Hybrid Planning Technique for Pediatric Craniospinal Irradiation with an O-ring Linac

Introduction

Methods

Simulation and Contouring
All patients were immobilized in the head-first supine position, which enables airway access for anesthesia, using a thermoplastic fixation head mask and a vacuum mattress. Computed tomography (CT) images of patients were acquired using Philips Big Bore CT (Philips Medical Systems, Highland Heights, OH, USA) with a 3 mm slice thickness. The whole brain clinical target volume (CTV_brain) was contoured such that the inferior border was at least 0.5 cm below the foramen magnum skull base and included the frontal lobe, cribriform plate region, and superior orbital fissure. PTV_brain was created by adding a 3 mm setup margin.[] The whole spine CTV_spinal was contoured to include the entire thecal sac and extend to the subarachnoid space along the intervertebral foramina and nerve roots. A minimum of 1 cm margin was given to cover the recesses of all vertebral bodies on both sides. The lower border was evaluated with the spinal MRI. The lower border is 2 cm below the end of the subdural space and extends to the S2-S3 interval. PTV_spinal was created by adding a 5 mm setup margin.[] The PTV margins were defined separately due to regional differences in setup uncertainties. A 3 mm PTV margin was used for the cranial region because of more rigid immobilization and higher setup accuracy. In contrast, a 5 mm PTV margin was applied for the spinal region due to the longer treatment field, increased geometric variability associated with patient positioning, and greater setup uncertainty. The final PTV was created by summing PTV_brain and PTV_spinal. The delineated various OARs, including the brainstem, optic chiasm, optic nerves, eyes, lenses, parotids, thyroid, larynx, heart, esophagus, bowel, liver, stomach, kidneys, and lungs. To assess the dose to non-target tissue, the structure named body-PTV was generated by excluding the PTV from the whole-body contour.

Treatment Planning
VMAT, IMRT, and HYBRID plans were generated for each patient with 6 MV FFF photon beams from Halcyon linac in the Eclipse v17.01 (Varian, Palo Alto, CA) treatment planning system (TPS) by the same medical physicist. Halcyon has stacked and staggered dual-layer multileaf collimators (MLC). This MLC configuration offers an advantage in terms of minimizing leaf transmission. The maximum speed of the leaves is 5 cm/s, which is two times faster than Millennium 120-leaf MLC. The dose rate was 800 MU/min. The photon optimizer was used for plan optimization, and final dose calculations were performed using the anisotropic analytical algorithm (AAA), which is the clinically validated and routinely used algorithm in our institution. The calculation grid size was selected as 2.5 mm. The total dose of 36 Gy in 20 fractions was prescribed for PTV.

In VMAT plans, triple coplanar arcs, each with a length of 358° arc, were utilized for cranial fields. The arcs were set at 181° to 179° (clockwise), 179° to 181° (counter-clockwise), and 181° to 179° (clockwise) with collimator angles of 30°, 330°, and 90°, respectively. Two full coplanar arcs were employed with avoidance sectors of 250°-110° in the clockwise direction, and 110°-250° in the counter-clockwise direction for each isocenter of the spine fields since the study emphasized that the 140° posterior arc length is sufficient for spine PTV.[] Furthermore, using a partial arc instead of full arc for spine fields offers the advantage of reducing the low doses received by the body.[] The collimator angles of the arcs were defined as 5° and 355° for spine fields. The number and positioning of isocenters were determined to create an overlapping length of approximately 10 cm between adjacent fields. In all plans, three isocenters were used because the PTV length ranged from 50.2 cm to 59.1 cm. The first isocenter was placed in the brain, and the second and third were in the superior and inferior spine, respectively. There was just a longitudinal axis difference between the isocenters due to their placement. A sample for the placement of isocenters and arc arrangement is shown in Figure 1. All arcs were optimized together at the same time. To improve dose homogeneity inside the field junction, a TPS-integrated automatic feathering (AF) algorithm was utilized during plan optimization. Furthermore, a manual normal tissue objective (NTO) was employed to manage dose fall-off outside the PTV borders, considering the following parameters: Distance from the target border 0.3 cm, start/end doses 105%/60%, and fall-off 1. To prevent hotspots from forming outside of the PTV, a shell structure with a 3 mm separation from the PTV was used. Priorities of shell and NTO, similar to those of OARs, were set at 150 while the PTV was 200. The optimization was done until the dose to the OARs was as low as possible without compromising the target coverage.

Fig 1: The arc arrangement (a) and placement of isocenters (b).

In IMRT plans, an attempt was made to obtain a dose distribution similar to the 3D-CRT technique by using two opposing lateral fields in the brain and a single posterior field for each isocenter in the spinal region, as in the 3DCRT technique. The collimator angle of all fields was set as 0°. The isocenters placement and optimization approach were the same as those of the VMAT plans. The IMRT plans were generated using inverse planning in the treatment planning system. The automatic (NTO) was utilized to manage dose fall-off outside the PTV borders. No shell structure was used. The hot spots were delineated as avoidance structures for subsequent optimizations to diminish them. Additionally, these spots were avoided by using fluence editing. The planner can adjust the fluence distribution of a field in beams eye view (BEV) using the fluence editor. Thanks to this editor, hot spots can be reduced.

In this study, HYBRID plans were generated by independently optimizing VMAT and IMRT plans within the treatment planning system. The total prescribed dose was equally divided between VMAT and IMRT. A dose of 18 Gy in 10 fractions was assigned to each technique, and their sum represented the HYBRID plan.

The plan acceptance criteria for all plans were that at least 95% of PTV received 95% of the prescription dose (D95% ≥ 95% of the prescribed dose), and the doses to OARs were met according to the QUANTEC table of dose limits.[] The mean dose (Dmean) to the lungs, kidneys, heart, and lens should be below 7 Gy, 15 Gy, 26 Gy, and 7 Gy respectively; the heart volume receiving 25 Gy (V25) should not exceed 10%.

Dosimetric Evaluation
All data from 30 treatment plans were collected via dose volume histograms (DVHs) calculated using TPS. VMAT, IMRT, and HYBRID plans were evaluated regarding target coverage, dose homogeneity, conformity, and OAR doses. The following formulas were used to calculate the conformity index (CI) and homogeneity index (HI).[,] The ideal value of the CI is 1.

Fig 7: Conformity index (CI) and homogeneity index (HI) calculation formulas.

In the CI formula, TVRI means the volume covered by reference isodose, TV means the target volume, and VRI means the volume of reference isodose. In the HI formula, Dx% means the dose received by x% volumes of the PTV. A HI value of 0 means that the dose distribution is homogeneous.

Doses of 95% (D95%, target coverage), 98% (D98%, significant minimum), 2% (D2%, significant maximum), Dmean, and volume receiving 107% of the prescribed dose (V107%) of PTV were assessed. The percentage volume receiving at least 2, 5, 10, 15, 20, 25, 30, 34.2, and 36 Gy (V2, V5, V10, V15, V20, V25, V30, V34.2, and V36) of body-PTV was analyzed. A dose of 1 cc (D1cc) and the Dmean were examined for OARs. The percent lung volumes receiving 5, 10, and 20 Gy (V5, V10, and V20) were also compared. Additionally, the total monitor units (MU) for all three techniques were recorded and analyzed.

IBM SPSS software package version 20.0 was used to conduct statistical analysis of the collected data. The analyses were carried out using the non-parametric Wilcoxon signed-rank test due to the limited sample size. A level of P values below 0.05 were considered statistically significant.

Results

Table 1: Dosimetric parameters of PTV and doses for BODY-PTV

Fig 2: The dose distribution of (a) VMAT, (b) IMRT, and (c) HYBRID. The dose color wash legend displays a dose from 50% of the prescribed dose to the plan's maximum value.

Non-Target Tissue (Body-PTV)
The lowest V2 and V5 for body-PTV were obtained with IMRT (p1<0.01 for VMAT vs. IMRT, p3<0.01 for IMRT vs. HYBRID), followed by HYBRID and VMAT plans. There was no significant difference between IMRT and VMAT plans in body-PTV V10. This value was significantly higher in HYBRID plans than in IMRT and VMAT. The other dosimetric parameters calculated in the current study for body-PTV were significantly lower in VMAT plans compared to IMRT and HYBRID. The graphical representation of the doses received by the body-PTV is shown in Figure 3.

Fig 3: Doses received by body-PTV for each plan.

OARs
Table 2 and Figure 4 exhibit a comparison of the mean doses to the OARs for each plan. The mean dose constraints for the heart, lungs, and kidneys were achieved across all three techniques; however, the lens dose constraint was not met in any of the evaluated plans. The ALARA (As Low As Reasonably Achievable) principle was considered for the lens, and efforts were made to minimize the dose as much as possible despite anatomical and planning limitations. Figure 5 shows the lung V5, V10, and V20 for each plan. The Dmean of the lenses and kidneys and the V5 of the lung in IMRT plans were significantly lower than those in VMAT and HYBRID plans. HYBRID plans showed significantly better sparing than VMAT plans in terms of the Dmean of the lenses and kidneys and V5 of the lung. No significant dose difference in the Dmean of the right eye, parotids, larynx, stomach, and lung was observed among the three radiotherapy plans. In comparison with IMRT and HYBRID, the V10 and V20 of the lung, V25 of the heart, and the Dmean of the left eye, thyroid, heart, esophagus, bowel, and liver in the VMAT plans were significantly reduced. Additionally, there was a significant reduction of these OAR doses with HYBRID plans compared to the IMRT plan.

Table 2: Dmean for the OARs of each plan

Fig 4: Dmean of the OARs of each plan.

Fig 5: The lung V5, V10, and V20 for each plan.

Figure 6 demonstrates the D1cc of the OARs for each plan. The lowest D1cc of the right eye, the left optic nerve, parotid, thyroid, heart, lung, esophagus, liver, stomach, and left kidney were obtained with VMAT, followed by HYBRID and IMRT. The IMRT plans were superior regarding the D1cc of the lens and brainstem than HYBRID and VMAT. The D1cc of the left eye and right kidney were minimized in HYBRID plans compared to VMAT and IMRT. In the pairwise comparison for planning techniques, the differences were statistically significant. There were no significant dose differences among the three radiotherapy plans in terms of the D1cc of the right optic nerve, optic chiasma, larynx, and bowel.

Fig 6: D1cc for the OARs of each plan.

Monitor Units
The MU values are presented as median (range) as follows: 1031.7 (1003.1-1086.3) for ARC, 1429.6 (1336.9-1562.5) for IMRT, and 1242.4 (1173.4-1304.5) for HYBRID. IMRT plans exhibited the highest MU values, whereas VMAT plans had the lowest. The MU values of hybrid plans were significantly higher than those of VMAT plans but significantly lower than those of IMRT plans (all comparisons, p<0.01).

Discussion

Conclusion

References