocenters can increase conformality, conformality is achieved at the expense of dose homogeneity (more hot and/or cold spots) within the target. This inhomogeneity within the target (which may affect tumor control rates) and surrounding normal tissue contrasts with the homogeneity that can be obtained with IMRT, which does not typically use stereotactic immobilization.

GammaKnife

First developed by Leksell for SRS in 1968, the GammaKnife system (Elekta AB, Stockholm, Sweden) concentrates radiation from 201 intersecting beams arising from Cobalt-60 sources mounted within a hemispherical assembly. Typical treatment plans are developed in a forward planning method, and these plans are limited in conformality, as the beams are all circular. To treat a complex lesion, multiple isocenters, or dose spheres, are treated to encompass the lesion. Lesion localization is accomplished with a stereotactic localization frame bolted to the patient's skull.

Arc-based linear accelerator radiotherapy has also been developed. At the University of Michigan, a technique using segmental, conformal stereotactic radiotherapy using standard MLCs set at various positions along the arc gives comparable conformality to GammaKnife SRS while improving on homogeneity (¹³).

GammaKnife and linear accelerator-based SRS may be used to treat intracranial lesions of less than 10 mL, such as metastatic lesions, meningiomas, AVMs, acoustic neurinomas, pituitary tumors, and skull base tumors. The size of lesions treated is limited because as a lesion becomes larger, the inhomogeneity of the treatment increases (undesirable in terms of tumor control), and the dose fall-off is less steep (undesirable in terms of side effects).

CyberKnife

An image-guided robotic stereotactic system called the CyberKnife (Accuray, Sunnyvale, CA) combines a linear accelerator tube, real-time image guidance, and an industrial robot that has six-axis range of motion (¹⁴). The robotic maneuverability is significantly better than that of a conventional linear accelerator, allowing for the implementation of a wider range of treatment plans than any other system. Real-time orthogonal x-ray fluoroscopy tubes determine the patient's position relative to the desired reference position, allowing the robotic arm to make any necessary adjustments, essentially tracking the target. No invasive stereotactic head frame is necessary for this system, as it uses the skeletal structures as a reference frame. The CyberKnife, as seen in Figure 5, can be used to treat the same lesions as the GammaKnife or linear accelerator-based systems.

Fig. 5

The CyberKnife plans are generated either with forward 3-D conformal planning or with inverse planning. Although this system is not considered IMRT, it produces comparably conformal plans. For an unusual geometric target such as a helically shaped tumor wrapped around the spine, the CyberKnife may produce better conformality than IMRT. However, the treatment times are much longer because the robotic arm needs to move to each beam angle.

Extracranial Stereotactic Radiation Therapy

The extracranial application of stereotactic principles, or extracranial stereotactic radiotherapy (ESR), involves the delivery of highly conformal radiation treatment to sites outside the skull, using the principles of stereotaxy for target localization. The recent evolution in tumor imaging, 3-D treatment planning, gated radiotherapy, and tumor tracking have made it physically possible to deliver highly conformal dose distributions to body targets in the lung, pancreas, liver, and spine.

The CyberKnife can be used for ESR. For tumor tracking outside the skull, markers visible to the x-ray imagers can be implanted adjacent to the tumor or directly in it. External surrogate markers visible to an infrared camera can also be placed for additional tracking information. If a patient is breathing, the treatments can be gated, or only turned on when the markers are in the desired treatment position. Other equipment vendors, including BrainLab, Elekta, Philips, Siemens, and Varian, have linear accelerator systems that combine an array of products such as external tracking, aggressive immobilization, and/or real-time imaging, to achieve the principles of stereotaxy necessary for ESR.

TomoTherapy

Another system capable of delivering IMRT is TomoTherapy (TomoTherapy, Inc., Madison, WI). In place of MLCs, a narrow fan collimator and beam rotate isocentrically around the patient. As this rotation occurs, one-dimensional intensity modulation is created and continually modified by shooting vanes of attenuating lead into the aperture at right angles to the direction of the slit. Additionally, the treatment table moves continuously through the field as the modulated beam rotates, thus creating a spiral beam pattern as seen in Figure 6. In addition to the IMRT delivery, the same radiation beam can be used to image the patient for near real-time CT patient information. Treatment can then be altered to account for any target motion caused by bodily functions such as breathing. With TomoTherapy, treatment plans are created with an inverse planning algorithm. The TomoTherapy treatment system can be used for most intracranial and extracranial lesions currently treated by linear accelerators.

Fig. 6

FUTURE DIRECTIONS

With the ability to deliver ever more conformal plans, eliminating patient motion will become especially critical. Tumor-tracking and real-time imaging methods are still in development and further work is needed to realize their full potential. Techniques such as active breathing control can fix the patient at a point within the breathing cycle, avoiding the positional influence of the diaphragm (¹⁵).

Geometric anatomic information can now be fused with functional anatomic information to allow new targets to be established for radiotherapy. Imaging modalities such as functional magnetic resonance imaging and spectroscopy, positron emission tomography, and molecular imaging techniques are likely to reveal the extent of disease much better than CT or MRI, and may even be able to predict the response of target and normal tissues to radiation during treatment. With the addition of these modalities, a new era of image-guided therapies will emerge.

Another recent technical advance in radiation oncology is proton beam-based therapy. Proton beams have a much more favorable beam dose profile (the Bragg peak) when compared with photons. Most of the proton dose is deposited at a depth related to its energy and without the exit dose seen with photons, as shown in Figure 7. Proton beams can create a more conformal treatment compared with photons, but photon-based IMRT has been able to accomplish most of these objectives without the significant expense of a proton treatment facility. In the future, by combining intensity-modulation technology with proton therapy (IMPT), it may be possible to exceed the capabilities of all current systems. But because proton facilities require extremely large and costly cyclotrons (more than $100 million compared with less than $10 million for photon IMRT systems), only a few centers have been able to use them. Proton beam therapy will only come into wide clinical use when its costs are reduced dramatically.

Fig. 7

Another priority will be to develop a better understanding about normal tissue tolerances to partial organ irradiation. Clinicians will then be better able to weigh the trade-offs between dose to target and dose to normal tissue. The resources to analyze the vast amounts of data generated during each treatment are only now beginning to be mobilized for the above priority.

Critical assessment will be needed to prove the benefit of technologies such as IMRT, which are more time-consuming and costly than conventional methods. If, as expected, these new technologies increase tumor control and reduce side effects, large benefits are in store for our patients.

Acknowledgments

The authors thank Steven Kronenberg, Lon Marsh, and Karen Vineberg for assistance with figure preparation.