Introduction

Major advances in surgery, radiotherapy (RT), and chemotherapy have established multimodal curative treatment options for oesophageal cancer (EC)1. Definitive concurrent chemoradiotherapy (CCRT) is the preferred treatment for inoperable patients, cases with unresectable EC, or patients who decline surgery2. However, the locoregional failure rate remains high (41–50%) after definitive CCRT using a radiation dose of 50.4 Gy for EC, with the vast majority of failures involving the primary tumour (86–90%)3,4. Thus, the standard radiation dose (50–50.4 Gy)1 may be suboptimal, and some studies have suggested that a higher RT dose to the primary tumour might improve local control and survival5,6.

The standard radiation dose for definitive CCRT is based on the INT-0123 trial, which compared high and standard doses (64.8 Gy vs. 50.4 Gy) for locally advanced EC7, and was prematurely closed because of multiple deaths in the high-dose group and no differences in survival or local control. However, 7 of the 11 treatment-related deaths in the high-dose group occurred before the radiation dose reached 50.4 Gy, and the high-dose group also received a significantly lower dose of fluorouracil, which might have affected the outcomes in that group. Moreover, the investigators used a conventional two-dimensional technique to deliver RT, rather than relatively safer modern RT techniques, such as three-dimensional conformal radiotherapy (3DCRT), intensity-modulated radiotherapy (IMRT), or volumetric modulated arc therapy (VMAT). Several dosimetric studies have explored the use of IMRT or VMAT to maintain the therapeutic ratio by administering a high RT dose to the tumour while minimizing the dose to the surrounding organs8,9,10,11. This single-centre retrospective study aimed to evaluate the clinical benefit of high-dose RT using modern techniques for treating EC.

Results

Table 1 shows the baseline characteristics of the groups that received an RT dose of <66 Gy (n = 44; “the <66 Gy group”) or ≥66 Gy (n = 71; “the ≥66 Gy group”). The median patient age at diagnosis was 61 years (range: 32–86 years). The patients predominantly were male (93%), and had squamous cell carcinoma (95%). The median tumour length was 7 cm (range: 2–20 cm). The main reason for not undergoing surgery was an unresectable tumour (59%). Eighty-five patients (74%) received maintenance chemotherapy. The RT techniques included VMAT (49 patients, 43%), IMRT (43 patients, 37%), and 3DCRT (23 patients, 20%).

Table 1 Comparing the patients’ clinical characteristics.
Full size table

The median ages were 62 years (range: 32–85 years) for the <66 Gy group and 60 years (range: 39–86 years) for the ≥66 Gy group (p = 0.523). The ≥66 Gy group had a significantly higher proportion of patients with a good performance status (Eastern Cooperative Oncology Group performance status [ECOG PS] of 0) (p = 0.027). There were no significant differences between the <66 Gy and ≥66 Gy groups in terms of their sex, tumour histology, grade, location, length, clinical T stage, clinical N stage, reasons for no surgery, chemotherapy regimens, maintenance chemotherapy status, and RT technique.

The median follow-up durations were 12 months (range: 2–103 months) for all patients and 52 months (range: 12–100 months) for the 18 surviving patients. The 3-year and 5-year overall survival (OS) rates for all patients were 26.6% and 18.4%, respectively. The 3-year rates of disease progression-free survival (DPFS), local progression-free survival (LPFS), and distant metastasis-free survival (DMPFS) were 17.8%, 62.8%, and 51.5%, respectively.

Table 2 shows the results of the univariate analyses for OS, DPFS, LPFS, and DMPFS. The comparisons of the <66 Gy and ≥66 Gy groups revealed significant differences in the 3-year rates of OS (117.9% vs. 32.1%, p = 0.026) and LPFS (46.1% vs. 72.1%, p = 0.005), but not in DPFS (11.4% vs. 21.9%, p = 0.059) and DMPFS (49% vs. 52.6%, p = 0.852). The median OS values in the ≥66 Gy and <66 Gy groups were 13 months and 10.2 months, respectively, which corresponded to 5-year OS rates of 24.1% and 10.2%, respectively. The survival curves of each group are plotted in Fig. 1. In the univariate analyses, poor OS was associated with older age (≥60 years), ECOG PS of 1–2, tumour length of >5 cm, clinical N2–3 stage, and not receiving maintenance chemotherapy.

Table 2 Univariate predictors of overall survival, disease progression-free survival, local progression-free survival, and distant metastasis-free survival.
Full size table
Figure 1
figure 1

Kaplan-Meier estimates of (A) overall survival, (B) disease progression-free survival, (C) local progression-free survival, and (D) distant metastasis-free survival according to radiation dose.

Full size image

Table 3 shows the results of the multivariate analysis, which included the significant factors from the univariate analyses. These results indicated that a RT dose of ≥66 Gy independently predicted better OS (hazard ratio [HR]: 0.646, 95% confidence interval [CI]: 0.425–0.982), and better LPFS (HR: 0.432, 95% CI: 0.216–0.865), but not better DPFS (HR: 0.727, 95% CI: 0.485–1.091) or DMPFS (HR: 1.044, 95% CI: 0.524–2.079). Poor OS was also independently associated with ECOG PS of 1–2, clinical N2–3 stage, and not receiving maintenance chemotherapy. Patients with ECOG PS of 1–2 independently had inferior DPFS, and LPFS. Advanced clinical N stage independently predicted inferior DPFS and DMPFS.

Table 3 Prognostic factors in multivariate analysis of overall survival, disease progression-free survival, local progression-free survival and distant metastasis-free survival.
Full size table

The subgroup analysis excluded 3 patients who received an RT dose of <41.4 Gy, but still revealed a significantly higher 3-year OS rate in the ≥66 Gy group than in the 41.4–65.9 Gy group (32.1% vs. 17.1%, p = 0.027) (Fig. 2A). We also calculated the OS from the latest date of RT to death or the last follow-up, which still revealed a significantly better 3-year OS rate in the ≥66 Gy group than in the 41.4–65.9 Gy group (32.2% vs. 14.6%, p = 0.037) (Fig. 2B). We further excluded 13 patients with RT dose <50 Gy. The 3-year OS rate was significantly higher in the ≥66 Gy group relative to the 50–65.9 Gy group, with rates of 32.1% and 16.1%, respectively (p = 0.035) (Fig. 3).

Figure 2
figure 2

Kaplan-Meier curves of overall survival estimated from the (A) beginning date or (B) the latest date of RT stratified based on radiation doses of 41.4–65.9 Gy versus ≥66 Gy.

Full size image
Figure 3
figure 3

Kaplan-Meier curves of overall survival stratified based on radiation doses of 50–65.9 Gy versus ≥66 Gy.

Full size image

The treatment-related toxicities are shown in Table 4. The only significant difference between the <66 Gy and ≥66 Gy groups was observed for acute dermatitis (3 patients [7%] vs. 20 patients [28%], p = 0.009), with a trend towards more acute grade ≥2 dysphagia in the ≥66 Gy group (41% vs. 58%, p = 0.073). No significant differences were observed for acute anaemia (p = 0.120), acute neutropenia (p = 0.605), pneumonitis (p = 0.190), tracheoaortic and tracheoesophageal fistula (p = 0.558), or benign oesophageal stricture (p = 0.431).

Table 4 Toxicities according to radiotherapy doses of <66 Gy or ≥ 66 Gy.
Full size table

Discussion

The present study investigated the relationship between RT dose and survival among patients with non-metastatic EC who underwent definitive RT with chemotherapy. The results indicate that a total RT dose of ≥66 Gy at the primary tumour was associated with significantly better OS and LPFS than a dose of <66 Gy, and this difference persisted after excluding patients who received an inadequate RT dose (<41.4 Gy or <50 Gy). It is possible that the results were related to the immortal time bias12, as we initially evaluated survival from the start date of RT for both groups. However, we also evaluated the survival outcomes from the last date of RT, which eliminated the immortal time bias and revealed that the significant difference in OS persisted between the two groups (Fig. 2B).

Although various studies have evaluated RT dose escalation for treating EC, the potential clinical benefits remain controversial5,6,13,14,15,16,17. Brower et al. used the National Cancer Data Base in the US to evaluate RT dose escalation for patients with stage I–III EC who underwent CCRT13, and reported that higher RT doses (51–54, 55–60, or >60 Gy) did not improve OS relative to a dose of 50–50.4 Gy. However, their cohort only included a limited number of patients who underwent IMRT or 3DCRT (39%), and the predominance of cases with an unknown RT modality (61%) might have influenced the apparent effect of RT dose on survival in the modern era of RT. Recently, Xu et al. reported an abstract of a randomized study to evaluate the clinical benefit of high-dose RT using modern techniques14. They found no significant differences in local/regional progression-free survival, DPFS, OS and toxicity between the high-dose (60 Gy) and low-dose (50 Gy) group. Chen et al. performed a population-based propensity score-matched study using Taiwanese registry data from patients with EC who underwent IMRT or 3DCRT15, and reported that an RT dose of ≥60 Gy may provide better 5-year OS than 50–50.4 Gy (22% vs. 14%, p < 0.05). He et al. retrospectively reviewed 193 patients with EC who underwent CCRT16, and reported that the high-dose group (>50.4 Gy) had a significantly lower local failure rate (17.9% vs. 34.3%, p = 0.024) than the low-dose group (≤50.4 Gy), but there was no significant difference in the 5-year OS rates (41.7% vs. 33.0%, p = 0.617). Cao et al. evaluated outcomes among 115 patients with squamous cell carcinoma of the cervical oesophagus who underwent definitive RT with or without chemotherapy17, and reported significantly higher rates of distant failure-free survival and OS for their ≥66 Gy group than for their <66 Gy group. Kim et al. investigated 236 patients with stage II–III EC who underwent definitive CCRT (<60 Gy for 120 patients and ≥60 Gy for 116 patients)6, and reported that the high-dose group experienced significantly better 2-year locoregional control (69.1% vs. 50.3%, p = 0.002), median PFS (16.7 months vs. 11.7 months, p = 0.029), and median OS (35.1 months vs. 22.3 months, p = 0.043). A recent systematic review by Song et al. has also indicated that CCRT with an RT dose of >60 Gy provided better tumour response, locoregional control, and OS5. Our results also support the association between improved OS and higher RT doses (≥66 Gy), relative to doses of <66 Gy or 41.4–65.9 Gy.

The present study revealed that many radiation oncologists in our department did not follow the NCCN guidelines’ recommendations regarding the RT dose for treating EC, as most patients (96/115 patients) received doses of >50.4 Gy. It provided an opportunity to examine the survival and toxicity effects of escalated RT doses (≥66 Gy) in this setting. Kim et al. have reported a positive correlation between RT dose and locoregional control among patients with EC6. Furthermore, Lertbutsayanukul et al. evaluated 44 patients with locally advanced EC who underwent CCRT (>50 Gy) before esophagectomy18, and reported that an RT dose of >60 Gy was associated with better pathological complete remission than lower doses (59.1% vs. 36.4%). Moreover, Welsh et al. reported that unresectable EC with gross tumour volume (GTV) failure after CCRT was associated with shorter survival than cases without GTV failure, which suggested that local control helped improve survival3. Therefore, we assume that higher RT doses in the present study provided improved local control that translated into improvements in OS.

Regarding treatment-related toxicity, one systematic review has indicated that there were no differences in grade ≥3 acute or late esophagitis between high-dose RT (≥60 Gy) and standard-dose RT5. Furthermore, other toxicities are rare and moderately tolerable. Two retrospective studies have also evaluated treatment-related toxicities, with Kim et al. reporting no significant differences between the <60 Gy and ≥60 Gy groups6. However, relative to a low-dose group (≤50.4 Gy), He et al. reported that higher doses (>50.4 Gy) were associated with higher rates of grade 3 skin reactions (12.5% vs. 2.2%, p < 0.001) and oesophageal stricture (32.1% vs. 18.2%, p = 0.037)16. In the present study, the only dose-related difference in treatment-related toxicity was observed for acute dermatitis (Table 4), although only 1 patient in the ≥66 Gy group experienced grade 3 acute dermatitis.

The present study has several limitations. The first is the retrospective design and the variable chemotherapy regimens. The second is the different patient characteristics for each group, with the high-dose group having more favourable performance status, which might have influenced the results, although a multivariate Cox proportional hazard model was used to adjust for potential confounding factors. Some may argue that patients in the low-dose group were treated with more palliative intent. However, we analysed the chemotherapy regimens and maintenance chemotherapy status between the <66 Gy and ≥66 Gy groups. No significant differences were found between the 2 groups in terms of chemotherapy regimens and maintenance chemotherapy status, which means that the patients in the low-dose group were not treated with more palliative intent. The third is the regional differences in the histological type of EC, with our cases predominantly involving squamous cell carcinoma (95%). Similar results were reported in the 2015 report by the Taiwan Department of Health19, although squamous cell carcinoma is declining in North America and Europe, with concurrent increases in adenocarcinoma of the distal oesophagus and gastroesophageal junction (approximately 70% of oesophageal carcinomas in the US)20. Thus, the results of the present study might not be generalizable to regions where the dominant tumour histology is adenocarcinoma. The fourth is the hidden bias which is a common issue for observational study21. For example, body weight loss, cachexia, or nutrition status was not controlled in the current study22.

In conclusion, the present study’s results suggest that higher RT doses (≥66 Gy), when administered using modern RT technique (3DCRT, IMRT, or VMAT), are feasible for patients with inoperable and non-metastatic EC. In addition, the higher RT doses were associated with significantly better LPFS and OS than for RT doses of <66 Gy. The two groups had similar toxicity incidences and severities, with the only significant differences observed for acute dermatitis, which is a manageable event. Furthermore, inferior OS was independently predicted by poor performance status, clinical N2–3 stage, and not receiving maintenance chemotherapy. Nevertheless, these findings require validation in prospective randomized trials using modern RT techniques.

Methods

Data source

We retrospectively evaluated 115 patients with histologically-confirmed EC who underwent definitive RT with chemotherapy at our department between July 2003 and November 2016. All patients had initially undergone a physical examination, chest radiography, barium swallow, chest computed tomography (CT), abdominal sonography, and upper gastrointestinal panendoscopy and/or fluorodeoxyglucose positron emission tomography (PET-CT). Patients were excluded if they had undergone primary surgery or had distant metastasis. Data were collected regarding demographic characteristics, performance status, pathology, imaging results, disease stage, reason for not undergoing surgery, RT, chemotherapy, and follow-up results. Disease staging was based on the 7th version of the American Joint Committee on Cancer guidelines (7th AJCC, 2009). The patients who were diagnosed within 2003–2009 were retrospectively staged in accordance with the 7th AJCC guidelines by reviewing their medical records and images. The retrospective study protocol was approved by the institutional review board of the Tri-Service General Hospital in Taiwan (1-101-05-041). All methods were performed in accordance with the relevant guidelines and regulations.

Radiotherapy technique

An Elekta linear accelerator was used to deliver RT with 15-MV photons. The GTV contained the primary tumour and metastatic lymph nodes, and was determined using CT, barium swallow, and endoscopy with or without PET-CT. The clinical target volume (CTV) was obtained by expanding the GTV by 3–5 cm at the cranial and caudal margins and by 0.3–2 cm at the transversal margins. The planning target volume included the CTV in the initial phase or the GTV in the boost phase, with additional margins of 0.3–1 cm in all directions to account for organ movement. All patients had undergone VMAT, IMRT, or 3DCRT, with sequential boost or simultaneous integrated boost approaches being used in our department to treat EC. During the sequential boost approach, the irradiated field encompassed the CTV and the regional lymph node drainage area, with a dose of 41.4–50.4 Gy in 1.8–2-Gy fractions, which was followed by a total RT dose of 50.4–70 Gy to the GTV. During the simultaneous integrated boost approach, VMAT or IMRT was used to simultaneously deliver a high dose (50.4–70 Gy) to the GTV and a lower dose (45–60 Gy) to the CTV and the regional lymph node drainage area. The irradiated tumour length was identified using the CT, barium swallow scans, endoscopy and/or PET-CT. The total RT dose included the doses to the primary tumour from the external beam RT, and patients were grouped according to GTV RT doses of <66 Gy (median: 54 Gy, range: 14.4–65 Gy) or ≥66 Gy (median: 66 Gy, range: 66–70 Gy).

Chemotherapy

The chemotherapy regimens were selected based on the patient’s age, general physical condition, and the oncologist’s preference. Most patients (n = 95) received cisplatin 60–100 mg/m2 on day 1 plus 5-fluorouracil (5-FU) 500–1000 mg/m2 daily on days 1–4 through continuous intravenous infusion every 3–4 weeks. Two cycles of cisplatin plus 5-FU were administered concurrently during RT. The other regimens during RT were cisplatin monotherapy (n = 14) with 30–40 mg/m2 weekly or 60–100 mg/m2 every 3–4 weeks, and 5-FU monotherapy (n = 6) with 500–1000 mg/m2 daily on days 1–4 every 3–4 weeks. Patients received additional maintenance chemotherapy if a medical oncologist determined that their medical condition and disease status would allow this.

Follow-up and toxicity assessment

Toxicity assessments were performed weekly during the RT. The first follow-up evaluation was performed at 1 month after RT and follow-ups were then continued at 3–6-month intervals. During the follow-up, the patients underwent physical examinations, blood and biochemical testing, chest CT, barium scans, and abdominal sonography. Upper gastrointestinal panendoscopy with biopsy or PET-CT were performed if clinically indicated. Toxicities were scored using version 4.0 of the Common Terminology Criteria for Adverse Events, and toxicities were considered acute if they developed within 1 month after completing RT.

Statistical analysis

The OS interval was calculated from the start of RT to death or the last follow-up. The DPFS interval was calculated from the start of RT to the first instance of recurrence, death, or the last follow-up. The LPFS and DMPFS intervals were calculated from the start of RT to the last follow-up or the first instance of local or distant recurrence, respectively. Categorical variables were analysed using the chi-square test, and survival analyses were performed using the Kaplan-Meier method. Univariate and multivariate analyses were performed using the log-rank test and a Cox proportional hazard regression model, respectively. A Cox proportional hazards model was used to determine the HR and 95% CI values. Differences were considered statistically significant at a p-value of <0.05. All analyses were performed using SPSS software (version 18.0; SPSS, Chicago, IL).