Calibration As Dose-To-Water Or Dose-To-Muscle

Calibration protocols such as TG-51¹ describe how to calculate dose-to-water and thereby provide the link between MU and dose-to-water. However, there is a disconnect when a patient is involved because the patient comprises (primarily) soft tissue, not water. Soft tissue has a different chemical composition than water and therefore different interaction cross sections. (Soft tissue also has a different density than water, but this by itself does not affect dose deposition because dose is energy per unit mass—more energy is deposited in the more dense soft tissue, but this is offset by the increased mass). The difference in chemical composition, however, means that there is approximately 1% less dose deposited in soft tissue than in water.

Historically, this 1% difference was often incorporated into the calibration of the linac. That is, after calibrating the unit as dose to water, a 0.99 factor was applied to this calibration. Such an approach was reasonable to ensure dose was calculated to tissue, and was also consistent with clinical trial standards as the NCI and IROC² support a dose-to-muscle standard.

However, there are several problems with trying to account for differences between soft tissue and water by using a manual 1% correction during the linac calibration. First, this is not done consistently by the medical physics field; 25% of IROC-monitored institutions apply this 1% correction, while the remainder do not. Second, this is a simplified correction that only approximates the difference in dose; for example, it doesn’t account for changes in attenuation with depth between water and muscle. Third, and most importantly, it overlooks the impact of the treatment planning system in the dose calculation process. If the planning system incorporates material composition (not just density) into the calculation algorithm, then the calibration in water will automatically be converted to dose-to-tissue in the dose calculation. However, many algorithms do not incorporate chemical composition into the dose calculation (just density), in which case the algorithm simply calculates dose-to-water (albeit water of varying density).

To understand the impact of the algorithm, and how it relates to the application of a correction factor during the calibration, there are several scenarios that can be considered. If a 0.99 correction is applied to a linac calibration and the planning system algorithm accounts for the fact that the patient is not water, then this difference has been double counted and the delivered dose will be incorrectly low by 1%. On the other hand, if a 0.99 correction is not applied to the linac calibration, and the planning system algorithm treats the patient as made of water, then the delivered dose will be incorrectly high by 1%.

Although this is only a 1% correction, this is a systematic issue that affects calibration and therefore all patients treated. Therefore, to address this issue, and provide consistent guidance to the medical physics community, a report is being prepared through AAPM to provide practical guidance on this issue: "AAPM report on reference dose specification for dose calculations: dose-to-water or dose-to-muscle?" This report aims to provide a common standard (dose-to-tissue) to ensure consistent dose calculation across the practice of radiation oncology. This report, which is expected to be published soon, will recommend how to move from a calibration in water to a dose calculation in tissue by accounting for the differences in different commercial treatment planning system algorithms.