A Multicenter Quality Audit Standardizing Esophageal Contouring in Postmastectomy Radiotherapy

Background

Postmastectomy radiation therapy (PMRT) including comprehensive regional nodal irradiation in selected patients is an integral component of breast cancer management. The use of volumetric modulated arc therapy (VMAT) and immune-mediated radiotherapy (IMRT) techniques in PMRT planning has become more common because these methods deliver better conformity than and improved coverage over 3D-conformal radiation therapy (CRT). Some studies have shown IMRT/VMAT delivers higher average esophageal doses compared with 3D techniques.^1,2 This is because of the use of gantry tilt and multileaf collimator blocks used in 3D techniques to avoid the esophagus. In the study by Yaney et al, patients receiving IMRT/VMAT developed grade 2 esophagitis at a rate 3 times higher than patients treated with 3D-CRT (23.6% vs 10.9%), while identifying 11 Gy as the critical dose threshold for increased toxicity risk.²

In this study, the esophagus was contoured from the inferior border of the cricoid cartilage down to the level of the carina or aortic arch and not from cricoid to gastroesophageal junction to define dosimetric constraints. Our monthly internal case-based review found that most practitioners routinely used auto-segmented, full-length esophageal contours as the reference structure during treatment planning. This practice can underestimate the dose delivered to the clinically relevant upper esophageal segment encompassed by regional nodal irradiation fields, which is most closely associated with the risk of radiation-induced esophagitis. In response to this observation, the present audit was conducted to first quantify the dosimetric differences between whole-esophagus contouring and standardized upper esophageal (cricoid-to-carina) contouring approaches. Following implementation of the revised contouring standard operating procedure (SOP), a separate postaudit clinical cohort was subsequently evaluated to assess the incidence of clinically observed esophagitis. This audit was conducted to evaluate the incidence of esophagitis and the dosimetric difference between contouring the entire esophagus vs the upper esophagus as the denominator.

Methods and Materials

This study was a multicenter retrospective quality-improvement audit across 8 radiation oncology centers in a nationwide network. The audit was initiated following an internal educational forum that identified a discrepancy between published esophageal dose constraints and the lower mean doses routinely reported using the network’s auto-segmented full-length esophagus contours. The audit included the 43 most recent consecutive patients who underwent postmastectomy radiation therapy with regional nodal irradiation using VMAT between March 10 and May 5, 2025. All participating sites contributed anonymized treatment plans and dosimetric data; institutional ethics committee policies permitted retrospective plan review.

All patients had previously undergone mastectomy and received VMAT-based PMRT covering the chest wall and supraclavicular nodal regions, with or without internal mammary node irradiation. Only cases with complete plan data and structure sets were included. No restrictions were imposed on patient age, tumor laterality, or stage, reflecting the real-world design of the audit. Each center used its standard auto-segmentation workflow, which delineates the esophagus in its entirety, from the cricoid cartilage to the gastroesophageal junction. For the purpose of this audit, all sites were instructed to recontour the esophagus from the cricoid cartilage to the tracheal bifurcation (carina), consistent with the anatomical definitions used in recent dose-toxicity studies.^1,2 Both the original auto-contours and the recontoured structures were retained for comparison, and no other organs at risk or target structures were altered. Treatment planning was performed using institutional VMAT protocols with inverse optimization. Prescriptions ranged from 43.5 Gy in 15 fractions to 50 Gy in 25 fractions. Prior to this audit, no uniform esophageal dose constraint existed across the network.

For each patient, dose-volume histogram (DVH) parameters were collected for both contour sets, including the esophageal mean dose (D_mean), maximum dose (D_max), and V20 (volume of organ receiving ≥20 Gy). The published mean-dose threshold of less than 11 Gy was used as the clinically relevant benchmark. One center additionally reviewed clinical records from a separate cohort of 7 patients, treated after implementation of the standardized contouring SOP to document the incidence of grade 2 esophagitis, enabling a preliminary dose-toxicity correlation within a single institutional cohort. The primary variable of interest was the change in D_mean after recontouring. Secondary variables included changes in D_max and V20, the proportion of patients whose D_mean exceeded 11 Gy, and changes in rates of clinically observed grade 2 esophagitis in the separate post-SOP clinical cohort. All DVH data were collated centrally and independently reviewed by 2 radiation oncologists for accuracy. Statistical analysis was performed using the Wilcoxon signed-rank test for paired dosimetric comparisons and Fisher’s exact test for categorical clinical outcomes. A P value less than .05 was considered statistically significant. To ensure reproducibility, a random sample of recontoured esophagus structures underwent dual-review quality assurance, with discrepancies resolved by consensus.

Results

Forty-three VMAT-based PMRT plans from 8 centers were reviewed. Complete dosimetric data were available for 42 cases in the mean-dose analysis, due to 1 incomplete structure set in which D_mean could not be reliably recalculated, while D_max and V20 were available for all 43 patients. Except for 1 center, all sites used full-length auto-segmented esophageal contours in their routine workflow. A summary of case selection is shown in the Figure, including the primary 43-patient dosimetric audit cohort and a separate clinical pre-SOP/post-SOP cohort (n = 7 each).

Recontouring the esophagus from the cricoid cartilage to the carina consistently increased the reported esophageal dose across the network. The median D_mean rose from 5.7 Gy with full-length auto-contours to 9.6 Gy with standardized contours (P < .001). The corresponding networkwide mean increased from 6.1 Gy (IQR, 4.7-7.9) to 11.7 Gy (IQR, 8.6-15.1). This difference arose because the patient-level median D_mean (5.7 Gy vs 9.6 Gy) reflects the midpoint of individual values, whereas the network work D_mean (6.1 Gy vs 11.7 Gy) is the overall average across all patients. Reporting both offers insight into individual patient experiences and broader systemic impact. Using standardized contours, 16 of 42 evaluable patients (38.1%) exceeded the clinically relevant 11-Gy D_mean threshold. D_max increased from 27.4 Gy to 30.8 Gy (P = .047), and V20 increased from 5.1% to 9.4% (P = .031). These results are summarized in Table 1.

Clinical correlation from a pilot center included 14 patients (7 before and 7 after SOP implementation). Prior to the contouring change, 6 of 7 (85.7%) patients developed grade 2 esophagitis. No cases occurred after SOP adoption (P = .004). Pre-SOP symptomatic patients had recalculated mean esophageal doses of 12 Gy to 13 Gy based on the modified cricoid-to-carina contours, whereas post-SOP patients had doses of 7 Gy to 10 Gy. These findings are presented in Table 2.

All analyses followed the predefined statistical plan, using Wilcoxon signed rank testing for dosimetric comparisons and Fisher exact test for categorical outcomes (significance level P < .05). No outliers were detected, and quality-assurance review confirmed uniform use of the cricoid-carina contour definition.

Discussion

PMRT now routinely includes comprehensive chest wall and regional nodal irradiation, and modern techniques such as VMAT/IMRT are increasingly used to achieve dose conformity and organ- at-risk sparing.^3,4 The highly conformal radiation plans that achieve precise dose delivery to targets also distribute low to moderate doses throughout adjacent normal tissues. The esophagus, while not a target, lies within the neck and mediastinum, traversed by supraclavicular and internal mammary fields, and can receive an appreciable dose. The esophagus receives minimal radiation exposure from breast-only fields, according to a systematic review, with doses averaging below 2 Gy, but regional node inclusion increases the D_mean to 11.4 Gy.² The exposure level of 11 Gy in the esophagus creates a significant risk for developing radiation-induced esophageal cancer, according to Duane et al.¹ The incidence of breast radiotherapy–related esophagitis remains underreported, but studies show that nodal field treatment leads to significant symptoms. The study by Yaney et al showed that 16% of patients who received breast regional nodal irradiation (RNI) developed grade 2 or higher esophagitis, while IMRT/VMAT plans resulted in 23.6% esophagitis compared with 10.9% with 3D-CRT plans.² The authors established 11 Gy as the critical dose threshold for esophageal toxicity, which led them to propose D_mean and volume constraints for RNI planning.^2,3

The recent PMRT guidelines emphasize the need for precise target and organs-at-risk definition during treatment planning. The 2025 ASTRO/ASCO/SSO guideline requires CT-based 3D conformal planning for the chest wall and nodes but allows IMRT/VMAT only when standard geometry fails to deliver proper coverage.⁵ The recommendations require health care providers to establish standardized organ definitions for the esophagus and establish suitable safety limits. The default practice in many clinics involves using auto-segmentation to define the entire esophagus from its cervical to gastroesophageal junction. The practice of using full esophagus auto-segmentation produces incorrect D_mean results because it includes nonirradiated segments, which reduce the calculated D_mean and conceal the actual dose received by the midthoracic esophagus in treatment fields.

The network audit demonstrated that all participating centers routinely used auto-segmented full-length esophageal contours as the reference structure for dose reporting, resulting in reported mean esophageal doses of approximately 6 to 7 Gy, which fell below the commonly cited 11-Gy threshold.^3,6 The D_mean increased from 11 Gy to 12 Gy when the cricoid-to-carina contouring method was used for the Yaney et al study participants who exceeded the recommended constraint threshold. Our network conducted a quality assessment across 8 facilities to address the observed discrepancy. The audit revealed that the previous contouring method produced incorrect esophageal dose measurements, which led to a networkwide update of contouring SOPs. The implementation of the new standard at 1 center resulted in a significant reduction of symptomatic esophagitis according to preliminary clinical data. We highlight the practical impact of standardized esophagus contouring and reinforce the need to apply modern dosimetric constraints in PMRT planning, thereby minimizing avoidable toxicity.

Corresponding Author

Sushil Beriwal, MD

Email: sushilberiwal@gmail.com

Phone: +94-12370-3283

References

1. Duane FK, Kerr A, Wang Z, et al. Exposure of the oesophagus in breast cancer radiotherapy: a systematic review of oesophagus doses published 2010-2020. Radiother Oncol. 2021; 164:261-267. doi: 10.1016/j.radonc.2021.09.032
2. Yaney A, Ayan AS, Pan X, et al. Dosimetric parameters associated with radiation-induced esophagitis in breast cancer patients undergoing regional nodal irradiation. Radiother Oncol. 2021; 155:167-173. doi: 10.1016/j.radonc.2020.10.042
3. Wang DQ, Zhang N, Dong LH, et al. Dose-volume predictors for radiation esophagitis in patients with breast cancer undergoing hypofractionated regional nodal radiation therapy. Int J Radiat Oncol Biol Phys. 2023;117(1):186-197. doi: 10.1016/j.ijrobp.2023.03.060
4. Alsaihaty Z, Manan HA, Sabarudin A, et al. Hybrid Treatment Planning for Chest Wall Irradiation Utilizing Three-Dimensional Conformal Radiotherapy (3DCRT), Intensity-Modulated Radiation Therapy (IMRT), and Volumetric Modulated Arc Therapy (VMAT): A Systematic Review. Cureus. 2024 May 3;16(5): e59583. Doi: 10.7759.
5. Jimenez RB, Abdou Y, Anderson P, et al. Postmastectomy radiation therapy: an ASTRO-ASCO-SSO Clinical Practice Guideline. J Clin Oncol. 2025;43(30):3292-3311. doi:10.1200/JCO-25-01747
6. Wang DQ, Yan L, Wu HF, et al. Prospective cohort study to validate esophageal dose constraints and predictive models for esophagitis in patients with breast cancer undergoing hypofractionated regional nodal radiotherapy. Radiother Oncol. 2025; 212:111120. doi: 10.1016/j.radonc.2025.111120