A retrospective study comparing definitive chemoradiotherapy vs. chemoradiotherapy followed by surgery in T4 esophageal squamous cell carcinoma patients who were downstaged after neochemoradiotherapy

The outcome of patients with T4 esophageal squamous cell carcinoma (ESCC) is extremely poor. Two distinct therapeutic options are currently available for T4 esophageal cancers: neochemoradiotherapy followed by surgery (CRT-S) and definitive chemoradiotherapy (D-CRT). This study aimed to investigate the clinicopathologic characteristics of T4 ESCC in Chinese patients and compare the survival between the two therapeutic options. We retrospectively analyzed 125 patients with clinically unresectable T4 ESCC in Tianjin Medical University Cancer Institute and Hospital from January 2010 to December 2020. Overall survival (OS), progression-free survival (PFS) and associated factors were analyzed. A total of 106 of 125 T4 ESCC patients were downstaged of the tumor by neoadjuvant CRT. Among 106 patients, 32 patients underwent CRT-S, and 74 patients underwent D-CRT. Patients in the CRT-S group had a higher OS (20.4 months vs. un-reached median OS, p = 0.037) and PFS (8.6 months vs. 21.0 months, p = 0.008) than those in the D-CRT group. In multivariate analysis, treatment was an independent predictor of PFS. After propensity score matching (PSM), 50 patients (CRT-S = 25; D-CRT = 25) were matched. Among these 50 patients, patients in the CRT-S group had a higher OS (15.6 months vs. un-reached median OS, p = 0.025) and PFS (7.2 months vs. 18.8 months, p = 0.026) than those in the D-CRT group. In multivariate analysis, treatment was an independent predictor for PFS. We demonstrated that CRT-S was superior to D-CRT for T4 ESCC patients who were downstaged by neo-CRT with respect to longer OS and PFS. Randomized controlled trials involving large population samples are needed to define the standard treatment for T4 ESCC.


Introduction
Esophageal squamous cell carcinoma (ESCC) is one of the leading causes of cancer-related mortality worldwide, with a high prevalence in Asia [1,2]. The outcome of patients with T4 esophageal cancer, defined as a tumor that invades neighboring structures (e.g., aorta, trachea, bronchus, and lung), is extremely poor [3]. Despite advances in surgical treatment, these tumors are usually considered inoperable; however, surgery alone has not improved the prognosis of patients with T4 esophageal tumors [4,5].
Two distinct therapeutic options are currently available for T4 esophageal cancers: chemoradiotherapy followed by surgery (CRT-S), which comprises esophagectomy following downstaging of the tumor by CRT, and definitive chemoradiotherapy (D-CRT), which is designed to avoid esophagectomy by using maximum doses of irradiation [4,6]. Until now, D-CRT has been the standard alternative curative management for patients with locally advanced disease who are not eligible for or refuse surgery [7]. To our knowledge, there is little information on the differences in the clinical outcomes of patients with T4 ESCC who undergo D-CRT and those who receive CRT-S. In this study, we discuss these two treatment modalities.

Patients
We retrospectively enrolled patients with ESCC who had been treated with CRT from January 2010 to December 2020 at our center, including 8 cases from a single-arm, single-center, investigator-initiated, exploratory, phase II clinical study (ClinicalTrials.gov NCT04137679) at Tianjin Medical University Cancer Institute & Hospital in Tianjin, China. The inclusion criteria were as follows: (1) squamous cell histological type; (2) staged as unresectable T4 disease by a multidisciplinary team based on biopsy and imaging pretreatment examination data; (3) downstaged to T3 or lower T stage by neoadjuvant CRT; (4) without distant organ metastases; and (5) Eastern Cooperative Oncology Group (ECOG) performance status (PS) of ≤ 1. All patients were staged according to the 8th edition of the American Joint Committee on Cancer (AJCC) staging manual.
Adjacent organ invasion was diagnosed by esophageal endoscopy ultrasound (EUS), endobronchial ultrasonography (EBUS), bronchoscopy, and/or computed tomography (CT) according to the following criteria: 1) EUS showed invasion of the trachea, bronchus, aorta, pleura, pericardium, and/or other peripheral organs; 2) EBUS showed invasion of the trachea and/or bronchus; 3) bronchoscopy showed protrusion of the esophageal tumor into trachea and/or bronchi or abnormal tracheal mucosa; and 4) if the patient could not undergo EUS, adjacent organ invasion was defined by CT or positron emission tomography (PET)-CT. Generally, invasion of the peripheral organs was diagnosed based on the loss of fat planes between the esophagus and peripheral organs. Invasion of the aorta was defined as > 90 degrees of the aorta surrounded by tumor in more than one CT slice. Invasion of the pericardium was defined as the disappearance of the fat line at the focus level and local pericardial thickening.
The following clinicopathologic parameters for each patient were also collected: sex, age at diagnosis, smoking history, drinking history and TNM stage in line with the 8th edition of the esophageal cancer staging system. The PFS and OS of patients diagnosed from January 2010 to December 2020 were recorded based on a follow-up clinic visit or a telephone call.

Treatment and response
Patients received conventional fraction radiotherapy using a 6 MeV linear accelerator. The radiation dose was calculated by intensity-modulated radiotherapy (IMRT). All plans were based on 3-or 5-mm CT scan images obtained in the treatment position before radiotherapy. The gross tumor volume (GTV) included the esophageal tumor found under CT and esophageal endoscopy and the enlarged locoregional lymph nodes found by CT before treatment. GTV also included distant lymph nodes with metastasis for patients with cervical or abdominal distant lymph node metastases. The clinical target volume (CTV) was added to the esophageal tumor 5-6 mm laterally and 2-3 cm in the cephalo-caudal direction, and CTV also includes positive lymph nodes and its drainage area. The planning target volume (PTV) was added to the CTV 5 mm in the cephalo-caudal and lateral directions. Cone-beam computed tomography (CBCT) was used to guarantee the treatment position during the whole radiotherapy process once a week. The induction and concurrent chemotherapy with radiation included cisplatin, 5-FU or taxane-based regimens. The patients' response to treatment was assessed by CT scan, endoscopy and upper gastrointestinal contrast in accordance with Response Evaluation Criteria in Solid Tumors Keywords: Esophageal squamous cell carcinoma, T4, Neo-chemoradiotherapy followed by surgery, Definitive chemoradiotherapy 1.1 (RECIST 1.1) criteria and the method reported in our previous study [8].
Patients in the CRT-S group received a 40-Gy radiation dose (2 Gy once daily in 20 fractions, 5 days a week) and were scheduled to undergo resection 4-6 weeks after having completed induction chemoradiation. Patients in the D-CRT group received a 60-Gy radiation dose (2 Gy once daily in 30 fractions, 5 days a week). An intensitymodulated radiotherapy dose planning system was used.

Propensity score matching analysis
In this study, propensity score matching (PSM) was used to reduce bias due to an imbalance in observed variables between the CRT-S and D-CRT groups. Three baseline characteristics (sex, T-stage and segment) were selected as covariates in the PSM model, and the match tolerance was set to 0.01. Propensity scores of individuals were calculated with logistic regression analysis (SPSS version 22.0, Chicago, IL), and then, the optimal 1:1 matching between CRT-S and D-CRT patients was produced based on propensity scores. After matching, the distribution of the remaining observed variables was similar in the CRT-S and D-CRT groups.

Statistical analysis
Pearson's χ 2 test was used to investigate the correlations between 2 categorical variables. PFS and OS distribution was analyzed using the Kaplan-Meier method, and logrank tests were employed for comparison of PFS or OS between 2 categories in univariate analysis. Multivariate survival analysis was conducted using Cox proportional hazards regression to identify independent prognostic factors. Data were statistically analyzed using SPSS 22.0 (Abbott Laboratories, North Chicago, IL, USA). Statistical significance was set at p < 0.05.

Ethics statement
This study was approved by the institutional review board at Tianjin Medical University. Written informed consent was obtained from each patient to allow their biological samples to be genetically analyzed. The experimental protocol of this study was performed strictly in accordance with the guidelines.

Results
From January 2010 to December 2020, 125 T4 ESCC patients were treated with CRT in our center; of these, 106 patients were down staged to T3 or lower T stage. The patients' demographics and tumor characteristics are listed in Table 1. After multi-disciplinary treatment (MDT) of department of radiation oncology and department of esophageal cancer, all patients were suitable for surgery or definitive chemoradiotherapy after neochemoradiotherapy. Of the whole 106 cases, 8 cases were from a single-arm, single-center, investigator-initiated, exploratory, phase II clinical study (ClinicalTrials. gov NCT04137679) at Tianjin Medical University Cancer Institute & Hospital in Tianjin, China. The choices of the 8 patients were randomly made according to the random system. For the rest of the 98 cases, doctors provided the two possibilities of treatment ways equally and patients and doctors made the final decision together. Of the 106 downstaged T4 stage patients, 32 patients received surgery after neo-CRT, and 74 patients received D-CRT (Fig. 1). These patients had a median followup duration of 17.3 months (range: 2.6-90.6 months). The cohort consisted mostly of males, 92 male patients (86.8%) and only 14 female patients (13.2%). The majority of enrolled patients had upper (36; 34.0%) or middle (42; 4.0%) thoracic ESCC. In comparison with the D-CRT group, the CRT-S group had a lower proportion of T4b stage (p = 0.063), and more patients were in the middle or lower segment (p < 0.001). No other significant differences were found between the patients in the CRT-S and D-CRT groups with respect to age, sex, smoking history, drinking index, Karnofsky Performance Status (KPS) score, N stage or weight loss (Table 1).

Discussion
In the present study, we retrospectively investigated the clinical characteristics and prognosis of patients with T4 stage ESCC who were downstaged after neo-CRT. We   independent predictor for PFS. In the present report, the median OS of the CRT-S group was not available at the data cutoff (37.6 months median follow-up time). However, our findings showed that the 24-month OS rate was 69.4%. The therapeutic strategy for locally advanced inoperable ESCC is controversial because the outcome differs among institutions [3,6,[9][10][11][12]. For locally advanced inoperable ESCC, although concurrent chemoradiotherapy (CCRT) is recommended [13][14][15], neo-CRT is usually performed [10,16]. Real-world data reported neoadjuvant chemoradiotherapy (44.9%) and definitive chemoradiotherapy (36.0%); however, 27.9% of patients undergoing neoadjuvant chemoradiotherapy did not receive planned esophagectomy [17]. In retrospective studies of T4b EC, patients who underwent surgicalbased therapy had the longest median OS. However, the grouping was not random. Well-selected responders to chemotherapy (CMT) and/or radiotherapy (RT) may be able to undergo resection and numerically prolong survival, but patient selection remains paramount [18,19]. The median OS in the CMT, RT and surgery-based therapy groups was 6.0, 12.7, and 43.9 months (P < 0.001), respectively. Nonsurgical treatment was associated with poorer OS (P < 0.05). A similar problem also existed in a study from Japan. A small sample study from Japan reported that the OS of T4b EC patients who underwent DCF-RT was 50% at 3 years compared to 37.5% for all T4b patients [18]. The mean interval was 17 months compared to 14.3 months for all included T4b patients. The 5-year survival rate was 19% in the D-CRT group [20,21].
However, some studies showed negative results. A study from Japan enrolled 71 patients with T4 EC with tracheobronchial invasion (TBI); 58 underwent dCRT, and 13 underwent iCRT-S. There was no significant difference between the dCRT and iCRT-S groups. Clinical LN negativity and later treatment period were significantly good prognostic factors for T4 EC with TBI [10]. The median survival times (MSTs) of patients with locally advanced esophageal cancer with adjacent organ invasion after definitive CRT, bypass surgery plus CRT and CRT followed by esophagectomy were 10.4, 11.0 and 16.4 months, respectively; MST did not differ significantly between patients [22]. To our knowledge, our study is the first to compare CRT-S and D-CRT in locally advanced inoperable ESCC patients who were downstaged after neo-CRT. The choice of the 8 patients from a single-arm, single-center, investigator-initiated, exploratory, phase II clinical study (ClinicalTrials.gov NCT04137679) was randomly made according to the random system. For the rest of the 98 cases, doctors provided the two possibilities of treatment ways equally and patients and doctors made the final decision together. It is a basic principle to respect the patients right of consent. After propensity score matching (PSM), no significant differences between the CRT-S and D-CRT groups were observed in the patient characteristics. Although we acknowledge the limitations of retrospective studies, surgery after neo-CRT may probably improve survival and decrease recurrence compared with D-CRT for locally advanced inoperable ESCC patients who are downstaged after neo-CRT. The first failure sites were categorized as local, regional nodal, or distant. In previous studies of locally advanced esophageal cancer patients who underwent chemoradiation followed by esophagectomy, the rates were 4.5-67% local, 10-57.9% regional nodal, and 21-32% distant failures [23][24][25][26]. Local failure was correlated with fewer lymph nodes (LNs) assessed and close/positive margins. Regional nodal failure was correlated with fewer LNs assessed and larger pretreatment tumor size. Distant recurrence was correlated with higher pathologic nodal stage, ulceration, perineural invasion, residual disease, and higher posttreatment PET SUV max. Patients with a pathologic complete response were less likely to experience distant recurrence [24]. In our study, all cases were evaluated as down staged to operable by CT, endoscopy and upper gastrointestinal contrast. Most cases in the CRT-S group had distant metastasis (8/32, 25.0%), and the local progression rate was 3.1% (1/32). The low local progression rate may be partially related to a high pathological complete response (pCR) rate (37.5%, 12/32), fewer lymph nodes (N0, 23/32; N1, 3/32; N2, 4/32; N3, 0/32), high negative margin rate (32/32, 100%) and high R0 resection rate (30/32, 93.75%) [24]. In the D-CRT group, the distant metastasis rate was only 16.2% (12/74). Most patients in the D-CRT group had local progression (27/74, 36.5%), 6 of whom were complicated with esophageal fistula, and 2 of whom had massive hemorrhage (Table 5). Our data indicated that CRT-S reduced the local recurrence rate and the incidence of esophageal fistula (P = 0.097) and massive hemorrhage (p = 0.348), which was partially consistent with previous studies [3,27].
In a randomized study comparing D-CRT and CRT-S for patients with locally advanced operable esophageal carcinoma, preoperative chemoradiotherapy improved survival among patients with potentially curable esophageal or esophagogastric junction cancer. The regimen Regional lymph node metastasis 6 19 Distant metastasis 8 12 Stable was associated with acceptable adverse event rates [28][29][30]. Some studies discussed the therapeutic options of T4b EC; although CRT-S improved the survival rate, perioperative complications were also increased. Anastomotic leakage after surgery was 25%, the recurrent laryngeal nerve paralysis rate was approximately 50% [21], the reoperation rate was approximately 7.7% [23], and the operative mortality was 5-23.4% [25]. In the present study, all adverse events were no more than grade 3 and manageable. Leukocytopenia incidence (p = 0.038) in the CRT-S group was lower than those in the D-CRT group. Although there were no differences of statistics in neutrophilic granulopenia incidence (p = 0.58), thrombocytopenia incidence (p = 0.087) and radiation esophagitis incidence (p = 0.080) in the CRT-S and D-CRT group, grade 3 radiation esophagitis incidence was lower in the CRT-S group than the D-CRT group. In the CRT-S group, anastomotic leakage, post thoracotomy pulmonary infection and operative mortality were lower than those in previous reports in patients with advanced ESCC (Table 5) [20,21,23,25].