Detection of secondary upper gastrointestinal tract cancer during follow‐up esophagogastroduodenoscopy after gastrectomy for gastric cancer

Abstract Aim Esophagogastroduodenoscopy (EGD) may contribute to early detection of secondary cancer in the upper gastrointestinal tract although the clinical relevance of follow‐up after gastrectomy remains unclear. This study aimed to elucidate the effectiveness of follow‐up EGD by investigating the incidence of secondary cancer in any part of the upper gastrointestinal tract. Methods Data from 1438 patients who underwent curative partial gastrectomy for primary gastric cancer between 2008 and 2014 and follow‐up EGD at least once during a 5‐year follow‐up period were retrospectively reviewed. Incidence rates of remnant gastric cancer, laryngeal cancer, and esophageal cancer detected after follow‐up EGD were determined, and risk factors for secondary cancers were examined. The characteristics of clinicopathological diagnoses of secondary cancers were reviewed and compared according to the frequency of follow‐up EGD. Results The average annual frequency of EGD was 0.7, while the 5‐year cumulative incidence rates of remnant gastric cancer and secondary laryngeal and esophageal cancers were 2.9% and 1.3%, respectively. Risk factors for remnant gastric cancer included heavy smoking, proximal gastrectomy, and tumor size ≥ 30 mm. All secondary cancers were resectable upon diagnosis, with endoscopically resectable cancer accounting for 81.0% of cases. Our results found a significantly higher proportion of endoscopically resectable cancers during regular follow‐up than during infrequent follow‐up. Conclusions Follow‐up EGD can be a useful modality for detecting secondary upper gastrointestinal tract cancer, likely leading to curative treatment for secondary cancer. Focusing on patients presenting with risk factors may increase the value of follow‐up EGD after gastrectomy.


| INTRODUC TI ON
Curative resection is the standard treatment for gastric cancer, and surgery combined with adjuvant chemotherapy improves survival rates. 1 However, the significance of and optimum protocol for follow-up after surgery remain unclear. Follow-up after gastrectomy is generally conducted to diagnose post-gastrectomy syndromes and postoperative complications, and promptly detect secondary cancer and recurrence. 2 Some studies have shown that early detection of secondary cancer or recurrence during follow-up after gastrectomy improved overall survival; however, other studies have shown that early detection of recurrence after intensive follow-up examinations did not improve overall survival. [3][4][5] Five-year cumulative incidence rates of remnant gastric cancer after gastrectomy have been estimated in the range of 1.4%-6.8%, [6][7][8] with patients having early remnant gastric cancer showing relatively good survival outcomes. [9][10][11] Esophagogastroduodenoscopy (EGD) is useful for the detection of early remnant gastric cancer, 10  Although several studies on follow-up EGD and remnant gastric cancer are available, only a few studies have investigated secondary cancers in the upper gastrointestinal tract. Patients with gastric cancer may have a high risk of secondary head and neck cancers and esophageal cancers. 13 We examined the incidence rates of secondary cancer This study aimed to evaluate the effectiveness of regular follow-up EGD after gastrectomy by analyzing 5-year cumulative incidence rates of secondary cancer in any part of the upper gastrointestinal tract and associated treatment outcomes. Moreover, risk factors for secondary cancer and characteristics of patients most likely to benefit from follow-up EGD were identified.

| Patients
A total of 2104 patients underwent curative gastric resection (R0 surgery) for primary gastric cancer between January 2008 and December 2014 at the Shizuoka Cancer Center. A total of 1438 patients were included after 511, two, and 153 patients who underwent total gastrectomy, had a history of surgical resection of the larynx or esophagus, and had not undergone EGD within 5 years after gastrectomy, respectively, were excluded ( Figure 1).

| Follow-upperiodandEGDexamination
According to the 2018 Japanese Gastric Cancer Treatment Guidelines, 2 EGD examination was performed 1, 3, and 5 years after surgery, although the timing of EGD and medical facilities where patients received EGD were determined for each patient by the attending surgeons. Data on EGD examinations performed at the clinic were extracted from medical records, patient referral documents, and EGD reports. Atrophic gastritis was evaluated using the Takemoto-Kimura classification, 14 with moderate or severe atrophic gastritis defined as C-3, O-1, O-2, and O-3.

| Statisticalanalyses
The Fisher exact test and Mann-Whitney U test were used to compare categorical and continuous variables between groups, respectively. We regarded death as a competing risk in the analysis of secondary upper gastrointestinal cancer cumulative incidence rates.
Risk factors for secondary cancer were analyzed using the Fine and Gray model, 16 in which the strength of the association between variables and secondary cancer risk was assessed using the subdistribution hazard ratio (SHR). Risk factors were examined in the multivariate analysis using variables with P-values of <.1 in the univariate analysis. All statistical analyses were performed using R Statistics

| Follow-upEGDexamination
The median follow-up duration for EGD was 1700 days (

| Clinicopathologicalcharacteristics
The clinicopathological characteristics of the patients at initial surgery are summarized in Table 1. Forty-three patients had a history of upper gastrointestinal cancer, and all previous upper gastrointestinal cancers were curatively treated using endoscopy or chemoradiotherapy.

| Cumulativeincidenceratesand characteristics of secondary upper gastrointestinal cancer
Remnant gastric cancer, lower laryngeal cancer, and esophageal cancer were found in 44, eight, and 15 patients, respectively. Remnant gastric cancer and lower laryngeal cancer were found in one patient, and laryngeal cancer and esophageal cancer were diagnosed in two patients during the same EGD examination. The cumulative incidence rates of remnant gastric cancer at the first, third, and fifth year after surgery were 0.3%, 1.4%, and 2.9%, respectively, and those of secondary laryngeal and esophageal cancers were 0.3%, 0.8%, and 1.3%, respectively ( Figure 2).  Table 1. Proximal gastrectomy was significantly more frequent in patients with secondary upper gastrointestinal cancer than in those who underwent other surgical procedures (P = .010).

| Riskfactors
The risk factors for remnant gastric cancer are shown in Table 2.  (Table S1).

| Pathologicalstaging
The treatment for secondary upper gastrointestinal cancer is shown in  Pathological staging findings of each cancer type are summarized in Figure 4. All cancers were resectable, with endoscopically resectable cancer accounting for 51 (81.0%) of the 63 patients who underwent curative treatment.   (Table S2).  (Table S3). Cumulative incidence rates of secondary upper gastrointestinal cancer at the first, third, and fifth year after surgery were 0.7%, 2.5%, and 4.6% in the regular follow-up group and 0.0%, 1.7%, and 3.3% in the infrequent follow-up group, respectively ( Figure S2). No significant difference in the cumulative incidence rates was observed between the groups (SHR, 1.261; 95% CI, 0.713-2.216; P = .421).
Recurrence of remnant gastric cancer was found in two patients in each group (P = .238).

| DISCUSS ION
Among 1438 patients who underwent gastrectomy, 2.9% were found to have remnant gastric cancer, while 1.3% developed laryngeal and esophageal cancer. All secondary cancers were resectable, and most of them underwent curative endoscopic resection. Death is considered a competing risk in the analyses of cumulative incidence that included patients who died during the follow-up period. In the Kaplan-Meier and Cox regression analyses, patients with competitive risks should be excluded, as these methods overestimate the incidence of secondary cancer. In contrast, the competitive risk model enables the identification of incidence rates of and risk factors for secondary cancer in all patients, including those with advanced cancer. 18,19 In the present study, the 5-year cumulative incidence rates of secondary laryngeal and esophageal cancers were 1.3%.
Meanwhile, a previous study reported the 5- Several guidelines recommend follow-up EGD after gastrectomy, but only a few define the optimum follow-up duration and intervals between examinations. 29,30 The present findings suggest that regular follow-up EGD may contribute to early detection of secondary upper gastrointestinal cancer and enable less invasive curative treatment; in addition, annual follow-up EGD may benefit patients with risk factors for secondary upper gastrointestinal cancer.
The current study had several limitations. First, we did not analyze the cost-effectiveness of follow-up EGD. Given that EGD is relatively inexpensive in Japan, the cost may not restrict clinical implementation. Secondly, we could not collect information regarding Helicobacter pylori infection, which is a known risk factor for gastric cancer. 31 Our hospital does not routinely assess for Helicobacter pylori infection before gastrectomy; therefore, atrophic gastritis before initial surgery was substituted for Helicobacter pylori infection.