First, exposure dose with backscatter was measured using a single set of kV X-ray tube and a flat panel detector (FPD) mounted on an O-ring shaped gantry [9] , a cylindrical ionization chamber (DC300, IBA Dosimetry, Germany), and water-equivalent phantoms (Tough Water, KYOTO KAGAKU, Japan). Figure 1 represents a photograph of measurement system of exposure dose. The ionization chamber was rigidly attached on the phantom of 200 mm in thickness. Then, the ionization chamber was positioned to 150 mm upper from the isocenter along the beam axis. Subsequently, exposure dose to a fully opened field of 171 × 226 mm² on the isocenter plane was measured at a speed of 5 fps with X-ray tube current of 200 mA, and X-ray tube voltage of 60, 80, 100, 120 kV, respectively. The nominal exposure time was set to 5 ms.

Next, half-value layer (HVL) of aluminum was measured using a spherical ionization chamber (Exradin A4, Standard Imaging Inc., USA) to estimate the effective energy and absorbed dose conversion factor of soft tissue of kV X-ray.

Then, the skin dose was calculated using the following formula:

D = X × F , (1)

where D is the skin dose (Gy), X the exposure dose with backscatter (C/kg), and F the absorbed dose conversion factor of soft tissue (Gy∙(C/kg)⁻¹) [10] .

First, the top surface of the couch was positioned at the isocenter in the vertical direction. Then, the water-equivalent phantoms (Tough Water) of 0, 50, 100, 200 mm in thickness were put on the couch, respectively. Next, the ionization chamber (Exradin A5, Standard Imaging Inc., USA), was rigidly fixed at the midpoint between the FPD and the isocenter, in the air using carton paper. For the irradiation field of 171 × 226 mm² (fully opened field) and 50 × 50 mm² which is just involving the ionization chamber, the exposure doses to FPD as the amount of ionization were measured with X-ray tube voltage of 60, 80, 100, 120 kV, respectively (Figure 2). Then, the ratio of the amount of ionization for the fully opened field subtracted by the amount of ionization for the field involving the ionization chamber to the

amount of ionization for the field involving the ionization chamber was computed as the scattered dose ratio in FPD.

Under the same condition as described above, amount of ionization in front of FPD was measured with a variety of X-ray tube voltage and thickness of the phantoms of 0, 50, 100, 200 mm for the fully opened field of 171 × 226 mm². Then, the corresponding FPD images were stored. From the FPD images, the ratio of the standard deviation of pixel values within the predetermined region of interest (ROI) with the size of 128 × 128 pixels to the mean within different ROI having little scattered dose from the chamber with the same size was computed using Image-J software (National Institutes of Health, USA) as image noise .

Table 1 represents half-value layer of aluminum, the corresponding effective energy, the absorbed dose conversion factors of soft tissue, and skin doses calculated using Equation (1) for kV X-ray tube voltage of 60, 80, 100, 120 kV, respectively.

Figure 3 shows variations of skin dose for a single exposure as a function of X-ray tube current-time product (mAs). There were good linear correlations between tube current-time product and skin doses for each tube voltage. Then, skin doses per mAs, D / m A s ( v ) ( v | tube voltage, v ∈ 60, 80, 100, 120 kV) (mGy), were calculated as the gradient of the regression lines in Figure 3, resulting in 0.034, 0.078, 0.130, 0.205 mGy for the tube voltage of 60, 80, 100, 120 kV, respectively.

Figure 4 shows variation in the scattered dose ratio in FPD as a function of X-ray tube voltage for the phantoms with 0, 50, 100, 200 mm in thickness, respectively. We have observed that difference in scattered dose ratio in FPD for the phantom with the same thickness was insignificant with respect to the X-ray tube voltage while the scattered dose ratio in FPD became higher for the thicker phantoms. For 60 kV and 80 kV with 200 mm in thickness, the scattered dose ratio could be ignored because of little amount of ionization for the field involving the ionization chamber. In addition, the above tube voltages were out of the clinical conditions.

Figure 5 shows variations in amount of ionization in front of FPD as a function of the phantom thickness for kV X-ray irradiation of 1 mAs with tube voltage of 60, 80, 100, 120 kV, respectively.

On another hand, image noise as a function of amount of ionization in front of FPD for tube voltage of 60, 80, 100, 120 kV was shown in Figure 6. In addition, the corresponding phantom images under each imaging conditions were represented in Figures 6 (a)-(f).

From Figure 6, the approximation formula of the image noise in the FPD, N , was expressed as

N = 0.045 × ( 1 / Q F P D e n ) 0.479 , (2)

where Q F P D e n (pC) denotes amount of ionization in front of FPD.

The above approximation formula was not applicable for the patient with thickness of 50 mm or less because of greater structure mottle.

The kV X-ray source to isocenter distance in the Vero4DRT was very similar to that in the Synergy XVI. The skin dose in the Synergy XVI was 0.08 mGy/mAs for a tube voltage of 100 kV and 0.15 mGy/mAs for a tube voltage of 120 kV, respectively [11] . The corresponding skin doses measured in the Vero4DRT were 0.13 and 0.21 mGy/mAs, respectively. The skin doses in the Vero4DRT were greater by a factor of 1.6 and 1.4 than those in the Synergy XVI with the tube voltage of 100 and 120 kV, respectively. It is because the kV X-ray source in the Vero4DRT has no filtration while a 0.1-mm Cu filter is attached to the kV X-ray source in the Synergy XVI [11] . Our previous study [12] has demonstrated that the skin doses in the Vero4DRT with a 0.1-mm Cu filter attached to the kV X-ray source became 1.1 and 0.99 greater than those in Synergy XVI with the tube voltage of 100 and 120 kV, respectively. The added filtration to the kV X-ray induces a rise in image noise due to reduction of amount of ionization in front of FPD while reducing skin dose. Therefore, we need to take the amount of image noise as well as skin dose into consideration for determination of an optimal thickness of the added filter.

In the protocol of DTT treatment using the Vero4DRT, it is required to create a temporal 4D model between positions of the tumor and those of the abdomen just before the treatment as well as at the rehearsal of the treatment. Then, kV X-ray imaging in two orthogonal directions is performed to acquire the positions of the tumor. Furthermore, another kV X-ray imaging is performed to verify the accuracy of 4D model during the treatment [13] .

Table 2 shows estimated imaging dose for six patients who were enrolled in the DTT treatment for lung cancers in our hospital. Using the imaging conditions such as the tube voltage, the tube current-time product, and the exposure time in the system log file; and the measured skin dose described in 3.1., the accumulated skin dose throughout the DTT irradiation treatment course (48 Gy/4 fractions) was estimated for each kV X-ray tube, respectively. In this study, both angles of the gantry and the ring were fixed to 0 degrees for all ports although

those should be configured to various angles in clinical situation, leading to decrease of overlap of the kV X-ray irradiation fields. Therefore, the estimated skin dose throughout the DTT treatment course should be greater than clinical cases. The average accumulated skin dose among kV X-ray tube #1 and #2 was 0.50 Gy while the maximum accumulated skin dose was 0.93 Gy. The above estimated skin doses were less than 3 Gy in case of which the follow-up care is not forced for the heart IVR patients [3] , suggesting that DTT irradiation for lung cancers can be performed without special attention about the imaging dose although the summed dose of the skin dose and the treatment dose is required to estimate total dose.

Our result has demonstrated that scattered dose to FPD is not increased according to peak tube voltage; and therefore, the high tube voltage does not decrease image quality.

From Equation (2), amount of ionization in front of FPD, Q F P D e n ( N ) (pC) is expressed as

Q F P D e n ( N ) = ( 0.045 / N ) 1 / 0.479 , (3)

where N denotes image noise. Then, tube current-exposure time product Q ( N , t , v ) is expressed as

Q ( N , t , v ) = Q F P D e n ( N ) / Q F P D e n / m A s ( t , v ) . (4)

Here, t , v , Q F P D e n / m A s ( t , v ) denote phantom thickness, tube voltage, amount of ionization in front of FPD per mAs, respectively. Using skin dose per mAs, D / m A s ( v ) , skin dose, D ( N , t , v ) , is expressed as

D ( N , t , v ) = Q ( N , t , v ) × D / m A s ( v ) = ( 0.045 / N ) ( 1 / 0.479 ) / Q F P D e n / m A s ( t , v ) × D / m A s ( v ) . (5)

Next, D / m A s ( v ) was obtained from Figure 3 while Q F P D e n / m A s ( t , v ) was obtained from Figure 5. Using Equation (5), skin dose was computed for the tube voltage of 60, 80, 100, 120 (kV), respectively, as shown in Figure 7.

Figure 8 depicts variations in skin dose with the same amount of image noise as a function of phantom thickness for each tube voltage. Each skin dose was normalized to the skin dose with tube voltage of 120 kV. From Figure 8, the skin dose with tube voltage of 120 kV has analytically become lower than that with any other tube voltage in this phantom study, although skin dose during image guided radiation therapy may be taken into account because of lots of imaging number.