Influence of beamlet width on dynamic IMRT plan quality in nasopharyngeal carcinoma

Study design and patients

The retrospective study was approved by the medical ethics committee of the Third Affiliated Hospital of Sun Yat-sen University (Approval No. [2022]-02-043-01). The need for written informed consent was waived owing to the retrospective nature of the study.

A total of 20 patients with NPC aged 25–60 years and who received IMRT in the Department of Oncology Radiotherapy between February and April 2020 were evaluated. Among them, 16 and four patients were male and female, respectively.

Simulation and contouring

The Elekta synergy accelerator used for irradiation was originally equipped with an MLCi2 offering 40 pairs of leaves with one cm width. Figure 1 shows the division of the fluence map with different beamlet widths. All patients were fixed with head, neck, and shoulder stabilizers and polyurethane foam sealing agent in the supine position. Siemens computed tomography was used, and the scanning range was from the top of the head to two cm below the clavicle head. The reconstructed images were transmitted to the Monaco TPS. Following the International Commission on Radiation Units and Measurements (ICRU) report number 62, the gross target volume (GTV) was divided into the primary tumor gross target volume and neck metastatic lymph node gross target volume. The clinical target volume was divided into high-risk clinical target volume (CTV) and low-risk CTV, which are mainly used to prevent exposure. The PTV was defined as a uniform three-dimensional expansion of the CTV or GTV by three mm. The OARs included the lens, optic nerves, optic chiasm, eyes, spinal cord, brainstem, pituitary, parotids, temporal lobes, mandibular joints, and mandibles.

Planning

All plans used X-rays with an energy of 6 MV in the Monaco TPS for dynamic IMRT planning, which were set to nine beams at 160°, 120°, 80°, 40°, 0°, 320°, 280°, 240°, and 200°. The PTVs and OARs adopted naming rules as recommended by the American Association of Physicists in Medicine task group-263 report (Mayo et al., 2015). The prescribed doses of the primary tumor (PTVnx), neck metastatic lymph nodes (PTVnd), the high-risk clinical target region (PTV1), and the low-risk clinical target region (PTV2) were 70, 66, 64, and 56 Gy, respectively. The prescribed dose was required to contain at least 95% of the target volumes and the OARs, as shown in Table 1. The beamlet widths were set to 2, 4, 6, 8, and 10 mm, and the plans were named BL02, BL04, BL06, BL08, and BL10, respectively. The other parameters and constraint functions remained unchanged. To ensure the accuracy of the calculation, a 2-mm calculation grid was used, and the plans met the prescribed doses and limits of each OAR (Mnejja et al., 2020).

Plan evaluation

The quality score was used to compare the quality of the plans (Bohsung et al., 2005). However, the plan results obtained from some optimized conditions demonstrated that the protective effect of OARs was good, but the dose distribution of the tumors was poor. In such cases, it was difficult to evaluate the overall quality of the plan. To overcome this problem, we used a quality score system to analyze different plans. The plan quality score was used to compare the overall accomplishments of the same dose from the different plans. The group with the best results among the five groups was used as the reference group. A higher plan quality score indicated a less effective overall quality of the plan. A total score of 0 indicated that all indicators of the plan were better than those of the other plans. For example, the minimum dose of the brainstem in the BL10 mm group was 50.90 Gy, with the lowest being 49.10 Gy in the BL04 mm group. Using BL04 mm as the reference group, according to formulas (1) and (2), the score of the brainstems was 0.037 in the BL10 mm group and 0.000 in the BL04 mm group. Other targets involved in the evaluation were also evaluated using the same formula. The overall quality score of the plan was obtained by adding the scores for all the items. (1)${\displaystyle S_j=\left|\frac{M_j-{\mathrm{C}}_j}{C_j}\times P_j\right|}$ (2)${\displaystyle S_D=\sum _{j=1}^k{S}_j}$ where M_j is the actual dose to the j-th anatomical structure (PTV or OAR); C_j is the target dose; P_j is the weighting factor of the anatomical structure (1 in this study); S_D is the overall quality score of the plan; k is the number of targets included in the evaluation; and S_j is the j-th sub-item quality score.

Dose-volume histogram evaluation

For the target area, the evaluation parameters were target coverage (TC), homogeneity index (HI), conformity index (CI), and D_2%. Eq. (3) is the calculation formula for TC. HI and CI were calculated following the ICRU83 report using formulas Eqs. (4) and (5), respectively, (3)${\displaystyle \mathrm{TC}\left(\text{\%}\right)=\left(\mathrm{T}{\mathrm{V}}_{\mathrm{PI}}/\mathrm{TV}\right)\times 100\text{\%}}$ (4)${\displaystyle \mathrm{HI}=\left(D_{2\text{\%}}/{\mathrm{D}}_{98\text{\%}}\right)/{\mathrm{D}}_{50\text{\%}}}$ (5)${\displaystyle \mathrm{CI}={\left(\mathrm{T}{\mathrm{V}}_{\mathrm{PI}}\right)}^2/\left(\mathrm{TV}\times {\mathrm{V}}_{\mathrm{PI}}\right)}$

where D_2% is the dose containing 2% of the target volume, D_98% is the dose containing 98% of the target volume, and D_50% is the dose containing 50% of the target volume. The smaller the HI value, the more uniform the target dose. TV is the PTV, TV_PI is the volume of the target area included in the prescribed dose, and V_PI is the total volume of the prescribed dose line. The closer the CI value is to 1, the better the conformity of the target area.

The MUs of segment is an important parameter in IMRT planning as it can directly affect the plan’s quality and efficiency. Therefore, the influence of the MUs should be considered in plan optimization. When the minimum MUs of a segment is increased from four to eight, the number of segments is reduced by 36.5%; however, the dose of the PTVs and OARs did not change significantly. Therefore, the minimum MUs should be set to 5, considering the quality of the plan.

Statistical analysis

The results were presented as the mean ± standard deviation. The plan results of the optimal group were evaluated using the quality score system to obtain better dose parameters than those obtained using other alternative beam widths. The t-test for normally distributed data and the rank-sum test for non-normally distributed data, followed by Bonferroni’s correction, were used to the intergroup comparison for dosimetric parameters and measurement results. SPSS version 22.0 (IBM Corp., Armonk, NY, USA) was used for all statistical analyses, and p < 0.05 was considered statistically significant.

Overall quality score of plans

The TC, HI, D_2%, D_50%, PTV, and Dmax of the OARs were scored according to the planning scoring system, and the highest plan quality was determined according to the score. A lower score indicated a higher overall plan quality. The results for the overall quality scores of the dynamic IMRT plans with different beamlet widths are presented in Fig. 2. The overall PTVs and OARs scores of the BL02, BL04, BL06, BL08, and BL10 mm groups were 0.236, 0.087, 0.528, 0.1, and 1.249, respectively. The BL04 group had the lowest score (0.087), indicating that the overall quality of the plan was the most effective. For the TC of PTVs, the five groups scored 0.111, 0.009, 0.257, 0.548, and 0.807. The BL04 mm group had the lowest total target area score of 0.009, indicating that it had the highest target coverage. For OARs, the five groups scored 0.125, 0.079, 0.270, 0.457, and 0.441, and the BL04 mm group had the lowest total score, indicating that it had the lowest radiation dose.

PTV

Based on the results of the overall quality scoring system, the BL04 mm group had the lowest total score, indicating that it had the most effective overall quality of the plan. As such, the BL04 mm group was set as the reference group. Table 2 provides the PTVs of the CI, HI, TC, and D_2% for the BL02, BL04, BL06, BL08, and BL10 mm groups. Although the CI, HI and D_2% values were no significant differences among all five groups, the TC was highest with the plan using beamlet widths of four mm. Moreover, compared to the plan with beamlet widths of four mm, the p value of PTVnd for beamlet widths of 6, 8 and 10 mm reached statistical significance (p = 0.015, p = 0.014, and p = 0.042, respectively).

OAR

Table 3 provides the doses required in the OARs of IMRT plans with different beamlet widths for the 20 patients with NPC. The D_2% of the temporal lobes, mandibular joints, brainstem, parotids, pituitary, and eyes were the lowest in the BL04 mm group. The total score for BL04 mm was the lowest, indicating that the BL04 mm group plan had the most effective overall quality. Thus, the BL04 mm group was set as the reference group. The results of the BL08 mm and BL10 mm group changed slightly in the temporal lobes (p = 0.013, p = 0.015), parotids (both p = 0.001), and eyes (p = 0.016, p = 0.024) as evidenced by statistically significant differences, compared to the BL04 mm group, respectively. Though the scores of the optic nerves were significant different (p = 0.027) between BL08 mm and BL04 groups, there were no significant differences between other three group plans. Both compared with BL04 mm group respectively, doses to the optic chiasm between the BL06 mm and BL08 mm groups significantly increased (p = 0.021, p = 0.004), while the plan with beamlet widths of two mm and 10 mm did not reach statistical significance.

Delivery efficiency

The MUs of the BL04 mm group was 837, which was higher than that of the other groups. The BL04 mm group had the longest plan delivery time (PDT) of 358 s. However, the difference was not significant, and the range was within 30 s. The PDT tended to shorten with an increase in the beamlet width, as shown in Fig. 3. Furthermore, the ratio of the number of segments <5 MUs to the total number of segments was calculated. Table 4 shows the results of MUs, PDT, and the ratio of the number of segments <5 MUs for the five IMRT plans. Therefore, the BL04 mm group had the higher MUs than the other groups.