Half‐beam volumetric‐modulated arc therapy in adjuvant radiotherapy for gynecological cancers

Abstract Purpose The purpose of this study is to introduce half‐beam volumetric‐modulated arc therapy (HVMAT), an innovative treatment planning technique from our work, for reducing dose to the organs at risk (OAR) during adjuvant radiotherapy for gynecological cancers. Methods and materials Seventy‐two treatment plans of 36 patients with gynecological cancers receiving adjuvant radiotherapy were assessed. Among them, 36 plans were designed using HVMAT and paired with the other 36 traditional volumetric‐modulated arc therapy (VMAT) plans for each patient. The main uniqueness of the HVMAT designs was that it consisted of two opposite‐shielded half‐beam fields rotated inversely in two coplanar arcs, collocating with the specially‐devised avoidance structures to enhance the control of the OAR doses. The dose distributions in HVMAT and VMAT were evaluated and compared using the random effects model. Results The ratios of OAR doses in HVMAT compared with VMAT showed a comprehensive OAR dose reduction when using HVMAT (V 20Gy: bladder, 0.92; rectum, 0.95; V 30Gy: bowel, 0.91; femoral heads, 0.66), except for the ilium (V 30Gy: 1.12). The overall mean difference for each OAR across V 40Gy, V 30Gy, V 20Gy, and bowel V 15Gy was statistically significant (almost all p < 0.001). In addition, HVMAT promoted a better conformity index, homogeneity index, D 2%, and V 107% of the planning target volume (all p < 0.001). Conclusions HVMAT is capable of generating deep double‐concave dose distributions with the advantage of reducing dose to several OARs simultaneously. It is highly recommended for pelvic irradiation, especially for treating gynecological cancers in adjuvant radiotherapy.

occur subsequently. Genitourinary and gastrointestinal complications are the most common grade ≥ 3 morbidities after pelvic irradiation. 11,12 To investigate the clinical outcomes and effects of treatment on late morbidity, the EMBRACE studies from the GEC-ESTRO GYN working group and their relevant analyses offer reference materials for treatment parameters. [11][12][13][14][15][16] For organs close to the irradiated pelvic field, including the bladder, rectum, and bowel, they focus on the dose levels of 30 Gy and 40 Gy. 13 In addition to genitourinary and gastrointestinal complications, pelvic bone complications are a serious issue attributable to pelvic irradiation that should also be considered. 17,18 Patients with gynecological cancer undergoing pelvic radiotherapy have increased risks of bone marrow toxicity and pelvic insufficiency fractures. 17,18 Treatment planning therefore becomes complicated because several OARs must be spared simultaneously. However, volumetric-modulated arc therapy (VMAT) is generally associated with low-dose bath to healthy tissues. 19 This drives the motivation to develop a better planning technique to improve the disadvantages of the VMAT. The half -beam volumetric modulated arc therapy (HVMAT) is therefore designed for gynecological cancers treated with adjuvant radiotherapy.

Patients and target delineation
With the ethical approval of a local institutional review board, 36 patients with endometrial or cervical cancer receiving adjuvant radiotherapy at our institution were assessed. All patients had a disease status ranging from stages IA to IIIC and had undergone a radical surgery that included a total hysterectomy with only the vaginal cuff left. The treatment target was delineated based on images of computed tomography (CT) simulation with a full bladder set up by drinking 300-400 ml water 30-40 min prior to the scan. The clinical target volume (CTV) definitions were set according to the NRG oncology/RTOG consensus guidelines, including 3.5-4.0 cm of the proximal vagina as well as paravaginal and retracted parametrical tissue with considerations for internal organ motion; the common, external, and internal iliac nodal regions with a 0.7 cm margin around the vessels; and the obturator nodal region. For patients with cervical cancer, or endometrial cancer with cervical stromal invasion or known lymph node metastases, the presacral nodal region was also involved. 20 The planning target volume (PTV) was generated by expanding 0.7 cm from the CTV in all directions. 20 For VMAT, the arc rotations were the same as for HVMAT, but the fields were unshielded. Two arcs were optimized simultaneously in both types of planning. The treatment plans were designed using the Eclipse™ treatment planning system (TPS),version 11.5 (Varian Medical Systems, Palo Alto, CA, USA), for the linear accelerator, which was the Truebeam STx (Varian Medical Systems, Palo Alto, CA, USA) equipped with a high-definition 120-leaf multileaf collimator (MLC) with 2.5 mm leaves in the middle and 5 mm leaves on the sides. The maximum dose rate was set to 400 MU/min with 10 MV photon beams. The grid size for dose calculation was 2.5 mm. The anisotropic analytical algorithm, version 11.0.31, was used for volume dose calculation, and the progressive resolution optimizer, version 11.0.31, was used for planning optimization.

Planning optimization
In addition to the normal organs of bladder, rectum, bowel bag, ilium, femoral heads, and spinal cord, six avoidance structures were designed to enhance the control of the doses to organs at risk (OAR). Figure 2 illustrates the delineation of these six avoidance F I G U R E 2 Avoidance structures for planning optimization in addition to normal OAR. A 0.5-cm wide ring (purple) surrounds the entire PTV (red) with a distance of 0.5 cm. Another two 1.5-cm wide structures, named AS_low (cyan, ring shape) and AS_up (yellow, rainbow shape), surround the ring to strengthen the dose falloff and are separated by the superior border of the sacroiliac joint. Three individual 2-cm-wide structures are designed for avoiding dose spillage to the bladder, rectum, and bowel, named AS_bladder (green, arrow shape), AS_rectum (pink, stick shape), and AS_bowel (orange, rainbow shape), respectively. Abbreviations: OAR, organs at risk; PTV, planning target volume structures for planning optimization. They include a 0.5-cm wide ring structure to primarily increase the dose falloff, and another two 1.5-cm wide structures separated by the superior border of the sacroiliac joint and surrounding the ring to further strengthen the dose falloff. The other three 2-cm wide structures with specific shapes were individually designed for avoiding dose spillage to the bladder, rectum, and bowel. Among them, the structures for the bladder and rectum extended beyond the body surface to prevent radiation beams from the contralateral side. All structures were at least 0.5-cm away from the PTV to preclude dose conflicts between the target and avoidance structures during planning optimization. Table 1 summarizes the dose-volume constraints and priorities for planning optimization regarding all structures. This objective template was applied in every plan with no discrepancies to avoid systematic biases between different types of planning during optimization. The template was used for the whole optimization process (multi-resolution 1-4) followed by an additional optimization (multi-resolution 4 only). The main scheme was to achieve at least the D 95% of PTV ≥ 42.5 Gy, which was 95% of the prescribed dose, while keeping the bladder V 45Gy ≤ 35 % , the rectum V 30Gy ≤ 60%, and the bowel V 40Gy ≤ 30 % . 21 An experienced medical physicist made all plans with no individual adjustments for different patients.

Plan evaluation
The goodness of the HVMAT and VMAT plans was determined primarily by the dose distributions of OAR, including the bladder, rectum, bowel, ilium, and femoral heads. The representative dose-volume parameters were V 40Gy , V 30Gy , and V 20Gy . For the bowel, V 15Gy was additionally evaluated because of its importance regarding grade ≥ 3 toxicities. 22 Body V 20Gy was the parameter used for the comparison of the low-dose bath between these two techniques. Furthermore, to understand the benchmark of the target dose at which the OAR doses were analyzed, the conformity index (CI), homogeneity index (HI), D 2% , and V 107% of PTV were assessed as the second endpoint. 13,23,24

Statistics
The comparisons of OAR and PTV doses between HVMAT and VMAT plans were conducted using the random effects model with random intercepts, which can estimate the mean difference between HVMAT and VMAT and its p value across different dose-volume parameters, with adjustments for different patients under the assumption of independent and identically distributed random variables. 25,26 The statistical calculations were applied in R, version 4.0.4, a programming language and software environment for statistical computing and graphics supported by the R Foundation for Statistical Computing. A p value < 0.05 was considered significant. Figure 3 indicates that HVMAT facilitated more concentrated dose distributions, which reduced dose to the OAR. Table 2 demonstrates the result of a comprehensive OAR dose reduction in HVMAT compared with VMAT at the expense of a 10% higher ilium dose at V 30Gy and V 40Gy . The lowest mean ratio of the doses in HVMAT against VMAT for each OAR appeared at V 20Gy of the bladder (0.92) and rectum (0.95), and V 30Gy of the

Plan delivery
Under the 3%/3 mm criterion of gamma analysis, portal dosimetry was carried out and passed for all HVMAT plans. On average, the passing rate of area gamma < 1 was 98.9% with a tolerance limit of 97%, the maximum gamma was 1.9 with the tolerance limit of 3, and the average gamma was 0.3 with the tolerance limit of 0.5. The average monitor unit (MU) of HVMAT was 827.8 MU delivered in 2.1 min, which was more than that of VMAT, with an average of 582.6 MU delivered in 2.0 min.

DISCUSSION
The comprehensive dosimetric advantages of HVMAT are derived from several tailored designs that confine the dose to PTV with less dose to the OAR. The designs involve the half -beam fields, the avoidance structures, and the arcs used in HVMAT to generate deep doubleconcave dose distributions for preserving several OAR simultaneously.
Half -beam fields are the most essential part of generating double-concave dose distributions. Each field focuses only on one hemi-side of the body at a time to sculpt the dose to fit PTV and to shield the other hemiside of the body using jaws, meaning that the leaves of the MLC do not need to be moved back and forth across the midline frequently, which can facilitate better shielding of the OAR. This advantage can be observed in the difference in the average MU of HVMAT and VMAT. The average MU of HVMAT is 1.4 times more than that of VMAT,which means that the selection of MU is more delicate in HVMAT. Former studies showed that permanent shielding is beneficial in creating mono-concave dose distributions for anal cancers. 27 Other researchers found two arcs optimal. 28,29 In our HVMAT design, the innovation of two permanent opposite-shielded half -beam fields creates double-concave dose distributions that fit better for the geometry of PTV in gynecological cancers. Moreover, half beams reduce beam divergence that may also help beam focusing.
The avoidance structures also contribute much to the double-concave dose distributions. A single anterior avoidance structure was formerly proven to be a useful tool when planning rectal and anal canal cases. 30 For the more complicated dose distributions that gynecological cancers require,we apply three individual structures with special shapes to control the dose received by the bladder, rectum, and bowel respectively, with no dose conflicts between one another. In addition, structures extend outside the body, reducing dose from the contralateral side. The avoidance structures work well with the halfbeam fields to achieve deep double-concave dose distributions.
The bladder, rectum, and bowel are the ordinary OAR in pelvic radiotherapy. However, the important dosevolume parameters for each of them are not yet conclusive. In general, the dose levels being cared about for the bladder and rectum are usually higher than the dose levels for the bowel. [31][32][33][34][35][36][37] The range of dose levels in regard to grade ≥ 2 urinary incontinence, grade ≥ 1 hematuria, and grade ≥ 1 pain during urination is from 40 Gy to 75 Gy. 31 For grade ≥ 2 late rectal toxicity and grade ≥ 1 rectal bleeding, the range is between 21 Gy and 73 Gy. 32 On the other hand, concerning bowel toxicity such as grade ≥ 2 diarrhea and grade ≥ 2 enteritis, the range is as low as 5 Gy but as high as 50 Gy in some literature. [33][34][35][36][37] Although the prescribed dose only goes up to 45 Gy in our study, it is dangerous if the care of low-dose bath, which is a general drawback of VMAT, is not taken during treatment planning and plan evaluation. HVMAT is therefore a valuable planning technique because it intrinsically reduces OAR doses by means of its designs.
The ilium and femoral heads are complementary normal structures during planning optimization, and the doses are evaluated in our study. Bone complications are not uncommon after pelvic radiotherapy. The proportion of patients with radiation-induced pelvic bone and femoral head complications is approximately 31.1%, including 13.9% with pelvic insufficiency fracture, 12.3% with red bone marrow changes, 4.1% with radiation osteitis, and 0.8% with avascular necrosis of the femoral head. 38 The median time from the end of radiotherapy to the diagnosis of bone complications is around 25 months. 38 However, the doses can be decreased if attention is paid to them. Using the bone structures during planning, optimization effectively reduces the doses without increasing the doses to the bladder, rectum, and bowel. 39 HVMAT can be applied properly with care and attention to detail.
Using daily image guidance such as daily cone beam CT is strongly recommended in combination with HVMAT plans. One reason is because of the tight dose distributions generated by HVMAT. The other is because of the tight vaginal PTV margin used in our study. A vaginal PTV margin of 0.6-0.8 cm can be safe if considering internal organ motion and applying daily cone beam CT image guidance to ensure accurate setup and bladder fill and rectal distension. 20 A margin of 1.5-2.0 cm or greater is recommended if no or minimal image guidance is employed. 20 The use of HVMAT in other tumor groups, such as gynecological tumor without surgical resection and prostate cancer requiring pelvic nodal irradiation, is feasible. Double-concave dose distributions are also needed in these clinical scenarios. Although the OAR doses reduced in a definitive setting are not as much as in an adjuvant setting because the tumor exists, HVMAT still demonstrated dosimetric benefits in our additional examination. Seminal vesicle irradiation is another issue in treating prostate cancer. The elongation of seminal vesicles laterally and superiorly from the prostate between the bladder and rectum can cause a significant dose increase around this area, especially because the dose prescribed for prostate cancer is usually higher. Hence, despite not being the main subject in this study, using HVMAT in other tumor groups ha s potential. This study has some limitations. First, it was a planning comparison study. Although the results were clear and statistically significant, it only yielded logical advantages of HVMAT over VMAT in dosimetric data. Radiotherapy is carried out for better clinical outcomes of patients rather than merely for better treatment planning parameters. Therefore, further clinical follow-up is essential to examine whether the dosimetric advantages can be finally translated into clinical significance, particularly for those with small differences between HVMAT and VMAT. Second, the half -beam design requires the TPS supporting manually opened half -beam fields with at least fixing the jaw at the side of permanent shielding. For the TPS that the jaw opens automatically according to the target, the half -beam design could not be achieved successfully. Third, the extension of avoidance structures exceeding the body surface is a strategy limited to certain planning algorithms. For those with the optimization and calculation volume only within the patient body, this strategy may be unnecessary. Fourth, our study only focused on patients needing pelvic irradiation. If para-aortic irradiation is also demanded, the duodenum, kidneys, and possible transposed ovaries are also OAR abutting the irradiated volume. 13 In this case, the ability of HVMAT to avoid dose spillage anteriorly to the bowel may decrease because these OAR contribute much dose restriction from the lateral side. More attention must be paid to bowel toxicity in this situation.

CONCLUSIONS
HVMAT is an innovative treatment planning technique designed for gynecological radiotherapy. By means of its crafted designs involving the two opposite-shielded halfbeam fields and the specially-devised avoidance structures, it is capable of generating deep double-concave dose distributions with the advantages of reducing dose to the OAR comprehensively also facilitating a better PTV dose. In clinical applications, it is highly recommended for treating gynecological cancers in adjuvant radiotherapy that requires the preservation of several nearby OAR simultaneously.

AC K N OW L E D G M E N T S
This work was supported by the Cathay General Hospital Institutional grant (CGH-MR-B10914). The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. The present study is ethically approved by institutional review board of the Cathay General Hospital. The reference number is CGH-P108083. The research was conducted in the Department of Radiation Oncology, Cathay General Hospital, 280 Renai Rd. Sec.4, Taipei, Taiwan.

C O N F L I C T O F I N T E R E S T
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

AU T H O R S ' C O N T R I B U T I O N S
PCY participated in the design of the study, data collection, and paper writing. CJW involved in the acquisition of data and patient care. HHN conducted the research project. LTL involved in the acquisition of data and patient care. SS involved in the acquisition of data and patient care. YLT participated in the revising of the manuscript critically for important intellectual content. All authors read and approved the final manuscript.