Dosimetric risk factors for radiation esophagitis in patients with breast cancer following regional nodal radiation

Abstract

BACKGROUND

Radiation esophagitis (RE) is one of the most common clinical symptoms of regi-onal lymph node radiotherapy for breast cancer. However, there are fewer studies focusing on RE caused by hypofractionated radiotherapy (HFRT).

AIM

To analyze the clinical and dosimetric factors that contribute to the development of RE in patients with breast cancer treated with HFRT of regional lymph nodes.

METHODS

Between January and December 2022, we retrospectively analysed 64 patients with breast cancer who met our inclusion criteria underwent regional nodal intensity-modulated radiotherapy at a radiotherapy dose of 43.5 Gy/15F.

RESULTS

Of the 64 patients in this study, 24 (37.5%) did not develop RE, 29 (45.3%) developed grade 1 RE (G1RE), 11 (17.2%) developed grade 2 RE (G2RE), and none developed grade 3 RE or higher. Our univariable logistic regression analysis found G2RE to be significantly correlated with the maximum dose, mean dose, relative volume 20-40, and absolute volume (AV) 20-40. Our stepwise linear regression analyses found AV30 and AV35 to be significantly associated with G2RE (P < 0.001). The optimal threshold for AV30 was 2.39 mL [area under the curve (AUC): 0.996; sensitivity: 90.9%; specificity: 91.1%]. The optimal threshold for AV35 was 0.71 mL (AUC: 0.932; sensitivity: 90.9%; specificity: 83.9%).

CONCLUSION

AV30 and AV35 were significantly associated with G2RE. The thresholds for AV30 and AV35 should be limited to 2.39 mL and 0.71 mL, respectively.

Key Words: Breast cancer, Dosimetric parameters, Radiation esophagitis, Esophagitis, Hypofractionated radiotherapy

Core Tip: Radiation esophagitis (RE) is a common clinical symptom among patients with breast cancer receiving regional lymph node radiotherapy. RE commonly affects patients’ quality of life and can even lead to termination of treatment. Hypofractionated radiation therapy (HFRT) is becoming a common treatment for patients with breast cancer. However, there are few studies on the occurrence of RE in patients with breast cancer undergoing HFRT. This study aimed to provide clinical guidance by analyzing the clinical and dosimetric factors contributing to the development of RE in patients with breast cancer treated with HFRT to the regional lymph nodes.

INTRODUCTION

The risk of recurrence in breast cancer with regional lymph nodes is reduced by 33% by postoperative radiotherapy[1]. Hypofractionated radiotherapy (HFRT) is as safe and effective as conventional fractionated radiotherapy (CFRT)[2-4]. In both patients who have undergone breast-conserving surgery and those who have undergone a mastectomy, HFRT has been utilized extensively for breast cancer over the past few decades[4,5]. The incidence rates of acute radiation reactions, including radiodermatitis and radiation pneumonitis, indicate that HFRT is as or less toxic than CFRT[3,6]. However, there have been few studies on the occurrence of radiation esophagitis (RE) in patients with breast cancer who have received HFRT.

Irradiation of regional lymph nodes such as the axillary, supraclavicular, and internal mammary lymph nodes is known to increase the risk of RE[7]. The pain and difficulty eating caused by RE can significantly affect a patient’s quality of life. In this study, we investigated RE in patients with breast cancer receiving hypofractionated intensity-modulated radiotherapy (IMRT). We aim to provide clinical guidance on the optimal regional nodal radiation dosimetric parameters for the avoidance of RE in breast cancer patients.

MATERIALS AND METHODS

Patients selection

This study retrospectively analysed 64 breast cancer patients who received regional lymph node adjuvant radiotherapy at our institution between January and December 2022 and met the inclusion criteria. These criteria were as follows: ≥ 18 years of age; receipt of mastectomy or breast conservation surgery and sentinel lymph node biopsy/axillary dissection; negative surgical margins; absence of supraclavicular and internal breast lymph node metastases. All of the patients signed a consent form for HFRT treatment and approved by the Medical Ethics Committee of the Affiliated Hospital of Hebei University of Engineering (2021[K]088). The study was conducted in accordance with the tenets of the 2013 revision of the Declaration of Helsinki.

Treatment and dosimetric parameters

All patients received radiotherapy within 8 wk of surgery (without planned adjuvant chemotherapy) or after completion of adjuvant chemotherapy. Computed tomography analog scanning was used for positioning and body thermoplastic wrap was used to fix the position. The planning system was used for the digital transmission of images. In patients who had undergone a mastectomy, a 5-mm-thick silicone pad was placed on the surface of the chest wall for the first six irradiations. The irradiated area included the entire breast/chest wall, the supra/infraclavicular fossa, and the group II axillary lymph nodes.

Per the Radiation Therapy Oncology Collaborative Group (RTOG) guidelines, we outlined the esophageal area from the lower edge of the cricoid cartilage to the lower edge of the aortic arch (Figure 1). Hypofractionated IMRT was administered at a dose of 45 Gy/15 F (2.9 Gy/F, 1/d, 5 F/w). The guidelines stipulate that the prescribed dose should cover 95%-110% of the planned target volume. Dose-volume histograms were used to determine the dosimetric parameters for the esophagus, including the mean dose (Dmean), maximum dose (Dmax), total volume, relative volume, and absolute volume (AV) for the esophagus treated with 10-40 Gy (RV10–RV40 and AV10–AV40). Our institution limits the irradiated dose to the esophagus to a Dmax < 30 Gy for breast cancer patients treated with regional lymph node radiotherapy as we have found an increased probability of RE when the dose exceeds this range. Therefore, we investigated the V30-V40 interval in more detail. The concurrent use of targeted therapy and radiotherapy was permitted if the patient had a HER2 overexpression molecular subtype, but concurrent chemotherapy or hormonal therapy was not permitted.

Figure 1 The esophageal area from the lower edge of the cricoid cartilage to the lower edge of the aortic arch. A and B: The esophagus is outlined in the front and side view; C: An illustration of the supraclavicular nodal outlined and dose distribution; D: An illustration of the chest wall and axillary nodal outlined and dose distribution after mastectomy; E: An illustration of the breast/tumor bed outlined and dose distribution after breast-conserving surgery. The colors corresponding to the prescription dose from low to high are blue, green, yellow, and red.

Evaluation

The presence and severity of RE were evaluated using the acute RE grading criteria of the RTOG. Evaluations were conducted during radiotherapy and at 1 wk, 2 wk, and 3 months follow-ups after the end of the treatment course. This study was analysed for grade 2 RE (G2RE) since it was the highest level observed.

Statistics analysis

Our data analyses were based on the G2RE. Univariate logistic regression analyses were conducted first to identify the clinical variables and dosimetric indicators associated with G2RE. Multivariate logistic regression analysis was then used to identify the independent clinical variables from the univariate logistic regression with significance levels P < 0.05. Due to the covariance between RE dosimetry parameters, stepwise linear regression analysis was finally applied in this study. Receiver operating characteristics (ROC) analysis was combined with the optimal Youden index to identify the esophageal dosimetry parameter thresholds associated with RE risk in the stepwise linear regression analyses. P < 0.05 were considered statistically significant. All analyses were performed using SPSS for Windows, v. 26.0 (IBM Corp., Armonk, NY, United States).

RESULTS

This study comprised 64 patients with breast cancer who underwent postoperative regional nodal radiation. All patients completed the full course of their radiotherapy. The clinical characteristics of the patients are shown in Table 1. Of the 64 patients, 24 (37.5%) did not develop RE, 29 patients (45.3%) developed grade 1 RE (G1RE), 11 (17.2%) developed G2RE, and none of the patients developed grade 3 RE (G3RE) or higher. In all RE patients, G1RE became apparent after the fifth radiation treatment, while G2RE appeared after the seventh radiation treatment. After 10-15 radiation treatments, G2RE reached its peak and then disappeared within 30 d (Figure 2). Univariable logistic regression analyses of the clinical characteristics of those participants with G2RE found no correlations (Table 2).

Figure 2 Proportion of patients with radiation esophagitis over time. G0: No radiation oesophagitis; G1: Grade 1 radiation oesophagitis; G2: Grade 2 radiation oesophagitis.

Table 1 Patient characteristics.

Variable	G0	G2	χ2	P value
Age
< 60	41	7	0.92	0.57
≥ 60	12	4
Laterality
Left	28	5	0.20	0.66
Right	25	6
Type of surgery
Mastectomy	40	6	1.97	0.16
Lumpectomy	13	5
Histologic type
Ductal invasive carcinoma	45	11	1.90	0.59
Lobular invasive carcinoma	3	0
Mucinous carcinoma	2	0
Others	3	0
Stage group
Ⅱ A	31	6	2.12	0.35
Ⅱ B	10	4
Ⅲ	12	1
Molecular subtype
Luminal A	18	2	6.389	0.172
Luminal B	8	0
Luminal HER2-positive	9	1
HER2 overpressing	6	2
Triple-negative	12	6
Chemotherapy
Neoadjuvant chemotherapy	19	5	0.69	0.71
Adjuvant chemotherapy	32	6
No chemotherapy	2	0
Anti-HER2 targeted therapy
No	37	9	0.65	0.66
Yes	15	3

G0: No radiation oesophagitis; G2: Grade 2 radiation oesophagitis; HER2: Human epidermal growth factor receptor 2.

Table 2 Univariable logistic regression analyses of clinical characteristics and grade 2 radiation oesophagitis.

Variable	OR (95%CI)	P value
Age
< 60	1.00	0.34
≥ 60	1.95 (0.49-7.81)
Laterality
Left	1.00
Right	1.34 (0.37-4.95)	0.66
Surgery
Lumpectomy	1.00
Mastectomy	0.39 (0.10-1.49)	0.17
Histologic type
Ductal invasive carcinoma	1.00
Others	0.00 (0.00-)	1.0
Stage
Ⅱ	1.00
Ⅲ	0.34 (0.04-2.94)	0.33
Anti-HER2 targeted therapy
Yes	1.00
No	1.05 (0.25-4.51)	0.95
Chemotherapy
Yes	1.00
No	0.00(0.00- )	1.0

OR: Odds ratio; CI: Confidence interval; HER2: Human epidermal growth factor receptor 2.

The data for the dosimetric parameters are presented in Table 3. Our univariable logistic regression analyses found G2RE to be significantly correlated with Dmax, Dmean, RV20-40, and AV20-40 (P <0.05). Our stepwise linear regression analyses found AV30 and AV35 to be significantly correlated with G2RE. We also found significant covariance between RV30 and RV35 (VIF = 14.51 and 36.97, respectively). The optimal thresholds for these statistically significant dosimetric parameters were calculated using ROC analysis. The optimal threshold for AV30 was found to be 2.39 mL (P < 0.001), with an area under the curve (AUC) of 0.996, sensitivity of 90.9%, and specificity of 91.1%; and that for AV35 was 0.71 mL, (P < 0.001), with an AUC of 0.932, sensitivity of 90.9%, and specificity of 83.9% (Table 4).

Table 3 Degree of variability of dosimetric parameters in patients.

Variable	G0			≥ G2
	Mean	Range	Std Dev	Mean	Range	Std Dev
Total volume	8.59	(6.01-11.75)	1.28	8.90	(6.65-11.19)	1.40
Dmax	39.60	(34.35-46.38)	3.24	42.78	(39.40-46.81)	2.69
Dmean	16.11	(10.5-25.28)	3.02	19.44	(14.30-22.49)	2.71
RV V10 (%)	59.35	(28.83-87.50)	16.41	66.45	(37.87-95.19)	17.45
RV V20 (%)	37.62	(19.45-71.20)	10.09	45.27	(31.06-54.15)	7.26
RV V30 (%)	16.45	(5.10-38.69)	8.81	36.45	(26.28-44.83)	5.81
RV V35 (%)	4.59	(0.00-27.09)	5.93	17.11	(7.01-27.95)	7.91
RV V40 (%)	0.58	(0.00-9.57)	1.57	3.02	(0.00-11.55)	4.36
AV V10 (mL)	5.10	(2.34-8.97)	1.48	5.73	(3.54-7.34)	1.34
AV V20 (mL)	3.23	(1.42-6.55)	0.96	4.03	(2.51-5.48)	0.91
AV V30 (mL)	1.39	(0.08-2.79)	0.71	3.25	(2.01-4.73)	1.34
AV V35 (mL)	0.37	(0.00-2.03)	0.47	1.49	(0.61-2.33)	0.67
AV V40 (mL)	0.05	(0.00-0.88)	0.14	0.27	(0.00-1.03)	0.40

G0: No radiation oesophagitis; G2: Grade 2 radiation oesophagitis; RV: Relative volume; AV: Absolute volume.

Table 4 Comparisons of esophageal dosimetric parameters between patients without and with grade 2 radiation oesophagitis.

Variable	Univariable logistic regression		Stepwise linear regression
	OR (95%CI)	P value	Adjust R2	Std. error	F	P value
Total volume	1.20 (0.72-2.00)	0.48
Dmean	1.44 (1.12-1.85)	< 0.01
Dmax	1.38 (1.09-1.76)	0.01
RV V10	1.02 (0.98-1.06)	0.29
RV V20	1.08 (1.01-1.16)	0.03
RV V30	1.36 (1.13-1.63)	< 0.01
RV V35	1.22 (1.10-1.36)	< 0.01
RV V40	1.36 (1.06-1.75)	0.01
AV V10	1.38 (0.89-2.13)	0.16
AV V20	2.30 (1.14-4.64)	0.02
AV V30	2.97 (1.88-5.69)	< 0.01	0.48	0.48	59.1	< 0.01
AV V35	2.43 (1.45-3.96)	< 0.01	0.51	0.76	33.9	0.03
AV V40	2.52 (1.96-3.60)	0.01

OR: Odds ratio; CI: Confidence interval; RV: Relative volume; AV: Absolute volume.

DISCUSSION

RE is among the most common radiotoxic responses to regional nodal irradiation for breast cancer. Although dosimetric analyses of the incidence of RE in patients with lung cancer have been conducted[8-10], there have been few studies addressing HFRT-induced RE in breast cancer patients. In this study, we found that the dosimetric parameters of the esophagus are closely related to the incidence of RE and should be limited to AV30 < 2.39 mL and AV35 < 0.71 mL.

Among our breast cancer cohort, the incidence of G1RE was 45.3% and that of G2RE was 17.2%. There were no instances of higher grade RE. The prevalence of G2RE has been found to range from 7.4% to 44.8%, whereas that of G3RE ranges from 0% to 1.5%[11-20]. In a prospective study that delivered 43.5 Gy/15 F of radiation to regional lymph nodes, there was a 40.9% incidence of G2RE and a 0.3% incidence of G3RE[20]. In the same study, radiation doses of 40.5 Gy/15 F to the supraclavicular nodes resulted in a 7.4% incidence of G2RE and a 3.7% incidence of G3RE[15].

In the present study, we found the highest prevalence of RE in the final week of radiotherapy and the subsequent week. At this time point, the esophagus had been exposed to 20.3 Gy of radiation. Other studies of HFRT therapy have observed similar timing of RE onset[15,20]. Traditionally, RE peaks one week after the end of a course of CFRT[11] while HFRT accelerates RE onset. Using these data can allow physicians to predict when RE treatment is likely to be needed.

We found no correlations between RE and the clinical characteristics of our patients. In a prospective study of breast cancer patients receiving regional lymph node irradiation at a dose of 43.5 Gy/15 F, left-sided breast cancer patients were more likely to develop G2RE[20]. However, another study of breast cancer patients receiving regional lymph node radiotherapy, but at a dose of 50 Gy/25 F, found the incidence of G2RE to be comparable for left-sided and right-sided cancer[13]. This is likely because the esophagus is located on the left side, so when cancer affects the left breast that side is more susceptible to radiation exposure. Ahn et al[21] studied radiotherapy of regional lymph nodes in lung cancer patients (Dmean: 66 Gy) and found advanced age to be a risk factor for RE. The small number of patients in our study and their relative youth (75% were younger than 60 years) was something of a limitation and should be addressed in future research. It should be noted that the upper esophagus is more sensitive to radiation due to differences in sensory nerve axon distribution and physiology[22,23]. There is a significant increase in the probability of G2RE when the pharynx is ≥ 1 cm[11]. Therefore, the variables that contribute to RE risk may be multifactorial. Overall, our results and the findings of previous studies strongly suggest that the dose of irradiation to the esophagus should be a matter of some concern.

Our univariable logistic regression analyses found that G2RE was significantly associated with Dmax, Dmean, RV20-40, and AV20-40. When we analyzed these dosimetric parameters using multivariate logistic regression analyses, we found inconsistent significance levels for Dmax and Dmean. Moreover, there was a strong covariance between RV30 and RV35 (VIF = 14.51 and 36.97, respectively). To address this discrepancy, we applied stepwise linear regression analyses to our data. Based on our findings, we concluded that limiting AV30 and AV35 to > 2.39 mL and > 0.71 mL, respectively would reduce the incidence of G2RE in this patient population. A multivariate logistic regression analysis of Dmean, Dmax, and V5-V35 by Yaney et al[13] identified Dmean as the factor most strongly associated with RE. This is supported by multiple studies that have reported a significant correlation between Dmean and RE[11-13,20,24,25]. West et al[11] found that a Dmean > 31 Gy led to a three times higher risk of ≥ G2R. In the present study, the maximum Dmean value was under 31 Gy (range: 14.30-22.49 Gy), which may explain the inconsistent significance of the correlation between Dmean and RE.

A study of patients with breast cancer treated with CFRT (50 Gy/25 F) who had undergone breast-conserving or radical treatment, as well as those with recurrent breast cancer found Dmax, Dmean, and V10-V50 to be associated with the development of G2RE[13]. However, Dzul et al[12] found only Dmean to be significantly associated with RE. The reason for these conflicting results may be that only 54% of the patients in the Dzul et al[12] study underwent CFRT, while the rest were treated with HFRT. Amin et al[15] treated breast cancer patients requiring supraclavicular regional lymph node irradiation with HFRT (40.5 Gy/15 F) and found Dmean, V10, and V20 to be significantly associated with ≥ G1RE. However, these were the only three indicators analyzed. Wang et al[20] used a radiation dose of 43.5 Gy/15 F and found Dmax, Dmean, RV20-RV40, and AV35-AV40 to be significantly associated with G2RE. While patients in the present study were treated with IMRT only, Wang et al[20] used different radiotherapy techniques with different patients, including hybrid volumetric modulated arc therapy, IMRT, tomography, and electrons. IMRT has been found to carry a higher risk of RE than three-dimensional conformal radiotherapy in patients receiving regional lymph node irradiation for breast cancer[13,14]. In line with the study by Wang et al[20], we placed only one parameter at a time in the multivariate model to avoid multicollinearity between the esophageal dosimetry parameters. However, we also utilized stepwise linear regression to avoid covariance. This may account for some of the differences between the two studies. However, there has been little research on this topic.

HFRT is becoming a more common treatment for patients with breast cancer. In the present study, we included patients with regional nodal breast cancer who had undergone either conservation surgery or mastectomy and were treated with HFRT. All of our patients were treated with IMRT to avoid inconsistent results due to different radiotherapy techniques. This approach also provided evidence of the dosimetric indicators associated with G2RE. However, the study had some limitations. First, this was a retrospective study, so the completeness and homogeneity of the data cannot be guaranteed. In the future, we intend to conduct a prospective study with carefully controlled experimental conditions. Second, our sample size was relatively small and the patients were all treated at one institution. A multi-center clinical study with a larger number of patients is needed to verify our findings. Given the results of the present study, we recommend focusing on AV30 and AV35 in clinical practice. This will help to reduce the toxicity of the treatment and improve patient compliance. In cases where AV30 and AV35 exceed 2.39 mL and 0.71 mL, oral mucosal protectants, anti-inflammatory, and analgesic drugs may be given to reduce the risk of G2RE or higher.

CONCLUSION

In conclusion, AV30 and AV35 are the most sensitive predictors of G2RE. Limiting AV30 to < 2.39 mL and AV35 to < 0.71 mL may reduce the incidence of RE in breast cancer patients following regional lymph node radiotherapy, thereby enhancing their quality of life.