Validity of upfront surgery for patients with unsuspected lymph node metastasis in esophageal cancer: a propensity scoring matching study

We retrospectively compiled and analyzed data from patients who had undergone curative and complete surgery for intra-thoracic esophageal cancers at a single tertiary Korean hospital from January 2009 to December 2015. Inclusion criteria were complete (R0) and curative surgery cases, intra-thoracic esophageal cancer, preoperative biopsy-proven squamous cell carcinoma, and cases with WL or UL. Exclusion criteria were palliation or salvage cases, other uncured previous or current primary cancers, complete remission cases after neoadjuvant therapy, and operative mortalities (defined as patients who died during hospitalization or within one month after surgery). The preoperative evaluations included esophagogastroduodenoscopy, esophagography, chest CT, abdominal CT, PET- CT, endoscopic ultrasound, and bone scan. In the present study, lesions were regarded as malignant or metastatic when the maximum standardized uptake value was ≥2.5 and the size was ≥1 cm on PET- CT and endoscopic ultrasound. Neoadjuvant and adjuvant therapies were usually performed following the National Comprehensive Cancer Network guidelines or the recommendations from a multidisciplinary team who assessed cancer status and each patient’s condition [4, 17]. Neoadjuvant therapy usually consisted of two cycles of cisplatin and 5-fluorouracil, plus 25 fractions of radiation therapy (over five weeks) to a total of 41–45 Gray. Re-evaluation by PET- CT was performed four weeks after completion of neoadjuvant therapy, and further management was determined. In those cases receiving neoadjuvant therapy, surgery was usually preformed five or six weeks after neoadjuvant therapy completion. In the present study, preoperative stage in neoadjuvant cases was defined as clinical stage by PET-CT reevaluation after completion of neoadjuvant therapy and before surgery. Surgeries were performed by two thoracic surgeons using Ivor Lewis or McKeown procedure depending on cancer status. Patients were usually followed at 3-month intervals for one year after treatment completion, including adjuvant therapy, and then at 6-month intervals. Recurrence or metastasis was diagnosed based on imaging findings including PET-CT, brain MRI, and bone scan, or pathological confirmation when clinically feasible. Cancer stage was determined with regard to the seventh American Joint Committee on Cancer staging system.

UL status is defined as having pathologic evidence of metastatic esophageal cancer in LN after esophagectomy in preoperative cases without LN metastasis (i.e. cN0 on the preoperative evaluations, but pN1 on the postoperative pathologic findings).Since most patients were in an early cancer stage, the sample size was small, and various factors after recurrence can influence overall survival, we carried out a 5 -year disease-free survival (DFS) study instead of an overall survival study to clarify whether upfront esophagectomy without neoadjuvant therapy in UL is appropriate. To clarify the validity of upfront esophagectomy in UL, we compared DFS between WL and UL. In addition, we investigated the benefit of neoadjuvant therapy in UL.

The Student’s T test was used to compare the groups with continuous variables. The Chi-squared test or the Fisher’s exact test was used to evaluate the associations between the groups with categorical variables as appropriate. Disease-free intervals were measured from time of surgery to time of recurrence or last follow-up. DFS between WL and UL was compared using the Kaplan-Meier method with log-rank test. The influence of neoadjuvant therapy on prognosis in UL was also evaluated using the Kaplan-Meier method with log-rank test. To overcome heterogeneity in baseline characteristics between the groups, a matched-analysis based on propensity scores was then carried out to create a cohort of WL patients with clinical characteristics similar to those in the UL. Covariates included age, sex, pathological T stage, method of esophagectomy, and receipt of neoadjuvant therapy. All statistical analyses were performed using the Statistical Package of Social Sciences version 22.0 (Chicago, IL, USA). A p- value less than 0.05 (two- sided) was regarded as statistically significant. This study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital (IRB approval number: KC17RESI0293).

The incidence of UL among preoperative N0 patients was 25.6% and the incidence of UL cases who did not receive neoadjuvant therapy was 47.2%. We included 215 patients (male 199, female 19; mean age 62.8 ± 8.6 years) who met the study criteria. All subjects were stratified into either WL (160 patients) or UL (55 patients). Twenty nine of 55 patients in UL received neoadjuvant therapy before esophagectomy and all patients with LN metastasis received adjuvant therapy after esophagectomy. All cancer histologies confirmed squamous cell carcinomas. Tumors were located in the upper thoracic esophagus (25 cases), middle thoracic esophagus (109 cases), and lower thoracic esophagus (81 cases). The mean tumor length and size were 3.0 (±2.2) cm and 8.3 (±13.8) cm², respectively. The mean number of LN dissected was 24.3 (±11.0). The mean observation period was 29.3 (±20.3) months. Recurrence was found in 46 patients during the observation period. The overall clinico-pathologic characteristics of the study subjects are summarized in Table 1.

There were 160 WL and 55 UL patients. Fifty three patients received neoadjuvant therapy (24 in WL and 29 in UL). We analyzed DFS between WL and UL after propensity score matching to overcome heterogeneity between WL and UL. Covariates included age, sex, neoadjuvant therapy, method of surgery, and pathological T stage (Table 2).There were 55 WL cases and 55 UL cases after propensity score matching. There was no significant difference in DFS between WL and UL (Fig. 1. p = 0.242).

Twenty nine of 55 patients in UL received neoadjuvant therapy and all patients in UL received adjuvant therapy after surgery. To clarify the influence of neoadjuvant therapy in UL, we compared DFS between UL cases that received and did not receive neoadjuvant therapy after propensity score matching to overcome heterogeneity between the two groups. Covariates included age, method of surgery, and pathological T stage (Table 3). There were 26 cases each that receive and did not receive neoadjuvant therapy, respectively, after propensity score matching. There was no significant difference in DFS between cases that received or did not receive neoadjuvant therapy (Fig. 2, p = 0.769).