Anatomical determination of target volumes and organs at risk is a critical oncologic task and the resultant regions of interest (ROI) are used to design and evaluate a plan. Because contrast agents cannot be used for planning CT to secure correct interpretation of the CT number, ROI delineation is often difficult. Contrast-enhanced X-ray CT and magnetic resonance imaging (MRI) may give high-contrast image of patient anatomy. Positron emission tomography (PET) gives metabolic image of tumor as shown in clinical examples in this book. These imaging modalities are thus useful for identification of gross tumor volume (GTV) on the planning CT image, for which advancement of image-processing technologies has enabled computerized image registration and fused display even in the presence of body deformation.

To plan a beam for a tumor, a radiation oncologist first defines a clinical target volume (CTV) that includes the GTV and surrounding region of clinical margin for potential infiltration. The CTV is the volume to be treated in the planning CT. However, there are always differences between planning CT and treatment times due to physiological changes, organ motion, patient setup error, beam model error, etc. All these uncertainties must be considered at the time of planning. To treat the CTV with prescribed dose in reality, appropriate margins need to be added for a planning target volume (PTV), which is intended to receive the prescribed tumor dose in treatment planning. However, it is generally difficult to quantify these uncertainties. In the cases where the dominant uncertainties are expected in the clinical margins, the setup and internal margins may be disregarded. In such cases, the PTV may be defined as identical to the CTV.

Ideally, the margin against all uncertainties should be included in the PTV. Margin for setup error should expand the PTV in lateral direction while that for range error should expand the PTV in depth direction for each field. However, in common practice, the PTV is defined before the beams are set up, which makes it difficult to define field-specific PTV. Therefore, a common practice is to assign only minimum isotropic margin for PTV to account for the setup uncertainty, which is usually smaller than the range uncertainty. Therefore, the excess of the range uncertainty is usually considered by additional depth margin to cover the PTV with an SOBP of each beam.

As water is the reference material in radiation dosimetry, patients are modeled volumetrically as water of variable effective density (ED) to design beams and to calculate dose distributions. Precise beam-range control to cover the tumor site with SOBP is the essence of C-ion RT. For this purpose, the ED is defined by the stopping-power ratio of the tissue to water for carbon ions. In the planning CT, the CT number ideally represents the X-ray attenuation ratio of the body tissue to water. The conversion from CT number to ED is based on strong systematic correlation between X-ray attenuation and carbon-ion stopping power for human tissues as shown in Fig. 11.2. The CT-ED relation depends on tube voltage, X-ray filter, and hardening effects. Therefore, precise calibration of the conversion curve is necessary for each CT-scanning condition. The hardening effect is object-dependent and therefore considered to be dominant uncertainty for the conversion based on measurement with calibration phantom. Because the conversion uncertainty is generally assumed to be of the order of 1 %, the precision of conversion curves in the same calibration condition should naturally be below 1 %. As there are no available tissue-equivalent materials that are valid for carbon-ion beams, the construction of the CT-ED relationship requires knowledge of the X-ray energy spectrum and elemental composition of tissues, with which X-ray attenuation coefficient or CT number and stopping-power ratio or ED of body tissues can be associated.

In addition to energy loss of carbon ions, which is the major source of radiation dose, multiple Coulomb scattering and nuclear interactions also influence the dose distribution. The majority of body tissues are nearly equivalent to water, which is abundant in oxygen, while bone tissues are abundant in calcium and adipose tissues are abundant in carbon. These compositional differences are ignored in the ED-based patient model. Generally, this approximation may cause only marginal effects to dose distributions in clinical cases.

In general, common tools for treatment planning, such as isodose contours on a CT image and dose-volume histograms (DVHs), are useful also for C-ion RT, where the dose refers to the RBE-weighted dose for clinical evaluation.

For a treatment in multiple positions, one cannot simply sum individual dose distributions on different CT images using rigid image-registration techniques because the patient is usually deformed in different positions. Instead, deformable image-registration techniques may have to be used. Otherwise, for clinical dose evaluation purposes, a virtual plan is made for a virtual treatment machine that delivers equivalently oriented beams in one of the positions. The virtual plan may be useful for clinical review to assess the planned treatment of the patient.

During the review of treatment plans, uncertainties in the treatment planning and delivery process should be considered. The uncertainty in the dose distribution due to patient setup or range calculation may be assessed by calculating dose distributions for the hypothetic cases that include intentional perturbations for patient position and target depth by their estimated uncertainties. This procedure is often referred to as assessment of robustness.

In the scheme where the clinical RBE is an unambiguously formulated definition rather than estimation, no uncertainty is given to the RBE. In other words, the RBE-weighted dose is the basis of prescription as the most relevant quantity to clinical endpoints. Incidentally, the true RBE may be only derived from results of clinical studies. However, direct estimation of the RBE between carbon-ion and the reference X-ray radiations is often impossible due to differences of their standard treatment protocols in fractionation scheme.