Optimization of radiation therapy II: the critical-voxel model.

The Complication Factor (CF) is an objective function recently introduced for use in the optimization of radiation therapy X treatment planning. Unlike earlier objective functions based upon physical/geometrical criteria, such as tumor dose uniformity, minimal integral-dose, etc., the CF stems from a simple biological/probabilistic model of radiation damage in living organisms. The CF defines the integral-response of an organ as that fraction of it rendered non-functional by irradiation; this parameter is significant if the net amount of damage to the organ is of importance but details of its spatial distribution are not as, for example, might be nearly the case with liver. This approach does not work, however, if complications in any one individual volume-element are critical, as with spinal cord or tumor recurrence. Several authors have addressed the latter problem, and we find that the probabilistic argument common to their methods fits comfortably within the CF framework. Drawing attention to the distinct differences between the integral-response and critical-voxel cases hopefully will be of value in the further development of biological modelling, for application in radiotherapy and elsewhere.

Causes and consequences of inhomogeneous dose distributions in radiation therapy.

Reasons for delivering a non-uniform dose to the target volume are discussed. These include deliberate tailoring of dose to a non-uniform tumor burden or to a non-uniform expectation of the presence of disease, and undesired but unavoidable non-uniformities due to: technical factors; set-up uncertainty; and the need to avoid sensitive organs abutting the target volume. The consequences of non-uniform dose distributions are reviewed and it is suggested that: tumor control may be better characterized by the mean rather than the minimum target absorbed dose when the dose non-uniformity is not too great; modest dose deficits to small sub-volumes of the target volume may not be too deleterious; and modest dose increments to substantial sub-volumes of the target volume may be advantageous. Further modeling, and animal experiments in which tumors are non-uniformly irradiated are required to validate these hypotheses.

A diagnostic X ray field verification device for a 10 MV linear accelerator.

The quality of portal films taken at megavoltage energies is frequently poor, despite the use of various techniques such as lead screens to improve the image. When critical structures, such as the spinal cord, are to be blocked out of the treatment field, it is frequently difficult to determine whether the block is correctly placed. To solve this problem, a diagnostic X ray source has been mounted on the side of a 10 MV linear accelerator to provide accurate verification of patient positioning and location of shielding blocks. The coincidence between the mechanical, 10 MV and diagnostic X ray isocenters is about 1 mm. The system has been designed so that procedures to monitor this coincidence, and adjustment procedures to maintain it, are easily performed.

The simulation of electron beam therapy employing cone collimation.

A simple device is described which is used to determine treatment distance and beam direction in electron therapy employing cone collimation. A technique for the production of irregular field templates and localization films is given.

The accuracy of the line and point source approximations in Ir-192 dosimetry.

The dosimetric approximations used in computer-aided treatment planning of Ir-192 seed implants generally ignore individual seed dimensions and internal structure. Most commonly, each seed is approximated by an isotropic point source. Alternately, each ribbon assembly consisting of uniformly spaced seeds is replaced by an unfiltered line source. Using filtration corrections applicable to platinum- and steel-encapsulated seeds calculated by the Monte Carlo method, the dosimetric errors introduced by these models into two-and-three-dimensional dose distributions arising from multiple plane implants are analyzed. Our results demonstrate that when anisotropy correction factors of 0.96 and 0.99 are used for platinum- and steel-filtered seeds, respectively, the point source model is accurate within 2%. The accuracy of the line source approximation depends significantly upon the details of its implementation. If the linear density of the line is set equal to individual seed strength for an inter-seed spacing of 1 cm, and filtration correction factors of 0.94 and 0.97 are used for platinum- and steel-clad seeds, respectively, the accuracy of the line source model is 1.5% near the implant center. The method of dose-volume histograms is used to compare the predictions of the different models.

A test object for evaluation of portal films.

A test object has been designed for evaluation of the image quality of portal films in high energy photon radiation therapy. It consists of a pattern of notched polyvinyl chloride cylinders, fastened to a plastic sheet and immersed in water during exposure. In a pilot experiment, films produced with the test object were evaluated by a panel of observers. The results indicate that the use of the test object simulates the clinical application of portal films well. It is concluded that the test object can simplify studies of the efficacy of various methods to produce and view portal films.

Quality assurance in radiation therapy: future plans in physics.

Modern day radiation therapy has seen the impact of high technology resulting in more sophisticated computer augmented treatment delivery systems, treatment planning procedures and diagnostic imaging techniques. Much work has already been reported in the area of physics efforts related to quality assurance in radiation therapy. Future efforts in physics will have to address the new developments in each component of the whole radiation treatment process. Certain new developments, using both computer and imaging technologies, show promise in providing tools to verify the accuracy of the delivered radiation treatment. Areas receiving careful attention are: integration and registration of information from multiple sources of diagnostic studies; validation of the accuracy of treatment planning systems; assessment of relative merits of alternate dose distributions; improvement of portal and verification film image quality; real time monitoring using light emitting screens and coupled with TV systems; monitoring of treatment and machine parameters using "record and verify" computer systems. The medical physics community, primarily through the American Association of Physicists in Medicine (AAPM), will continue the development of methodologies for technology transfer in the area of quality assurance. Committees and task groups within the AAPM will address the new developments impacting on quality assurance and prepare appropriate protocols and documents to assist the practicing physicist. By necessity, the national Radiological Physics Center (RPC) and the regional Centers for Radiological Physics (CRP) will have to take a major role in the development of new quality assurance programs.

Lung corrections in photon beam treatment planning: are we ready?

This paper reviews reasons cited for and against the use of lung corrections. It is suggested that all the reasons cited for not making corrections are no longer viable. A phantom has been designed to simulate the thorax region of a patient at both CT and radiotherapy radiation energies. With this phantom, lung correction factors for the calculation of tumor dose have been measured for a typical lung cancer treatment regimen, and these results are shown to compare favorably with correction factors computed by all the commonly employed correction algorithms. Some algorithms are better than others, and one of the best is the readily hand-calculable generalized power-law TAR method. It is shown that failure to correct for lung transmission can severely limit the integrity of many interinstitutional studies, especially cooperative clinical trials. It is concluded that lung corrections for the calculation of tumor doses in the thorax region should be gradually introduced over the next several years.

Localization and protection of kidneys in radiation treatment planning using computed tomography.

Irradiation treatment portals of the upper abdomen must limit the dose to the kidneys. Sparing one-third of the parenchyma of each kidney will prevent late clinical sequelae. One hundred CT scans of the abdomen were studied to evaluate using the vertebrae as landmark for treatment planning. In lateral fields, using the anterior border of the vertebral column as a landmark for the posterior high isodose line will limit treatment to less than 60% (mean 22%) of a single kidney. Placing the edge of an anterior/posterior field 2 cm lateral to the vertebral column will limit the dose to less than 44% of a single kidney (mean 11%).

A simple system for manual image reconstruction from pairs of X ray films.

A simple mechanical back-projection system for X ray films is described which is easy to construct and implement. It enables mechanical simulation of the X ray geometry used when taking pairs of isocentric radiographs for reconstruction purposes. Such pairs may be conventional "AP and Lateral" sets but often it is preferable to take them in oblique directions on the order of 90 degrees apart. The device and the reconstruction method have proved to be very useful in determining target volumes for radiation treatment planning, especially if surgical clips and/or distinct anatomical structures are present. As an instructional tool it has advantages over an also locally developed computer-assisted method of reconstruction. The present system has proved to be highly useful especially in delineating the target volume for treatment planning of soft tissue sarcomas of the extremities and peripheral parts of the body, where detailed and accurate tailoring of shielding blocks is often of vital importance.

Three-field isocentric technique for breast irradiation using individualized shielding blocks.

The three-field technique is the most common method used for breast and regional node treatment after conservative surgery. Several variants of this technique, which are characterized by complex geometrical problems, have been described. A possible simplification of this technique and the use of individualized shielding blocks both for anterior and for tangential fields is proposed, thus allowing for the simultaneous shielding of the half beam and the critical areas. Advantages of isocentrical techniques are thereby maintained, but the number of mechanical movements required is minimized and collimators and couch rotations are not needed. Patient set-up time is also greatly shortened. The accuracy of this technique has been verified using both photographic methods and thermoluminescent dosimetry.

Multi-dimensional treatment planning: I. Delineation of anatomy.

We discuss the scope of a multi-dimensional treatment program designed to assist in planning radiation therapy. It includes: synthesis of diagnostic information; techniques for the assessment and delineation of anatomy; fully three-dimensional simulation of therapy; calculation and assessment of dose distributions; verification of treatment delivery; and assessment of the patient during and after treatment. In this paper we present details of techniques for the assessment and delineation of anatomy, including the display of CT information in three dimensions and the ability to draw on and edit the image displays.

Multi-dimensional treatment planning: II. Beam's eye-view, back projection, and projection through CT sections.

Three features of a fully three-dimensional treatment planning program are presented: (1) The beam's-eye-view provides the user with an accurate reproduction of anatomic features from the viewpoint of a treatment source. The source can be moved to any feasible position relative to the patient, permitting a choice which allows sensitive organs to be excluded from the beam. In this view a field defining aperture can readily be designed. (2) Back-projection of such an aperture shows the parts of the original transverse CT sections, or reconstructed sagittal or coronal sections, which may be covered by the selected beam. (3) Projection through the CT data from any desired origin provides an alignment film simulation which can be used to confirm accuracy of treatment, as well as help establish anatomic relationships relative to the margins of a treatment field.

The role of CT scanning in the radiotherapy planning of pelvic tumors.

In order to assess the place of computed tomography (CT) in radiotherapy planning, the tumor volumes are localized both by conventional techniques and with CT scanning under conditions simulating the radiotherapy. A comparison between the two methods has been made in a group of 55 patients with tumors in the pelvis. CT scanning was found to be of such value in 64% of the treatment series, in improving both the accuracy of localization of the target volume (48% of the patients) and the calculation of the dose distribution (31% of the patients) that its use is recommended during the radiotherapy planning of pelvic tumors.

Pion treatment procedures and verification techniques.

Procedures and techniques developed for the negative pi-meson (pion) radiotherapy program at the Los Alamos Meson Physics Facility, Los Alamos, NM, are reviewed and described. A particular pion patient is followed through the entire planning and treatment sequence to describe CT scanning procedures, bolus and collimator and treatment techniques developed to minimize positioning errors (less than 5 mm). Comparison of 2-D and 3-D isodose calculations developed at Los Alamos showed differences of less than 10% attributable to multiple scattering effects and the computational models used. Treatment verification methods using in vivo ion chamber dosimetry generally confirmed the prescribed dose delivery within 10% and using TLD within 18%.

Use of ultrasound scan in prostatic I-125 implantation.

The average dimension of prostate was measured by transabdominal ultrasonography preoperatively to compare with direct intraoperative measurements in 28 patients undergoing suprapubic I-125 seed implantation for the treatment of prostatic cancer. A highly significant correlation (r = 0.932) was found between these two measurements. Transabdominal ultrasonography is a relatively simple and accurate non-invasive method of determining the number of I-125 seeds needed to implant the patient with prostatic malignancy.

Lung dose calculations using computerized tomography: is there a need for pixel based procedures?

Radiotherapy of tumors within the thorax often requires radiation beams that traverse lung tissues. Not only do lungs consist of large volumes of low density tissues but, in addition, lung tissues are more radiosensitive than most other tissues. Computed tomography (CT) provides the detailed anatomic and geometric data ideally suited for accurate lung dose calculations. However, most radiotherapy centers do not have direct access to a CT scanner for planning purposes. In addition, it is argued by some that it is not necessary to use the CT pixel data directly for lung dose calculations but that it is sufficient to have the lung geometry outlined and to assume some average bulk density (not related to that specific patient) for the total lung volume. This study analyzes the potential errors in lung dose calculations when different lung density assumptions are made. The conclusions show that different levels of accuracy can be achieved with different levels of sophistication in lung density assumptions. However, the delivery of radiation absorbed dose to lung to an accuracy of 5% for all patients will require the detailed anatomical density data that are provided by CT.

Requirements for a treatment planning system for radioimmunotherapy.

Cancer-seeking antibodies carrying radionuclides can, in theory, be very powerful agents for the radiotherapy of cancer. However, as with all radiotherapy, the undesired dose to critical normal organs is the limiting factor that determines success or failure. The distribution of radiation dose in cancer and noncancer tissue is highly dependent on choices the therapist can make: choices of the antigens to be targeted, choices of the antibodies or antibody fragments to be used, choices of radionuclides, of amounts, of timing, and other electives. New technologies, especially of monoclonal antibody production, make the options myriad. Optimization of this therapy depends on a foreknowledge of the radiation dose distributions to be expected. The necessary data can be acquired by established tracer techniques, in individual patients, for particular treatment selections. These tracer techniques can now be implemented by advanced equipment for quantitative, tomographic radionuclide imaging and strengthened by dynamic modeling of the physiological parameters which govern radionuclide distribution, and hence radiation dose distribution.

CT scanning for radiation therapy treatment planning of hepatoma.

One hundred forty-five patients with hepatoma had CT scanning for radiation therapy treatment planning. In order to demonstrate the anatomical distortions that occur with hepatoma and its effect on treatment planning, a control group of 50 colorectal cancer patients with normal livers was analyzed for comparison. The objectives of planning were to deliver as homogeneous a dose to the whole liver as possible and not to treat more than one of two functional kidneys or more than one-half of both functional kidneys. Conventional AP/PA portals were defined by physical examination, intravenous pyelogram, and bowel gas patterns at simulation and were found to be inadequate for the treatment of 76% of patients with hepatoma and 10% of patients with normal livers. Among the control group patients with no hepatoma, only 10% required oblique portals and 6% could not be treated because of left hydronephrosis or a solitary right kidney. Because the distortion of the liver in hepatoma in relationship to the kidneys required portal modification in 76% of hepatoma cases; 39% required oblique planning, 24% AP/PA, 20% PA and left lateral portals, and 17% required 4-field, 3-field or other plans in order to meet the treatment planning objectives. We concluded that all patients receiving radiation therapy to the liver for hepatoma require CT scanning for optimum radiation therapy treatment planning because of the hepatic distortion that occurs in hepatoma and the requirements of renal tolerance.

Dosimetry of CT-guided volumetric IR-192 brain implant.

Over the past 2 years, an afterloading technique has been developed and refined to implant radioactive Ir-192 sources into brain tumors. The implantation procedure integrates a stereotaxic system with computerized tomography (CT), which provides tumor position, volume, and guides the placement of catheters. A radiolucent ring-frame immobilizes the head as holes are made at 1 cm intervals with the aid of a template. Catheters containing dummy sources 1 cm apart are then inserted to the desired depth, and their position verified in three dimensions to insure complete coverage of visible tumor volume as defined by contrast enhancement. Once catheters are secured, the anesthetized patient is moved to the intensive care unit where the dummy sources are replaced by ribbons of Ir-192 seeds (specific activity 0.6-1.0 mg Ra eq). CT scans with the dummy sources in place are used to designate spatial coordinates of the active sources. A computer program converts position data and source strength into isodose contours in any plane. The implant duration (70-100 hours) for the desired dose to the tumor periphery (60-120 Gy) is then calculated. Dose rate contours are superimposed on preimplant CT scans. Maximum and minimum doses are determined in each of the various planes. Verification dosimetry has been carried out with thermoluminescent dosimeters placed in a catheter located in a plane along the tumor periphery. In vivo isodose values compared to idealized plans agree within +/-5%-10%.