Total body irradiation (TBI) is a specialized radiotherapy technique that delivers a uniform dose of within ±10% of the prescribed dose to the entire body of the patient [1]. It is used to prepare a patient for bone marrow or stem cell transplantation by destroying the number of viable cells and suppressing the recipient’s immune system with total body exposure [2–6]. TBI usually uses the extended source-to-surface distance (SSD) due to three reasons. The first reason is to be able to deliver the beam with a low dose rate of less than 10 cGymin⁻¹ by reducing the beam efficiency. The second reason is to allow a uniform dose distribution along the beam axis. The final reason is to be able to cover the entire body of the patient with large field sizes (e.g. 40×40 cm²) [7–10].

The parallel-opposed lateral irradiation using extended SSD is the most preferred technique in TBI [11]. This stationary beam technique is very simple and is more comfortable for the patient than non-stationary beam techniques such as the translation beam and sweeping beam [12]. Moreover, the arms of the patient could be used as natural lung compensation for irradiation in the lateral direction in order to lower the risk of radiation-induced lung injury [12].

The tissue lateral effect is observed when a beam of insufficient energy is used for a patient which treated with parallel-opposed beams, and can give rise to an excessively higher dose to the peripheral regions compared to that of the midpoint [9]. The thicker the patient, the higher the beam energy required to produce acceptable dose uniformity for parallel-opposed fields [9].

In TBI, the target volume is extended and has an irregular shape; therefore, it is necessary to define one representative point at which the prescribed dose is administered. This point is placed at the patient’s midline and close to the abdomen or pelvis [13]. The TBI procedure delivers the prescribed dose to this reference point and should maintain the dose throughout the body to within ±10% of the prescribed dose. The thickness variation along the medial axis of the patient and the patient diameter along the path of beam also affects dose uniformity. To achieve dose uniformity along the sagittal axis, compensators such as lead or copper plates are designed for the head, neck, shoulder, chest, umbilicus, hip, thigh, knee, calf, and ankle areas, respectively.

In vivo dosimetry should be executed with a dosimeter to verify the dose calculation and compensator design [14]. The entrance and exit doses are measured by in vivo dosimetry during treatment and are used to estimate the delivered midline doses [10]. However, it is difficult to predict the midline dose from entrance and exit dose measurements since these surface doses change depending upon the thickness and beam energy [15]. Satory developed an empirical formula to calculate the midline dose from MOSFET (metal-oxide-semiconductor field-effect transistor) measurements of the entrance and exit doses. The dependency of the surface dose on the air gap between the spoiler and the surface was investigated with various phantom thicknesses. The entrance and exit doses were combined using an exponential attenuation formula to give an estimate of the midline dose and this was compared to the midline ion chamber measurement for a range of phantom thicknesses [10]. Although there are many advantages of using MOSFETs, such as its portability, reproducibility, and direct readout capability, it is a quite cumbersome to install and attach the device to the surface of the patient since the MOSFET device requires a cable (to apply a voltage to the detector) and a dedicated reader device.

OSLDs have been well established to be suitable dosimeters for in vivo measurements due to their reproducibility, low energy dependence, small size, and ease of use [16]. OSLDs have high sensitivity and can measure a surface dose of patient with a non-destructive readout (i.e. without removal the cover or packing) and reanalysis using a simple readout system. Moreover, measurement accuracy can be improved using OSLDs with fully filled deep electron/hole traps and optimal bleaching conditions [17, 18].

In this study, the entrance and exit doses were measured using OSLDs at strategic locations. The strategic locations means the point which patient thickness was measured to fabricate a compensator.

The midline dose was predicted from these measured doses. We investigated the relationship between the midline dose and the surface dose for a range of patient thicknesses. The relationship was evaluated to predict the midline dose from entrance and exit dose measurements for 6 MV and 15 MV.

Table 1 shows the tray factor, spoiler factor, lead attenuation ratio per 1 mm thickness, and source to axis distance factor for 6 MV and 15 MV. The maximum difference between the measured and calculated TMR data was within 2% and 1% for 6 MV and 15 MV, respectively.

Figure 2 shows the plot of the ratio of D_peak to D_m__mid as a function of the phantom thickness for 6 MV and 15 MV. As the thickness of the phantom was increased, D_peak also increased rapidly and the lower the energy, the faster the increase. At 6 MV, the ratio increased to 1.32 at a SWP thickness of 36 cm. When the SWP thickness was greater than 30 cm, D_peak was greater than 10% of D_m__mid at 6 MV. When 15 MV was used at a SWP thickness of 50 cm, D_peak increased to 125% of the midline dose. At a SWP thickness of 30 cm, the ratio was 1.12 for both energies. Therefore, if the thickness of the patient is greater than 30 cm, 15 MV should be used for treatment in order to achieve dose homogeneity [11].

Table 2 shows the values of D_en, D_ex, D_c__mid, D_m__mid, and the ratio of D_c__mid to D_m__mid for 6 MV. Den was increased and D_ex was decreased as the thickness of the SWP was increased. The ratio of D_c__mid to D_m__mid also increased with increasing SWP thickness. However, when the thickness of the SWP was increased from 10 cm to 16 cm, the ratio decreased slightly because the increase of the entrance dose was smaller than the decrease of the exit dose. This is because the entrance dose depends on the distance between the spoiler and the surface of the entrance. For OSLDs measurements, the measurement point was included in the build-up region. Even though the SSD of the measurement point of the entrance dose was decreased, the dose did not increase considerably [10].

The ratio was approximately 1.0 at a SWP thickness of 30 cm. For a pediatric patient, the thickness of most organs will be less than 30 cm. In the case of the neck or ankle, the thickness will be almost 10 cm. Therefore, we can estimate D_c__mid to be less than D_m__mid.

Table 3 shows the results of the 15 MV beam for the same data set as that of the 6 MV beam. The results show a similar tendency to those at 6 MV; at SWP thicknesses between 10 cm and 22 cm, the ratio decreased slightly. In this range, D_c__mid was approximately 85% of D_m__mid. For most adult patients, the thickness of the hips with both hands placed adjacent to the body is at least 40 cm. Therefore, the 15 MV beam was selected for treatment in consideration of the need for dose homogeneity. The thickness of the head and neck is slightly less than 25 cm. Therefore, D_c__mid at these areas of the body can fall below D_m__mid. When the SWP thickness was greater than 20 cm, the ratio increased in keeping with the SWP thickness. If the SWP thickness exceeded 46 cm, the ratio was greater than 1. Most obese patients have organ thicknesses of greater than 50 cm thickness at the shoulder, chest, umbilicus, and hip. D_c__mid of these points can exceed over 10% of the prescription dose. D_c__mid at 15 MV was estimated to be lower than that of 6 MV for same thickness. For 15 MV, a beam spoiler plate of 1 cm thickness is insufficient to produce the initial dose buildup region [10].

For TBI treatment, the acceptable range of dose homogeneity is 10% for most protocols. However, the predicted midline dose by entrance and exit dose measurement can be overestimated even though the real delivered dose is within the acceptable range. The midline dose cannot be predicted accurately from the entrance and exit dose measurements with OSLDs due to an insufficiently full build-up region and the lack of backscatter material. In addition, the tissue lateral effect is varies for different thicknesses at the calculation point.

Therefore, we need a correction to estimate the midline dose using the entrance and exit dose measurements. However, this correction changes for different thicknesses. The predicted midline dose can be corrected with the proper factor for each calculation point. Consequentially, we can predict a more accurate midline dose using the correction factor and surface dose.

In this study, the SWP we have used have discrete thicknesses due to the SWP combination limitations. Therefore, we need to investigate the relationship between the midline dose and the thickness of the SWP with finer resolution.