This is a period of unparalleled progress in medicine, with astounding advances in approaches to the treatment of cancer. New surgical methods using "stealth" techniques based on interactive imaging and endoscopic approaches with less morbidity are being introduced. New drug regimens based on novel biologic processes are rapidly being incorporated into the armamentarium of cancer treatment and diagnosis. With the mapping of the human genome completed, can genetic attacks on malignant disease be far behind? Not Your Father's Cobalt Treatment Unit The field of radiation oncology has undergone no less a renaissance. In fact, more advances have been made in radiation oncology in the past 5 to 7 years than in the previous 100+ years since the discovery of the x-ray and radiation-emitting nuclides. The separation and specialization of radiologic imaging and radiation oncology that occurred during the 1980s and 1990s is quickly being reversed. The new reliance on image guidance to establish targets for radiotherapy is forcing radiation oncologists and allied personnel to become versant in the information provided by computed tomography (CT), magnetic resonance imaging (MRI), and positron-emission tomography (PET). More and more, radiologists are entering the therapeutic venue with procedures such as image- guided chemoembolization and radiofrequency-based ablative therapy.[ 1] The new targeted radionuclide therapies-monoclonal antibody- tagged radionuclides that "seek and destroy" selective malignant cells- are being shared by both nuclear radiology and radiation oncology. Virtual Simulation One of the more important recent advances in radiotherapy has been the introduction of CT guidance in defining solid tumor targets and the surrounding normal tissue systems that must be excluded from radiotherapy fields. With conventional x-ray simulators and planar radiography to image the treatment portals, bony landmarks were frequently the only markers the radiation oncologist had to rely on in establishing the external linac beam path to the target. With virtual CT simulation, a complete CT image set allows contouring of the target volume on each axial CT slice along with normal tissue structures such as the spinal cord or kidneys and reformatting of the contoured dataset into coronal, sagittal, or three-dimensional (3D) representations with full surface information marking where the radiation beams will enter the body. Figure 1 shows several images from such a CT-reconstructed dataset. Illustrating the power of virtual simulation, the vertex view in this figure could not be obtained using conventional simulation techniques because a film could not be placed in the correct orientation to obtain this view. This technology led to the introduction of 3D conformal radiation therapy (CRT) and the use of portals that allow radiation to enter the body from superiorinferior oblique angles-a capability that is not available with conventional imaging methods. Dose-Volume Histograms
Computer planning systems (which calculate the dose patterns to targets so contoured in 3D) had to "keep up" with the technologic advance afforded by CT simulation of these targets. Fortunately, most treatment planning companies quickly developed software to import these images to computers and accurately portray the volume dose distributions to contoured targets and normal tissue structures. To adequately evaluate these 3D plans, a new graphic technique involving a "dose-volume histogram" (DVH) allowed the radiation oncologist and medical physicist to "see" the effect of selective beam blocking or small linac gantry angle changes on partial volumes being irradiated by the radiation field portals chosen.[2] Figure 2 illustrates such a DVH for a prostate treatment beam arrangement. The radiation oncologist can now prescribe, for example, "100% of the dose to the target while charging the medical physicist (or dosimetrist) to calculate the 3D plan to give only 50% of the dose to 10% or less of the volume of an adjacent critical structure." This ability has enhanced the protection of critical structures to an extent not previously available. Multileaf Collimation Placing lead-alloy blocks on each field each day of treatment (some plans called for six to eight individual fields) would prove taxing to the radiotherapist charged with daily treatment for up to 6 weeks-the time frame for radiotherapy frequently prescribed to such patients. Again, the linac manufacturers developed a clever solution. Instead of the rectangular collimators that variably blocked the beams exiting the head of the linac (along with shaped lead-alloy blocks further restricting the beam to the target), the solid tungsten collimators were "broken up" into individual leaves that could "slide" against one another and reproduce the shape of any of the external blocks used earlier.[3] Figure 3 is a photo of a modern linac used in radiotherapy today. Figure 4 is a view up into the head of a multileaf collimator-equipped linac showing the leaves in the required configuration to treat a particular target. These leaf positions are set automatically at the treatment console; they not only relieve the therapist from hoist- ing heavy blocks for each field set, but also allow greater throughput of patients undergoing radiotherapy, some using complex portal arrangements. Computer Verification Systems
At the same time that multileaf collimation was being developed, another advance was being made to ensure that all of the correct parameters associated with prescribed radiation treatments were being used for each treatment fraction. Verify-and-record computer systems that evaluated all aspects of the treatment unit (eg, linac, multileaf collimation shape, patient position, dose) were developed to electronically ensure that all parameters were correct for each treatment delivery.[4] If any of the parameters were not set properly, the software would not allow the radiation beam to be energized. This represents another level of patient safety afforded to workers in the field and enabled the confident use of more complicated radiation field arrangements. As with any computerized control system, however, care must be taken to input the correct data initially.[5] Stereotactic Radiosurgery The Leksell GammaKnife system for delivering large, single-fraction doses to benign lesions was developed early on to allow neurosurgeons and radiation oncologists to work together to control disease entities in the brain when conventional surgery was contraindicated.[ 6] The current version of this system consists of 201 small cobalt- 60 sources, all pointing to a common focal point upon which the target lesion will be placed using a stereotactic localization frame. Many disease entities have been treated with this method, including trigeminal neuralgia, arteriovenous malformations, meningiomas, and some metastatic lesions in the brain. A version of this technology was developed for linac-based systems by adding a "postcollimation" cylindrical device that could aim the radiation beam at a target in the brain while the linac gantry moved in an arc about the patient.[7] Several oblique arcs are typically used to "spread" small doses to normal brain tissue over larger areas in order to deliver higher doses to the target. New systems using mini- multileaf colllimators to treat similar targets are rapidly coming into vogue. These systems and their parallel treatment planning systems can provide conformal 3D target coverage, while keeping doses to normal brain tissue to an acceptable level.[8] Intensity-Modulated Radiotherapy In the early 1990s, Dr. Mark Carol, a neurosurgeon, had the clever idea that the large radiation beams being used in conventional radiotherapy could be "broken up" into many smaller beams, each of which could be opened when an appropriate target was "in view" or closed when a critical structure was in the beam's path.[9] This is a simple description of the first application of intensity-modulated radiotherapy (IMRT) using photon beams. The Nomos Corporation's Peacock delivery and planning system was developed to treat solid tumors and "paint a dose picture," much in the same way that CT is used to present "anatomic atlas" axial views of the inside of the body. A postcollimation device, termed the MIMiC, consists of 40 individual binary collimators (they can be opened or closed to block the radiation beam), in two rows of 20 each. Figure 5 depicts an en face view of the MIMiC with every other leaf open/closed. With this system, the tumor target is treated in axial slices along the body axis by rotating the gantry in a (usually) 270^o arc about the patient. At each gantry angle, the leaves may be open or closed depending on the controlling program of the MIMiC computer. This computer takes its instructions from a treatment plan developed earlier with a separate sophisticated treatment planning system. Inverse Planning
In contradistinction to conventional or 3D planning computer systems that use "trial and error" to develop the eventual plan and to a large exent are dependent on the experience and ability of the medical physicist or dosimetrist, the thousands of potential "opening and closings" of the MIMiC collimator varying by gantry angle are impossible to estimate from an a priori approach. Instead, "inverse planning" was developed, which involves using the axial slice CT image data, contouring targets, and normal tissue structures, and sending the information to a robust calculation engine to determine the MIMiC collimator configuration at each gantry angle that would deliver the optimized dose distribution to the target while keeping the dose to critical structures at an acceptably low level. Several iterations of the possible configurations are calculated until the system determines the "best" plan based on criteria selected by the medical physicist and radiation oncologist. Figure 6 shows a treatment plan developed with this inverse planning technique for a prostate carcinoma. Note that the system allows for conformation of the radiation dose around the anterior rectal wall, a capability not possible with conventional planning and delivery systems. This technology was initially applied to the brain, protecting the optical chiasm from irradiation when disease entities such as glioblastoma multiforme or meningioma were in close proximity. Because the high-dose region so tightly conforms in 3D to the target, a slight displacement of the high-dose volume could be dangerous to the patient. Hence, intricate immobilization devices had to be developed to ensure that targets were in the correct place from day to day. With this system, radiation oncologists could treat disease entities like head and neck tumors while sparing the uninvolved parotid glands (which frequently produced xerostomia with conventional methods) and paraspinal tumors in a wraparound configuration (which would not be possible with conventional radiotherapy methods). Subsequently, a patented device was added to the treatment couch that allowed nonaxial slices to be treated by this system. This improvement allowed higher conformality around the target and made possible the introduction of intensity-modulated radiosurgery into the armamentarium of radiation oncologists.[10]