Abstract
Background/Aim: The dose to the left anterior descending coronary artery (LAD) is associated with mortality in patients with esophageal cancer (EC) who underwent radiotherapy. The aim of this study was to compare the dose distributions to the LAD region achieved through volumetric-modulated arc therapy (VMAT) planning using a dynamic swing arc in OXRAY (DSA-VMAT) and conventional coplanar (Conv-VMAT) planning.
Patients and Methods: Ten patients with EC who had undergone radiotherapy (60 Gy in 30 fractions) at our Institution were selected for inclusion in the study. Two virtual plans (DSA-VMAT and Conv-VMAT) were created to compare the dose distributions of the LAD region, heart, lungs, and planning target volume (PTV). All plans were analyzed using paired t-tests.
Results: The mean values±standard deviation for 15 Gy to the LAD region (V15) were 10.48±13.04% for DSA-VMAT and 30.28±23.56% for Conv-VMAT. Compared with Conv-VMAT, DSA-VMAT significantly improved V15 of the LAD region (p=0.01). In addition, DSA-VMAT significantly reduced the mean heart dose (8.64±5.37 vs. 11.23±7.37 Gy), heart V40 (17.55±4.76% vs. 20.44±6.06%), lung V20 (14.87±5.93% vs. 17.81±7.70%), and lung V5 (57.27±8.24% vs. 61.15±9.97%) compared to Conv-VMAT (all p≤0.01). In contrast, there were no significant differences between the two groups in PTV dose coverage [D95 (p=0.61), D50 (p=0.62)], or conformity index (p=0.91).
Conclusion: Compared with Conv-VMAT, DSA-VMAT improved the dose distribution of the LAD region without impairing the PTV dose coverage. Thus, DSA-VMAT may reduce radiation-induced heart disease in patients with EC without loss of efficacy.
- OXRAY
- dynamic swing arc
- left anterior descending coronary artery region
- volumetric-modulated arc therapy
- esophageal cancer
Introduction
Esophageal cancer (EC) is one of the most common malignancies with a poor prognosis (1). Concurrent chemoradiotherapy (CCRT) is considered the definitive treatment option for patients with EC who are either ineligible for or refuse esophagectomy (2-4). Although CCRT is an effective treatment for EC, some patients develop severe adverse effects, including esophageal, pulmonary, and cardiac toxicities (5). Therefore, intensity-modulated radiotherapy, which administers irradiation that matches the shape of the tumor and reduces the dose reaching normal organs, is increasingly used in CCRT to treat EC.
Radiation-induced heart disease (RIHD) is a severe complication of heart irradiation. In patients with EC, RIHD occasionally occurs following CCRT and can affect clinical outcomes (6-8). Although initial studies evaluating RIHD examined the radiation dose to the entire heart, a growing number of recent studies have analyzed the radiation dose to cardiac substructures, such as the left anterior descending coronary artery (LAD), and RIHD-related effects on clinical outcomes after thoracic radiotherapy (9). Although a longer follow-up period is generally needed to detect RIHD, the prognosis of patients with EC has improved substantially in recent years (10-13). Therefore, to further improve treatment outcomes in patients with EC, CCRT treatment plans that reduce heart doses, particularly to the LAD, are required.
Recently, OXRAY (Hitachi, Ltd., Tokyo, Japan), a novel O-ring-type linear accelerator system, was released. The O-ring rotates the synchronized gantry and enables non-coplanar volumetric-modulated arc therapy (VMAT), which is an advanced, intensity-modulated radiotherapy technique, using a dynamic swing arc (DSA) without moving the patient’s couch. This novel OXRAY system has the potential to reduce heart doses, particularly to the LAD, without reducing the irradiation dose to the target volume. Therefore, in this study, we evaluated the impact of the OXRAY system in improving the dose distribution, particularly in the heart and LAD regions, in VMAT planning for EC.
Patients and Methods
Patients. This was a retrospective treatment-planning study. Computed tomography (CT) datasets from 10 consecutive patients with EC who received VMAT at our Institution between July 2023 and July 2024, and whose heart was included in the irradiated field were analyzed (Table I). The virtual plans were created solely for research purposes and were not used for actual treatment delivery. The study methods, including the investigation procedure and handling of patient information, were approved by the Institutional Review Board of National Cancer Center Hospital East (IRB no. 2018-076).
Patient characteristics (n=10).
All patients were immobilized with their hands down, using a suction brace covering their heads, shoulders, arms, and waist. Simulation computed tomography (CT) datasets were acquired using Aquilion One (Canon Medical Systems, Tochigi, Japan) with an image slice thickness of 2 mm. Simulation CT scans were performed after patients had fasted for at least 4 h. All patients underwent four-dimensional CT scans to measure the respiratory motion of the targets.
Contouring of targets and organs at risk. All contouring procedures were performed by Radiation Oncologists and reviewed by other Radiation Oncologists. The gross tumor volume was defined as the primary tumor and metastatic lymph nodes. The primary tumor was identified on the basis of CT, [18F]fluoro-D-glucose positron-emission tomography/CT, with/without endoscopic findings. The clinical target volume (CTV) for the primary tumor was defined as the tumor volume plus 2 cm margins in the craniocaudal direction. The CTV for metastatic lymph nodes was defined as any metastatic lymph node volume. In addition, the internal target volume was defined as the target volume of the CTV plus the internal margin. The planning target volume (PTV) was defined as the overall internal target volume plus 0.5 cm margins in all directions.
Organs at risk (OAR) were defined as the entire heart, the LAD region (LADR), lungs, and spinal cord. The structure of the entire heart was contoured using the cardiac contouring atlas for radiotherapy with appropriate window settings [window width=500 Hounsfield units (HU); window level=50 HU] (14). Similarly, the structure of the LADR was contoured using its contouring atlas (15). The lung structure was contoured using pulmonary windows (window width=1,500 HU, window level=−600 HU), whereas the spinal cord structure was contoured based on the bony limits of the spinal canal.
Virtual planning and evaluation. Two virtual VMAT plans with involved-field irradiation targeting only visible lesions on imaging studies were delivered at a total of 60 Gy in 30 fractions to the EC region (D98 ≥98%, i.e., at least 98% of the PTV received 98% of the prescribed dose), which is the most commonly used dose for definitive CCRT for EC in Japan. The two virtual VMAT plans were as follows: (i) DSA-VMAT with OXRAY using two arcs with continuous modulating ring angles in 11 modulation points with a 6-MV photon beam; and (ii) conventional coplanar VAMT (Conv-VMAT) with Halcyon (Varian Medical Systems, Palo Alto, CA, USA) using two arcs and collimator angles of 345° and 15° with a 6-MV flattening filter-free photon beam. All plans were created by well-trained Radiation Oncologists and Medical Physicists.
The prescribed doses were calculated using the collapsed Cone Version 5.8 algorithm of our treatment planning system. The treatment plans were created using the RayStation® Planning System (RaySearch, Stockholm, Sweden). In addition, the following dose–volume parameters were evaluated: PTV: D2, D50, D95, conformity index (CI), homogeneity index (HI); LADR: mean dose, volume receiving 10 Gy (V10), V15, V30, D0.03(ml); heart: mean dose, V20, V30, V40, V50, D0.03(ml); lung: mean dose, V5, V10, V20, V30, V40; and spinal cord: D0.03(ml). Dose constraints for optimization are presented in Table II. Table III summarizes information regarding the devices used for treatment planning.
Dose constraints for the volumetric-modulated arc therapy planning. The LADR dose was optimized to be as low as possible without affecting the distribution to the other structures.
Summary of the information on the devices used for treatment planning.
Statistical analysis. Dose distributions are summarized as means and standard deviations (SD). Paired t-tests were used to evaluate the association between the two VMAT plans (DSA-VMAT and Conv-VMAT). Statistical significance was set at p<0.05. All statistical analyses were performed using SAS (version 9.4; Cary, NC, USA).
Results
Comparison of LADR dose–volume parameters between DSA-VMAT and Conv-VMAT. An example of the LADR dose distribution is shown in Figure 1.
Differences in dose distribution with respect to the left anterior descending coronary artery region (LADR) between dynamic swing arc (DSA-VMAT) and conventional (Conv-VMAT) volumetric-modulated arc therapy. (A) Computed tomography images displaying the dose distribution of radiotherapy to esophageal cancer (60 Gy in 30 fractions). The green line in the figure represents the LADR, and the red line represents the planning target volume (PTV). In Conv-VMAT, the LADR appears almost fully irradiated with 15 Gy (brown line), whereas in DSA-VMAT, less than half of the LADR appears to be irradiated with this dose. (B) Dose-volume parameters of the PTV, LADR, heart, and lungs. Red line: PTV; yellow line: heart; blue line: lungs; green line: LADR. Compared with the Conv-VMAT plan, the DSA-VMAT plan can improve most of the important parameters of the LADR, heart, and lungs while also improving PTV coverage.
A V15 of >10% for LADR was found in 3/10 of the DSA-VMAT plans and 7/10 of the Conv-VMAT plans. The mean V15 was 10.48±13.04% for DSA-VMAT and 30.28±23.56% for Conv-VMAT. Compared with Conv-VMAT, DSA-VMAT improved V15 (p=0.01; Table IV). However, the mean V30 was 0.14±0.43% for DSA-VMAT and 1.40±3.98% for Conv-VMAT. DSA-VMAT did not significantly affect LADR V30 compared with Conv-VMAT (p=0.29; Table IV).
Paired t-test comparing the two volumetric-modulated arc therapy (VMAT) plans used in this study – dynamic swing arc VMAT (DSA-VMAT) and conventional VMAT (Conv-VMAT).
Comparison of PTV dose–volume parameters between DSA-VMAT and Conv-VMAT. The mean D95, D50 and CI were 99.32±0.82%, 102.40±0.67%, and 0.92±0.03 for DSA-VMAT and 98.51±1.10%, 102.54±0.46%, and 0.92±0.02 for Conv-VMAT, respectively. Compared with Conv-VMAT, DSA-VMAT significantly improved D95 (p=0.01, Table IV). However, DSA-VMAT did not significantly affect D50 and CI compared with Conv-VMAT (p=0.62 and 0.91, respectively; Table IV).
Comparison of dose–volume parameters between DSA-VMAT and Conv-VMAT for other OARs. Considering the dose parameters for the whole heart, the mean and V40 were 8.64±5.37 Gy and 17.55±4.76% for DSA-VMAT and 11.23±7.37 Gy and 20.44±6.06% for Conv-VMAT, respectively. Compared with Conv-VMAT, DSA-VMAT improved the mean and V40 (p=0.01 and p<0.01, respectively; Table IV).
Regarding the lung, the mean values of V20 and V5 were 14.87±5.93 and 57.27±8.24% for DSA-VMAT and 17.81±7.70 and 61.15±9.97% for Conv-VMAT, respectively. Compared with Conv-VMAT, DSA-VMAT significantly improved in lung V20 and V5 (p<0.01 and p=0.01, respectively; Table IV).
Discussion
In this study, we found that DSA-VMAT remarkably reduced V15 for the LADR without negatively affecting the PTV dose distribution compared with Conv-VMAT, achieving V15 <10% in 70% of patients with EC. Whole-heart and lung dose–volume parameters were also significantly improved.
RIHD is an important late adverse event in patients with EC who receive thoracic radiotherapy (6-8). Recently, the LAD dose has been widely recognized as a factor that affects RIHD (9). In patients with EC with a poor prognosis, a high LADR dose due to the proximity of the heart and LAD to the target volume has been associated with a greater occurrence of RIHD than in patients with other malignancies, such as lung cancer, lymphoma, or breast cancer (6, 16-18). Therefore, Wang et al. proposed a V30 of <10% as an appropriate cut-off value for the LAD in patients with EC (6). This dose constraint was acceptable for both the DSA-VMAT and Conv-VMAT systems with optimization to avoid LADR overdose in our study.
In several cancer types with favorable prognosis, a dose constraint of V15<10% of the LAD has been proposed as an important indicator for low mortality (19-21). This dose constraint was almost acceptable for DSA-VMAT only with optimization to avoid LADR overdose. Furthermore, the most important aspect was that DSA-VMAT reduced the LADR dose without worsening the PTV coverage. The treatment planner can adjust the swing angle in DSA-VMAT for each gantry angle. For a PTV located near the heart and LAD, using oblique beams from the upper and lower sides of the heart can improve the heart and LADR doses without negatively affecting PTV coverage. This improvement in hazardous distribution suggests that DSA-VMAT may be highly valuable for EC treatment plans.
In the VMAT era, the dose constraint for the whole heart can be safely achieved in both DSA-VMAT and Conv-VMAT plans in many treatment plannings. Therefore, the achievement of the LADR dose constraint has received more attention for RIHD as a risk factor (19-21). Even when considering the lungs, the dose constraint was safely achieved in both the DSA-VMAT and Conv-VMAT plans. The greatest advantage of DSA-VMAT is the reduction in LADR dose, which is impossible to achieve with the Conv-VMAT system. This DSA-VMAT system may become a game changer in VMAT radiotherapy for patients with EC, whose prognosis is improving.
Study limitations. Firstly, the sample size was small. Secondly, the planner’s skills in optimizing the treatment plan may have affected the quality of the treatment plan. Given that DSA-VMAT can select almost all omnidirectional beams, the most favorable beams may not have been selected. However, this suggests that there is potential for improvement in dose distribution in DSA-VMAT planning. Thirdly, owing to the limited quality of CT images due to free breathing, lack of cardiac synchronization, and absence of intravenous contrast medium administration, the LAD may have been inadequately evaluated. Therefore, we contoured the LADR as accurately as possible and assessed each plan accordingly. Because many improvements in imaging techniques have recently been reported, more accurate contouring of the LAD may be possible in practice in the future (23, 24). Finally, the use of DSA-VMAT with the OXRAY system was limited to cases with a field size <20 cm. Therefore, these results cannot be applied in clinical practice to all patients with EC. Despite these limitations, this study was markedly useful in verifying the effectiveness of DSA-VMAT with the OXRAY system in reducing the LADR dose in patients with EC, in whom LADR overdose has recently become an important issue.
Conclusion
DSA-VMAT with the OXRAY system dramatically improved the LADR dose without worsening the PTV coverage or other OAR dose distributions. Therefore, the clinical effect of DSA-VMAT in reducing RIHD should be evaluated in real-world clinical practice.
Acknowledgements
The Authors wish to express their sincere gratitude to Yuichi Nagai, the General Manager of the Radiation Technology Department, for supporting the research facilities and providing the environment for conducting this research. We express our deep gratitude to Takaki Ariji and Hajime Oyoshi, Chief of Radiation Technology, for sharing their knowledge and skills and providing guidance on the treatment planning method.
Footnotes
Authors’ Contributions
Concept and design: KM. Treatment planning creation data analysis and interpretation: KH, KM, and MW. Statistical analysis: MW. Important advice and critical discussion: KM, KT, and SK. Review of dose distributions (radiation oncologists): HH, KT and KM. Research management and supervision: MI and TS. All Authors read and approved the final version of the manuscript.
Conflicts of Interest
Kento Tomizawa received honorarium from Hitachi, Ltd (Tokyo, Japan). The other Authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Funding
This study was partly supported by JSPS KAKENHI Grant Number JP21K07661.
Artificial Intelligence (AI) Disclosure
During the preparation of this manuscript, a large language model (ChatGPT, OpenAI) was used solely for language editing and stylistic improvements in select paragraphs. No sections involving the generation, analysis, or interpretation of research data were produced by generative AI. All scientific content was created and verified by the Authors. Furthermore, no figures or visual data were generated or modified using generative AI or machine-learning-based image enhancement tools.
- Received May 9, 2025.
- Revision received May 26, 2025.
- Accepted May 28, 2025.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
