لخّصلي

Treatment planning
كتاب ريم ص 70
• Treatment planning for proton therapy typically follows a similar scan-plan-treat process that is used for external beam photon radiotherapy.
• The first step in the process is to determine a suitable method of setting up and immobilizing the patient for treatment.
• Immobilization methods are broadly similar to those used for photon radiotherapy, and a range of immobilization devices is commercially available (e.g. thermoplastic moulds, vacuum bags).

• Patients are then imaged in the treatment position on a computed tomography (CT) scanner using kilovoltage X-rays, and outlining of the target volume(s) and relevant organs-at-risk (OARs) is performed on this image.
• While other imaging modalities (e.g. magnetic resonance imaging (MRI), positron emission tomography (PET)) may be used to aid identification of the target volume, the treatment planning process and dose calculations are performed using the X-ray CT image alone.

• The computation of dose within the patient is crucial to any effective radiotherapy treatment.
• In both photon and proton radiotherapy, the X-ray CT image is used as the basis for treatment planning dose calculations.
• Each voxel in this image contains a numerical value (in Hounsfield units).
• In photon radiotherapy, the dose calculation requires that the relative electron density in each voxel is known, and this is easily obtained from the Hounsfield unit value.
• In proton therapy, the dose calculation instead requires that the proton stopping power is known, but the relationship between proton stopping power and Hounsfield unit is not simple.

Fig. 4.3 Illustrations of (a) passive scattering and (b) pencil beam scanning delivery systems

• uses conversion schemes such as the stoichiometric CT calibration method, where CT scans of materials of known elemental composition and density are used to generate a conversion table for HU to proton stopping power(4).
• The uncertainty in this conversion presents a source of uncertainty in the range of proton beams(5).
• Recently researchers have looked to the potential advantage of dual energy CT to provide a more accurate estimate of proton stopping power.
• Many theoretical studies have sought to quantify the potential advantage of dual energy CT.
• However, currently available dual energy CT scanners have not been designed specifically with proton therapy in mind and many of the theoretical gains have not been realized yet in clinical practice(6).
• Despite this, in looking for the best CT scanner for proton therapy, dual energy capabilities are an important consideration.
• Following the calculation of stopping powers from the CT image, proton dose can be computed by treatment planning systems (TPSs).
• As described earlier, dose is deposited by primary protons undergoing elastic Coulomb interactions and secondary particles produced by non-elastic nuclear interactions.
• Many TPSs use pencil beam models to calculate proton dose.
• Pencil beam algorithms convolve infinitesimal proton beamlets with fluence at the position of the beamlet.
• The TPS is used to determine how the prescribed dose to the target volume may be achieved using the delivery system.
• The process of creating a treatment plan has several steps, some of which are performed by the planner and others which are performed by the planning software.
• For those steps that are performed by the planner, departmental protocols may be used to give guidance to standardize plan quality.
• Proton therapy treatments can be delivered using either scattering or scanning technology and Fig. 4.3 shows a schematic diagram of both delivery systems.
• Scattering and scanning are distinct delivery methods, the choice of which affects the dose distribution that can be delivered.
• As such they often require different methods of treatment planning and are considered here separately

كتاب احمد ص 1089 – 1094
A. PRINCIPLES
• Basic principles of radiotherapy treatment planning for protons are essentially the same as for photons and electrons.
• These include acquisition of three-dimensional imaging data set, delineation of target volumes and organs at risk, setting up of one or more beams, selection of beam angles and energies, design of field apertures, optimization of treatment parameters through iterative or inverse planning, display of isodose distributions and dose volume histograms (DVHs), and so on, depending on the complexity of a given case.
•The planning system output for the selected plan includes the necessary treatment parameters to implement the plan (e.g., beam coordinates, angles, energies, patient setup parameters, isodose curves, DVHs, and digitally reconstructed radiographs).
• In the case of protons, additional data are provided for the construction of range compensators and other devices, depending on the type of accelerator and the beam delivery system.
• Because of the very sharp dose drop-off at the end of the beam range and laterally at the field edges and uncertainties in the computed tomography– based water-equivalent depths, calculated beam ranges, patient setup, target localization, and target motion assume greater importance for protons than for photons.
• So a major part of the treatment-planning process for protons consists of taking into account these uncertainties.
• For example, dose distributions are often computed at both the upper and the lower end of these uncertainties.
• Also, corrective techniques, such as “smearing” the range compensator, may be used to counteract the effects of some of the uncertainties (18,19).
• The smearing procedure consists of adjusting the compensator dimensions within the smearing distance, based on the geometric and target motion uncertainties, and thereby shifting its range profile to ensure target volume coverage during treatment (even at the expense of target volume conformality).
• The need and complexity of this procedure require that the proton beam treatment-planning systems must incorporate a smearing algorithm and provide details for the fabrication of the “smeared” range compensator.
• A combination of suitable margins around the clinical target volume and range smearing is essential to ensure target volume coverage at each treatment session.

B. TREATMENT BEAM PARAMETERS
• As discussed earlier, the proton beam is monoenergetic as it enters the treatment head or nozzle.
• The Bragg peak of such a beam, called the pristine peak, is very narrow in depth and is not clinically useful.
• The nozzle is equipped with a range modulation system that creates an SOBP by combining pristine peaks of reduced ranges and intensity (Fig. 27.10).
• Modulation of the proton beam in range and intensity is accomplished by a rotating modulation wheel (also called “propeller”).
• The wheel consists of varying thicknesses of plastic (e.g., polystyrene) with varying angular widths.
• The thickness is constant in a given segment but successively increases from one segment to the other.
• Whereas the water-equivalent range of the pristine peak is reduced by an amount equal to the water-equivalent thickness of the plastic in a segment, its intensity is reduced because of the increasing width of the segment (i.e., increasing beam-on time at that range position).
• As the wheel rotates, the combination of pristine peaks with successively reduced range and intensity creates the desired SOBP (Fig. 27.10).

Figure 27.10. Spread-out Bragg peak (SOBP) depth dose distribution, showing modulation width (distance between the distal and proximal 90% dose values) indicated by the vertical dashed lines. SOBP range is the depth of the distal 90% dose position. (From Kooy HM, Trofimov A, Engelsman M, et al. Treatment planning. In: Delaney TF, Kooy HM, eds. Protons and Charged Particle Radiotherapy. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:70-107.)

• A modern nozzle consists of many components for creating and monitoring a clinically useful beam (e.g., rotating range modulator wheel, range-shifter plates to bring the SOBP dose distribution to the desired location in the patient, scattering filters to spread and flatten the beam in the lateral dimensions, dose-monitoring ion chambers, and an assembly to mount patient-specific field aperture and range compensator).
• These nozzle components are not standard and may vary between different accelerators.
• The SOBP is specified by its modulation width, measured as the width between the distal and proximal 90% dose values relative to the maximum dose (indicated by vertical dashed lines in Fig. 27.10), and its range, measured at the distal 90% dose position.
• SOBP beam parameters are generated by the treatment-planning system for each treatment field.
• Lateral dimensions of the SOBP beam are shaped by a field aperture (corresponding to beam’s-eye-view projection of the field to cover the target), typically constructed from brass with equivalent wall thickness exceeding the maximum possible SOBP range by 2 cm.
• Thus, all the treatment beam parameters for each field, namely, beam energy, SOBP range and modulation, range compensator, field aperture, and dose, are designed by the treatment-planning system.

C. DOSE CALCULATION ALGORITHMS
• Several dose calculation algorithms for proton beam treatment planning have been developed.
• Based on the basic formalisms used, they fall into three major categories: (a) PB, (b) convolution/superposition, and (c) Monte Carlo.
• Some of these algorithms have been adopted by the commercial treatmentplanning systems (e.g., XiO by CMS, Inc., St. Louis, MO; Eclipse by Varian Medical Systems, Inc., Palo Alto, CA).
•The PB algorithm involves the calculation of dose distribution in infinitesimally narrow beams.
• The given field is divided into a fine grid and the PBs are positioned on the grid along ray lines emanating from the virtual source position defining the beam geometry (Fig. 14.52).
• Particles suffer energy degradation through inelastic collisions as well as lateral displacements through multiple elastic scattering, as discussed in Section 27.1B.
• As a result of the elastic scattering interactions, the PB dose distribution gradually expands in lateral dimensions as it traverses the medium until the particles have lost all their kinetic energy through inelastic collisions.
• The PB algorithm calculates the dose distribution in individual PBs, taking into account all the interactions and the medium heterogeneities.
• The dose at any point in the patient is calculated by summing the dose contribution of all the pencils to the point of interest.
• A number of PB algorithms (20–22) have been developed based on Molière’s theory of multiple scattering (23,24).
• Molière’s theory involves a mathematic formalism for the angular distribution of proton fluence as a PB of high-energy protons penetrates a medium.
• The angular distribution is described by a Gaussian function characteristic of multiple small-angle scattering (primarily by nuclei).
• Terms for large-angle scattering and other corrections are also included. Hanson et al. (25) and Deasy (22) have used “best fit” functions to represent Molière angular distribution by a single Gaussian term.
• The use of a Gaussian function to calculate lateral spread of proton fluence in a PB is analogous to the PB algorithm used for electrons (Chapter 14, Section 14.9).
• The computer implementation for protons almost parallels the PB algorithm of Hogstrom et al. (26) for electrons.
• The Gaussian distribution of proton fluence in a pencil is converted to dose distribution by multiplying the fluence by a measured or Monte Carlo–calculated broad-beam central axis depth–dose curve.
• For details of the PB algorithm for protons, based on Molière’s theory of lateral deflections, the reader is referred to Deasy (22).
• The convolution/superposition algorithm for photons was discussed in Chapter 19, Section 19.3. Petti (20,27) has described an analogous algorithm for protons in which dose at any point is determined by summing the dose from PB kernels, placed on the calculation grid.
• The PB kernel is precalculated in a water phantom using a Monte Carlo code. Heterogeneity corrections are made by scaling the kernel dose distribution by electron density in the convolution integral.
• It should be mentioned that the physics of particle scattering in heterogeneous media is not modeled in this algorithm.
• Therefore, the radial spread of particle fluence may not be accurately predicted in very dense or high-atomic-number materials.
• The Monte Carlo method is certainly the gold standard, but it is much slower than the analytic methods used for routine treatment planning.
• However, it is a valuable tool for testing the accuracy of these more practical algorithms.
• For further information on Monte Carlo codes for proton treatment planning, the reader is referred to references (28–30).

D. CLINICAL APPLICATIONS
Proton beam therapy has been used to treat almost all tumors that are traditionally treated with x-rays and electrons (e.g., tumors of the brain, spine, head and neck, breast, and lung; gastrointestinal malignancies; and prostate and gynecologic cancers). Because of the ability to obtain a high degree of conformity of dose distribution to the target volume with practically no exit dose to the normal tissues, proton radiotherapy is an excellent option for tumors in close proximity of critical structures such as tumors of the brain, eye, and spine. Also, protons give significantly less integral dose than photons and, therefore, should be a preferred modality in the treatment of pediatric tumors where there is always a concern for a possible development of secondary malignancies during the lifetime of the patient. For the same reasons, namely dose conformity and less integral dose, lung tumors are good candidates for proton therapy provided the respiratory tumor motion is properly managed. Steepness of the distal dose gradient of the SOBP beam is an attractive feature of protons, but in clinical practice this advantage is not fully realized. The accuracy of localizing the distal dose gradient is marred by several uncertainties: subjective element in target delineation, variations in patient setup, patient and internal organ movements during treatment, and accuracy limitations of the dose calculation algorithms. Therefore, adequate margins have to be added to the target volume to counteract the effects of these uncertainties. In addition to the margins, multiple isocentric beams are used to statistically minimize the uncertainties of adequate target coverage. Sparing of critical normal structures is limited by the same kind of uncertainties as in the dosimetric coverage of target volume. Although single and multiple static beams are often used in proton therapy, there is a trend toward adopting IMPT. Proton dose distributions can be optimized by the use of IMPT, achieving dose conformity comparable to IMRT but with much less integral dose. However, as discussed earlier, IMPT is very sensitive to target motion. Therefore, in cases where target motion is a problem, image guidance is essential to track target motion and ensure target coverage during each treatment. In the above discussion, we have only briefly touched upon various facets of treatment planning. For a comprehensive discussion of the physical and clinical aspects of proton radiotherapy, the reader is referred to the book by Delaney and Kooy (31) and the cited references