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.

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.

Modeling of normal tissue response to radiation: the critical volume model.

PURPOSE: A model for calculating normal tissue complication probability in response to therapeutic doses of radiation is presented. METHODS AND MATERIALS: The model which we call the "critical volume model" is based on a concept of functional subunits defined either structurally (e.g., nephrons) or functionally, and an assumption that normal tissue complication probability is fully determined by the number or fraction of surviving functional subunits composing an organ or tissue. The essential features of the model are that it takes into account variations in tissue radiosensitivity and architecture of an organ for a single patient and for a patient population, and predicts the normal tissue complication probability under conditions of 3-dimensional inhomogeneity of the dose distribution. The model can be used for Integral Response, or "parallel," organs (where all functional subunits are performing the same function in parallel and the output of the organ is the sum of the outputs of the functional subunits and for Critical Element, or "serial," organs (where damage to one functional subunit results in an expression of damage for the whole organ). The model combines into one compact scheme new concepts and several ideas and models which have been previously developed by other investigators. RESULTS: The behavior of the model is presented and discussed for the example of the kidney, with clinical nephritis as the functional endpoint. CONCLUSIONS: The model has the potential to be a useful tool for evaluation and optimization of 3-dimensional treatment plans for a variety of types of normal tissues.

Strategies for treating possible tumor extension: some theoretical considerations.

When there is a small possibility of cancer having extended to a region some distance from the main bulk of disease, it may be unclear whether to include that region in the target volume and, if so, what dose should be delivered to it. We have constructed a theoretical model that includes dose and volume relationships for both diseased and normal tissue. With this model one can calculate the change in tumor control probability (TCP) when varying doses are delivered to the regions of known and suspected disease. Values of TCP as a function of dose to the region of suspected disease have been calculated for a wide range of the variables on which the model depends. We conclude that the strategy of treating the region of suspected disease to about 70% of the dose delivered to the region of known disease is almost always better than not treating it at all, or treating both regions to a uniform but reduced dose designed to keep the probability of complication the same. The gain in TCP could be from 5 to 15% for situations of clinical interest.

Optimization of 3D radiation therapy with both physical and biological end points and constraints.

A new optimization model is described and its clinical usefulness is demonstrated. The optimization technique was developed to allow computer optimization of 3-dimensional radiation therapy plans with biological models of tumor and normal tissue response to radiation as well as with scores based on physical dose. The emphasis was placed on the optimization model, which should describe, as closely as possible, the goal of the radiation treatment, which is eradication of the tumor while sparing normal tissues. Since the statement of the goals may vary from case to case, a technique that allows a variety of objective functions and types of constraints was developed. The optimization algorithm is capable of handling nonlinear and even discrete score (objective) functions and constraints and effectively explores the vast space of feasible solutions in a relatively short time (minutes of MicroVax 3200 CPU time). An example of computer optimization of radiation therapy of a chordoma of the sphenoid bone using x-ray and proton beams is shown and compared with the best plans achieved by an experienced planner. Directions for future development of the algorithm, allowing optimization of beam orientation, are presented.

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.

Evaluation of volumetric differences through 3-dimensional display of distance of closest approach.

Treatment planning programs offer the ability to view quantitative information on serial CT sections. Many of these programs also offer the ability to view structures of anatomical interest from a beam's-eye view. Such displays may be of wire-frame objects or of shaded surface reconstructions derived from contours drawn in serial CT or MRI sections. Such surface reconstructions are not only shaded to give a 3-dimensional perspective, but can be colored in order to display information about some underlying property of a structure. Color has, for example, been used to demonstrate the distribution of dose on the surface of volumes of interest. We introduce a technique in which color is used to show the spatial differences between volumes of interest by displaying the surface of one volume and encoding in color the distance of closest approach of another volume. Regions in which the two surfaces are within a specified distance are shown in gray; regions in which the first structure lies outside the second are shown in reddish/yellow hues; and regions where the second is outside the first are shown in bluish/green hues with the hue being proportional to the distance apart. Such displays offer insight into anatomical relationships not otherwise easily obtained.

Energy of proton accelerator necessary for treatment of choroidal melanomas.

We have reviewed 94 patients with choroidal melanoma treated by proton beam therapy at the Harvard Cyclotron Laboratory. A beam penetration of f27 mm would be required to treat 90% of the lesions. We conclude that a machine energy of at least 55 and, preferably, 60 MeV would be necessary for a clinically viable therapy unit for the treatment of choroidal melanomas. An extracted beam current of 10(-9) A would be more than sufficient.

Fitting of normal tissue tolerance data to an analytic function.

During external beam radiotherapy, normal tissues are irradiated along with the tumor. Radiation therapists try to minimize the dose of normal tissues while delivering a high dose to the target volume. Often this is difficult and complications arise due to irradiation of normal tissues. These complications depend not only on the dose but also on volume of the organ irradiated. Lyman has suggested a four-parameter empirical model which can be used to represent normal tissue response under conditions of uniform irradiation to whole and partial volumes as a function of the dose and volume irradiated. In this paper, Lyman's model has been applied to a compilation of clinical tolerance data developed by Emami et al. The four parameters to characterize the tissue response have been determined and graphical representations of the derived probability distributions are presented. The model may, therefore, be used to interpolate clinical data to provide estimated normal tissue complication probabilities for any combination of dose and irradiated volume for the normal tissues and end points considered.

Histogram reduction method for calculating complication probabilities for three-dimensional treatment planning evaluations.

New tools are needed to help in evaluating 3-D treatment plans because of the large volume of data. One technique which may prove useful is the application of complication probability calculations. A method of calculating complication probabilities for inhomogeneously irradiated normal tissues is presented in this paper. The method uses clinical estimates of tolerance doses for a few discreet conditions of uniform partial organ irradiation, an empirical fit of a continuous function to these data, and a technique (the effective volume method) for transforming nonuniform dose-volume histograms into equivalent uniform histograms. The behavior of the effective volume histogram reduction method for various boundary conditions is reviewed. The use of complication probabilities in evaluating treatment plans is presented, using examples from an NCI 3-D treatment planning contract.

Accuracy of radiation field alignment in clinical practice.

We present an analysis of simulator and portal films of 71 patients. Twenty-five were analyzed retrospectively, 39 prospectively, but without changing routine filming practice, and 7 had daily portal films taken. Treatment-to-treatment variations in anatomy with respect to the field were determined by comparing sequential portal films. The standard deviation of the variations was approximately normally distributed with an average value of 3 mm independent of site and field shaping technique. Discrepancies between the portal and simulator films were greater and depended on the site of treatment. The mean worst-case discrepancy averaged over all sites was 7.7 mm; the lowest value was 3.5 mm in the head and neck region; the highest value was 9.2 mm in the thorax.

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.

Dose-volume histograms.

A plot of a cumulative dose-volume frequency distribution, commonly known as a dose-volume histogram (DVH), graphically summarizes the simulated radiation distribution within a volume of interest of a patient which would result from a proposed radiation treatment plan. DVHs show promise as tools for comparing rival treatment plans for a specific patient by clearly presenting the uniformity of dose in the target volume and any hot spots in adjacent normal organs or tissues. However, because of the loss of positional information in the volume(s) under consideration, it should not be the sole criterion for plan evaluation. DVHs can also be used as input data to estimate tumor control probability (TCP) and normal tissue complication probability (NTCP). The sensitivity of TCP and NTCP calculations to small changes in the DVH shape points to the need for an accurate method for computing DVHs. We present a discussion of the methodology for generating and plotting the DVHs, some caveats, limitations on their use and the general experience of four hospitals using DVHs.

Potential for improvement in radiation therapy.

A successful strategy for improving the efficacy of radiation therapy has been to improve dose distribution, that is, reduce treatment volume toward target volume. This is so as the smaller treatment volume has permitted a higher dose to the target (hence a high tumor control probability) and a lesser volume of non-target tissues being irradiated (consequently a reduced frequency and severity of treatment related morbidity). There are in place several important means for further improvements in dose distributions. These include: (a) 3D graphic reconstruction of the affected part with definition of the position of the tumor vis-a-vis the adjacent normal structures; (b) explicit inclusion in the treatment plan of the uncertainty band around each isodose contour; (c) on-line contrast enhanced visual monitoring of the target tissue during the individual treatment session; (d) gating of treatment so as to reduce the impact of patient motion on the needed treatment volume; (e) use of computer control systems to execute the treatment; and (f) use of treatment methods which achieve a reduced treatment volume. In an examination for sites for which treatment volumes might be decreased by a substantial factor we have compared treatment volumes for radical surgical and radiation therapy. Results are presented for carcinomas of the cervix (Stage IB), breast (Stage II), floor of mouth (Stage II). We describe a system developed here for on-line visual monitoring of the tissues covered by the treatment field. Brief descriptions are given of results of low LET charged particle radiation therapy and of intraoperative electron beam therapy. Also, the program developed here to use computer graphic techniques to display tumor and normal structures and isodose countours with uncertainty bands around each contour is mentioned.

Tumor and target delineation: current research and future challenges.

In the past decade, significant progress has been made in the imaging of tumors, three dimensional (3D) treatment planning, and radiation treatment delivery. At this time one of the greatest challenges for conformal radiation therapy is the accurate delineation of tumor and target volumes. The physician encounters many uncertainties in the process of defining both tumor and target. The sources of these uncertainties are discussed, as well as the issues requiring study to reduce these uncertainties.

Precise positioning of patients for radiation therapy.

We have developed a number of immobilization schemes which permit precise daily positioning of patients for radiation therapy. Pretreatment and post-treatment radiographs have been taken with the patient in the treatment position and analyzed to determine the amount of intratreatment movement. Studies of patients in the supine, seated and decubitus positions indicate mean movements of less than 1 mm with a standard deviation of less than 1mm. Patients immobilized in the seated position with a bite block and a mask have a mean movement of about 0.5 mm +/- 0.3 mm (s.d.), and patients immobilized in the supine position with their necks hyperextended for submental therapy evidence a mean movement of about 1.4 mm +/- 0.9 mm (s.d.). With the exception of those used for the decubitus position, the immobilization devices are simply fabricated out of thermoplastic casting materials readily available from orthopedic supply houses. A study of day-to-day reproducibility of patient position using laser alignment and pretreatment radiographs for final verification of position indicates that the initial laser alignment can be used to position a patient within 2.2 mm +/- 1.4 mm (s.d.) of the intended position. These results indicate that rigid immobilization devices can improve the precision of radiotherapy, which would be advantageous with respect to both tumor and normal tissue coverage in certain situations.

Tolerance of normal tissue to therapeutic irradiation.

The importance of knowledge on tolerance of normal tissue organs to irradiation by radiation oncologists cannot be overemphasized. Unfortunately, current knowledge is less than adequate. With the increasing use of 3-D treatment planning and dose delivery, this issue, particularly volumetric information, will become even more critical. As a part of the NCI contract N01 CM-47316, a task force, chaired by the primary author, was formed and an extensive literature search was carried out to address this issue. In this issue. In this manuscript we present the updated information on tolerance of normal tissues of concern in the protocols of this contract, based on available data, with a special emphasis on partial volume effects. Due to a lack of precise and comprehensive data base, opinions and experience of the clinicians from four universities involved in the contract have also been contributory. Obviously, this is not and cannot be a comprehensive work, which is beyond the scope of this contract.

Three-dimensional treatment planning for para-aortic node irradiation in patients with cervical cancer.

Three-dimensional treatment planning has been used by four cooperating centers to prepare and analyze multiple treatment plans on two cervix cancer patients. One patient had biopsy-proven and CT-demonstrable metastasis to the para-aortic nodes, while the other was at high risk for metastatic involvement of para-aortic nodes. Volume dose distributions were analyzed, and an attempt was made to define the role of 3-D treatment planning to the para-aortic region, where moderate to high doses (50-66 Gy) are required to sterilize microscopic and gross metastasis. Plans were prepared using the 3-D capabilities for tailoring fields to the target volumes, but using standard field arrangements (3-D standard), and with full utilization of the 3-D capabilities (3-D unconstrained). In some but not all 3-D unconstrained plans, higher doses were delivered to the large nodal volume and to the volume containing gross nodal disease than in plans analyzed but not prepared with full 3-D capability (3-D standard). The small bowel was the major dose limiting organ. Its tolerance would have been exceeded in all plans which prescribed 66 Gy to the gross nodal mass, although some reduction in small bowel near-maximum dose was achieved in the 3-D unconstrained plans. All plans were able to limit doses to other normal organs to tolerance levels or less, with significant reductions seen in doses to spinal cord, kidneys, and large bowel in the 3-D unconstrained plans, as compared to the 3-D standard plans. A high probability of small bowel injury was detected in one of four 3-D standard plans prescribed to receive 50 Gy to the large para-aortic nodal volume; the small bowel dose was reduced to an acceptable level in the corresponding 3-D unconstrained plan. An optimum beam energy for treating this site was not identified, with plans using 4, 6, 10, 15, 18, and 25 MV photons all being equally acceptable. Attempts to deliver moderate or high doses (50-66 Gy) to this region should be made only after careful analysis of the plan with techniques similar to those employed in this study.

Endocrine function following high dose proton therapy for tumors of the upper clivus.

The endocrine status of patients receiving proton radiation for tumors of the upper clivus was reviewed to evaluate the effect of high dose treatment on the pituitary gland. The fourteen patients had chordomas or low grade chondrosarcomas and were all treated by the same techniques. The median tumor dose was 69.7 Cobalt Gray Equivalent (CGE) with a range from 66.6 to 74.4 CGE. (CGE is used because modulated protons have an RBE of 1.1 compared to 60Co). The daily fraction size was 1.8-2.1 CGE. The median follow-up time is 48 months, ranging from 30 to 68 months. All treatments were planned using a computerized multi-dimensional system with the position of the pituitary outlined on the planning CT scan. Review of the dose distribution indicated that the dose to the pituitary ranged from 60.5 to 72.3 CGE, with a median of 67.6 CGE. One female patient had decreased thyroid and gonadotropin function at the time of diagnosis and has been on hormone replacement since that time. The other three females were all pre-menopausal at the time of radiotherapy. At this time four patients (3 males and 1 female) have developed endocrine abnormalities 14 to 45 months after irradiation. All four had evidence of hypothyroidism and two have also developed corticotropin deficiency. The three males had decreased testosterone levels; the female patient developed amenorrhea and hyperprolactinemia. All four are asymptomatic with ongoing hormone replacement.

Conservative treatment of uveal melanoma: probability of eye retention after proton treatment.

Enucleation was performed after proton treatment in 57 of 1006 (5.7%) uveal melanoma patients treated with proton beam therapy at the Harvard Cyclotron Laboratory between July 1975 and December 31, 1986. Only 2% of 99 patients with small tumors and 4% of 566 patients with intermediate size tumors underwent enucleation after treatment; 10% of 341 patients with large tumors lost the treated eye. No eyes were removed after 52 months, with 89% of enucleations performed during the first 36 months after treatment. Eye retention rates at 60 months were 89.1 +/- 3.0% for the entire group, and 97 +/- 3.7%, 92.7 +/- 3.1%, and 78.3 +/- 7.0% in patients with small, intermediate, and large tumors, respectively. Significantly greater enucleation rates were observed in patients with large tumors than in those with intermediate tumors (p = less than .0001), in patients with tumor height greater than 8 mm relative to those with tumors less than or equal to 8 mm, p = (less than .0001), with tumor diameter greater than 16 mm compared to less than or equal to 16 mm, (p = less than .0001), and with tumor involvement of the ciliary body compared to involvement of the choroid only (p = less than .0001). Possible strategies to decrease the likelihood of enucleation in patients at apparently increased risk of losing the eye after conservative therapy, that is, those with large tumors involving the ciliary body, might include a lower total dose, a more protracted treatment course, or a lower radiation dose and adjuvant treatment with chemotherapy and/or immunotherapy, with hyperthermia, or with other radiation sensitizers.