Dose evaluation indices for total body irradiation using TomoDirect with different numbers of ports: A comparison with the TomoHelical method

Abstract TomoDirect has been reported to have some advantages over TomoHelical in delivering total body irradiation (TBI). This study aimed to investigate the relationships between the number of ports and the dose evaluation indices in low‐dose TBI in TomoDirect mode using 2–12 ports and to compare these data with those for the TomoHelical mode in a simulation study. Thirteen patients underwent low‐dose TBI in TomoHelical mode from June 2015 to June 2016. We used the same computed tomography data sets for these patients to create new treatment plans for upper‐body parts using TomoDirect mode with 2–12 beam angles as well as TomoHelical mode. The prescription was 4 Gy in two equal fractions. For the TomoDirect data, we generated plans with 2–12 ports with approximately equally spaced angles; the modulation factor, field width, and pitch were 2.0, 5.0 cm, and 0.500, respectively. For the TomoHelical plans, the modulation factor, field width, and pitch were 2.0, 5.0 cm, and 0.397, respectively. D2, D98, D50, and the homogeneity index (HI) were evaluated to compare TomoDirect plans having 2–12 ports with the TomoHelical plan. Using TomoDirect plans, D2 with four ports or fewer, D98 with 10 ports or fewer, D50 with four ports or fewer and HI with five ports or fewer showed statistically significantly worse results than the TomoHelical plan. With the TomoDirect plans, D2 with seven ports or more, D50 with eight ports or more, and HI with eight ports or more showed statistically significant improvement compared with the TomoHelical plan. All of the dose evaluation indices of the TomoDirect plans showed a tendency to improve as the number of ports increased. TomoDirect plans showed statistically significant improvement of D2, D50, and HI compared with the TomoHelical plan. Therefore, we conclude that TomoDirect can provide better dose distribution in low‐dose TBI with TomoTherapy.


| INTRODUCTION
Total body irradiation (TBI) is widely used in conjunction with chemotherapy as a part of the conditioning regimen of hematopoietic stem cell transplantation (HSCT) for hematologic malignancies. [1][2][3][4][5][6] The most commonly applied total dose and fractionation schedule of myeloablative TBI is 12 Gy, twice daily, over 3 days. 7 Low-dose TBI has appeared as an effective form of conditioning in reduced-intensity HSCT for patients who cannot tolerate myeloablation because of their age or comorbidity. The prescription dose of low-dose TBI ranges from 2 to 8 Gy in one to four fractions. [7][8][9][10] TomoTherapy (Accuray, CA, USA) is the delivery of intensitymodulated radiation therapy using the rotational delivery of a fan beam in the manner of a computed tomography (CT) scanner. 11 Currently, TomoTherapy has two radiation modes, TomoHelical and TomoDirect. The former delivers treatment with 360°gantry rotation while the couch is translating through the gantry. In contrast, the latter delivers treatment with 2-12 predetermined, discrete angles using a fixed gantry as the couch passes through. 1,[12][13][14] Some groups have reported the roles and feasibility of TBI using the TomoHelical mode; these studies also showed that TomoHelical offers advantages compared with conventional linear acceleratorbased approaches in terms of dose homogeneity at the target lesion and organ at risk (OAR) sparing. [3][4][5][6][15][16][17] There are few reports of TBI using TomoDirect; however, Salz et al. 1 investigated TBI using Tomo-Direct with 12 static ports with equally spaced angles. They reported the potential advantages of TomoDirect over TomoHelical in TBI as follows. (a) It decreases the risk of interstitial pneumonitis because TomoDirect uses a maximum of 12 fields; since the lung irradiation time from the beginning to the end of treatment is extended, the average dose rate of the lung seems to be decreased. (b) Beam expansion on both edges by a maximum of five leaves each (3.125 cm at the isocenter) if the leaves on the edge of the multi-leaf collimator (MLC) are not used. Therefore, even if a set-up error up to 2 cm of the surface occurs, sufficient dose distribution is ensured. (c) The dose heterogeneity in the circulating blood is improved. 1,18 As far as we know, no report investigating the effect of the number of ports on the dose evaluation indices in TBI using TomoDirect has been published. In this study, we report the relationships between the number of ports and the dose evaluation indices in low-dose TBI with TomoDirect using 2-12 ports and compare the data with those for TomoHelical in a simulation study.

2.A | Patients
Thirteen patients underwent irradiation with 4 Gy in two equal fractions of TBI using the TomoHelical mode of TomoTherapy from June 2015 to June 2016 at Harasanshin Hospital. In this study, we only investigated patients who underwent low-dose TBI because there were too few cases of full-dose TBI (12 Gy/6 fractions) for myeloablative HSCT at our institution. We used the same CT data sets to replan using the modes of TomoHelical and TomoDirect with 2-12 beam angles for this study. The patient characteristics are shown in Co., Ltd., Nagano, Japan). Planning CT images were acquired in the supine position with 5-mm slices using a 64-slice CT (SCENARIA, Hitachi, Ltd., Tokyo, Japan). Since the TomoTherapy system has a couch with a limited translation length, two CT scans (head-first position and feet-first position) were performed for all patients. The headfirst position covered the range between the patients' cranial vertex and the middle of the femurs, and the feet-first position covered the range between the patients' toes and the middle of the femurs.
CTV consisted of an external body contour of the whole body. PTV1 consisted of CTV minus a 5-mm margin under the skin surface.

2.D | Treatment planning and plan evaluation
Treatment planning was performed on the TomoTherapy planning station. The prescription was 4 Gy to cover 85% of the volume of the PTV1, and the dose per fraction was 2 Gy. We did not use dose constraints for OARs, as we do in our clinical routine for low-dose TBI.
For the TomoDirect plans, we generated plans with 2-12 ports with approximately equally spaced angles 1 ; in all, 11 TomoDirect plans were created for each patient. The detailed beam angles of each TomoDirect plan are shown in Table 2 To compare these treatment plans, D2 (near-maximum dose), D98 (near-minimum dose), D50 (median dose), and the homogeneity index (HI) 19 of the PTV1 and the beam-on time were evaluated.
All the dose evaluations were performed on the PTV1. HI was defined by the following equation: HI = (D2-D98)/D50. An ideal value is equal to zero. 19    for all plans in this study.
Although we had two CT data sets because of the limitation of the couch motion with TomoTherapy, we used CT images for the head-first position only for all treatment planning in this study.

2.E | Statistical analysis
We   Fig. 1 show the dose evaluation index results for each TomoDirect plan with 2-12 ports and the TomoHelical plans.
All of the dose evaluation indices of the TomoDirect plans had a tendency to become better as the number of ports increased, but in the TomoDirect 2-port plan, the D2 values ranged from 4.72 to 5.11 Gy, and the D98 values in two patients were less than 3.6 Gy. Also, in the TomoDirect 4-port plan, the D98 value in one patient was 3.44 Gy. Thus, these plans did not meet the criteria of this study. Table 4 and Fig. 2

3.A | Comparison between TomoDirect and
TomoHelical plans Table 3 shows the results of the comparisons of D2, D98, D50, and HI between the TomoDirect plans with 2-12 ports and the TomoHelical plan. The D2 values of the TomoDirect plans with four ports or fewer were statistically significantly worse than that of the TomoHelical plan; however, the values for the TomoDirect plans with seven ports or more were statistically significantly better than that of the TomoHelical plan. The D98 values of the TomoDirect plans with ten ports or fewer were statistically significantly worse than that of the TomoHelical plan, and there were no statistically significant differences between the TomoDirect plans with 11 ports or more and the TomoHelical plan. The D50 values of the TomoDirect plans with four ports or fewer were statistically significantly worse than that of the TomoHelical plan, but those of the TomoDirect plans with eight ports or more showed statistically significant improvement over the TomoHelical plan. In terms of HI, the Tomo-Direct plans with five ports or fewer had statistically significantly worse values than the TomoHelical plan, but the TomoDirect plans with eight ports or more showed statistically significant improvement over the TomoHelical plan.

| DISCUSSION
We investigated the relationships between the number of ports and the dose evaluation indices in TBI using 2-to 12-port TomoDirect plans and compared the results with that of the TomoHelical plan.
We found that TomoDirect had a dosimetric advantage over TomoHelical; however, the beam-on time was longer in TomoDirect.
We should consider prolonging the treatment time with TomoDirect when considering clinical application.
In the TomoDirect 2-port plan, D2 (the near-maximum dose) was 5.11 Gy (4.72-5.34 Gy), and the minimum value of D98 (the nearminimum dose) in 13 patients was 3.52 Gy. Considering the practice guideline 8 and the criteria of this study, the TomoDirect 2-port plan may be unacceptable for clinical use. Furthermore, in the TomoDirect 4-port plan, the minimum value of D98 was 3.44 Gy; thus, this plan would also be unacceptable.
We found statistically significant improvements of D2, D50, and HI in the TomoDirect plans compared with TomoHelical plan. These differences may be caused by the thread effect in TomoHelical mode, which is known to be a dose-variation pattern that manifests as a ripple, which is the result of helical beam junctioning. [20][21][22] al. 22 investigated dose heterogeneity with TomoHelical at various skeletal regions due to the thread effect in total marrow irradiation, and found that the maximum left-to-right arm distance strongly correlated with the thread effect and resulted in dose heterogeneity, particularly in the bones of the arm. They discussed the absolute importance of homogeneous dose delivery even for the extremities, because relapses of hematological malignancies from the extremities have been reported. 23,24 They also mentioned that their findings were applicable for TBI and total skin irradiation. In this study, we observed dose variation patterns caused by the thread effect in the both arms when using the TomoHelical plans shown in Fig. 3. The pitch of 0.430 (pitch = 0.86/n, n: integer) was reported to minimize ripples 20 ; however, we observed these variation patterns to be more conspicuous, especially in the both arms than with a pitch of 0.397. Similar results were reported in a previous study. 22 The pitch of 0.397 was reported to minimize ripples in the setting of an off-axis distance = 20 cm and FW = 5.0 cm. 21 This study also showed a T A B L E 4 Beam-on time of TomoDirect and TomoHelical plans (mean ± SD).

Radiation mode
Beam-on time (s)

| CONCLUSIONS
We investigated the relationships between the number of ports and the dose evaluation indices in low-dose TBI using TomoDirect with 2-12 ports and compared the results to those obtained with TomoHelical. Statistically significant improvements of D2, D50, and HI but not D98 were found with the TomoDirect plans. Further investigations including dose verification and measurements in lowdose TBI using TomoDirect with fewer than 12 ports are necessary.