We reviewed the record of patients with histologically confirmed adenocarcinoma or poorly cohesive carcinoma of the stomach or gastroesophageal junction who underwent curative gastric resection at the Catholic University of Korea, Seoul St. Mary’s Hospital between January 2011 and December 2015. All patients had undergone potentially curative resection with D2 lymph node dissection and had no macroscopic or microscopic evidence of remnant tumors. D2 lymph node dissection was defined according to the guidelines of the Japanese Research Society for the Study of Gastric Cancer [19]. Patients with available electronic medical records were eligible if they met the following criteria: (1) 19 years of age or older; (2) histologically confirmed gastric or gastroesophageal junction adenocarcinoma or poorly cohesive carcinoma; (3) pathological stage I–III according to the American Joint Committee of Cancer, 7th edition [2]; (4) Eastern Cooperative Oncology Group (ECOG) performance status score 0–2; (5) completely recovered from their operation without any complication by 30 days post-surgery; and (6) adequate bone marrow and organ function (hemoglobin > 9.0 g/dL; platelet count > 100,000/mm³; total bilirubin < 1.5-fold higher than the upper normal limit; serum transaminase < 3-fold higher than the upper normal limit; and creatinine < 1.5-fold higher than the upper normal limit). Patients with the following conditions were excluded (1) received neoadjuvant chemotherapy; (2) history of another cancer within the past 5 years (except adequately treated in situ or non-melanoma skin cancers); (3) newly diagnosed with secondary malignancies after surgery for gastric cancer; and (4) history of human immunodeficiency virus infection, autoimmune disease that required treatment with immunosuppressive agents, and/or receipt of systemic corticosteroids 2 weeks prior to surgery.

For surveillance against tumor recurrence, all patients were evaluated every 3 months during the first year, every 6 months for the following 5 years, and annually thereafter. Imaging evaluations were performed using computed tomography and magnetic resonance imaging according to the aforementioned schedule and were analyzed by radiologists at our institution. Levels of carcinoembryonic antigen and carbohydrate antigen 19-9 were also checked at every visit. If evidence suggested the possibility of recurrence, additional imaging or biopsy studies were undertaken to investigate the presence of gastric cancer; any recurrent disease was confirmed by radiology and histology.

Baseline lymphopenia was defined as having an ALC < 1,000/μL within 2 weeks before surgery. If several preoperative ALC values were obtained, the count on the date closest to the surgery was considered. Ever lymphopenic and ever neutropenic were defined as having an ALC < 1,000/μL and an ANC < 1,000/μL, respectively, any time after the gastric cancer diagnosis. However, the following readings were excluded (1) those taken between the day of surgery and 30 days thereafter; (2) those acquired after the confirmation of gastric cancer recurrence; (3) those taken from patients during their last hospitalization if their admittance and death were for reasons other than gastric cancer; and (4) those taken during hospitalization for major surgery for reasons other than gastric cancer.

Overall mortality (OM) was defined as the percentage of patients who died of any cause during the study period. Gastric cancer-specific mortality (GCM) was defined as the percentage of patients who died due to gastric cancer during the study period. Disease-free survival (DFS) was defined as the interval between curative surgery and recurrence, death, or the date of the last follow-up.

A logistic regression model was used to determine whether baseline lymphopenia or ever lymphopenia was associated with age at diagnosis, cancer stage, human epidermal growth factor receptor 2 overexpression or amplification status, adjuvant chemotherapy administration, and ever neutropenia. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated. Variables with a p < 0.05 on the univariate analysis were subjected to multivariate analysis.

OM, GCM, and DFS for ever lymphopenic and never lymphopenic patients were calculated using Kaplan-Meier curves; differences between survival curves were tested for statistical significance using a two-tailed log-rank test. A multivariate Cox proportional hazards regression model was used to identify the effects of ever lymphopenia on survival. Data from patients who were alive on the last follow-up date were censored. Hazard ratios (HRs) and 95% CIs were calculated. All tests were two-sided and p < 0.05 were considered statistically significant. Statistical analyses were performed using IBM version 24.0 (IBM Co., Armonk, NY, USA) and GraphPad Prism version 8.0 (GraphPad Software Inc., San Diego, CA, USA).

This study was approved by the Institutional Review Board of the Catholic University of Korea, Seoul St. Mary’s Hospital (IRB No: KC19RESI0292), and was performed in accordance with Korean regulations and the Declaration of Helsinki. The Institutional Review Board waived informed consent for this study due to the retrospective nature of the analysis.

In total, 1,222 patients with stage I–III gastric cancer were eligible for inclusion in this study. The baseline characteristics of these patients are shown in Table 1. The mean age at diagnosis was 59 years, and two-thirds of the patients (63.7%) were male. Most patients (95.8%) had good performance status (0–1), and 74.1%, 13.8%, and 12.1% of the patients had stage I, II, and III disease, respectively. Most patients had tubular adenocarcinoma (82.6%), and the most common tumor location was the lower third of the stomach (62.8%). One-quarter of the patients (25.0%) received adjuvant chemotherapy; most (93.9%) were never neutropenic, although nearly one-quarter (24.8%) were ever lymphopenic, with a minimum ALC of 640/μL (Table 1). Over a median follow-up time of 4.9 years, 82 patients (6.7%) died, 65 (5.3%) of gastric cancer.

ECOG performance status and cancer stage were associated with baseline lymphopenia after univariate analysis. After multivariate analysis, a higher cancer stage (stage III vs. I: OR, 3.81; 95% CI, 2.05 to 7.09; p < 0.001) was found to be independently associated with baseline lymphopenia (Supplementary Table 1). Age at diagnosis, ECOG performance status, adjuvant chemotherapy administration, cancer stage, and ever neutropenia were associated with ever lymphopenia by univariate analysis; however, multivariate analysis revealed that only an older age at diagnosis (OR, 1.02; 95% CI, 1.01 to 1.04; p = 0.001), higher stage (stage III vs. I: OR, 3.01; 95% CI, 1.87 to 4.84; p < 0.001), and ever neutropenia (OR, 3.51; 95% CI, 2.08 to 5.93; p < 0.001) were independently associated with ever lymphopenia (Table 2). The incidence of ever lymphopenia in stage I, stage II, and stage III patients with gastric cancer was 19.0%, 34.3%, and 49.3%, respectively (Supplementary Table 2). Only higher cancer stage was positively associated with both baseline and ever lymphopenia.

On multivariable analysis, patients with ever lymphopenia had significantly higher OM than that of never lymphopenic patients (HR, 1.83; 95% CI, 1.17 to2.85; p = 0.008) (Table 3 and Fig. 1A). Other predictors of increased OM were higher stage (stage III vs. I: HR, 33.8; 95% CI, 17.1 to 66.7; p < 0.001) and older age at diagnosis (HR, 1.05; 95% CI, 1.02 to 1.07; p < 0.001). Receiving adjuvant chemotherapy was associated with lower OM (HR, 0.50; 95% CI, 0.29 to 0.88; p = 0.017). The 5-year OM rate among ever lymphopenic patients was 13.4%, whereas that among never lymphopenic patients was 4.4% (Fig. 1A).

Patients with ever lymphopenia also had significantly higher GCM (HR, 1.58; 95% CI, 0.96 to 2.60; p = 0.048) (Table 3 and Fig. 1B) compared to that of never lymphopenic patients. Higher stage (stage III vs. I: HR, 158.7; 95% CI, 55.5 to 471; p < 0.001) and older age at diagnosis (HR, 1.03: 95% CI, 1.00 to 1.05; p = 0.030) were also associated with higher GCM, whereas receiving adjuvant chemotherapy was associated with lower GCM (HR, 0.37; 95% CI, 0.20 to 0.66; p = 0.001). The 5-year GCM rate among ever lymphopenic patients was 10.9%, whereas that among never lymphopenic patients was 3.7% (Fig. 1B).

Univariate analysis showed that age at diagnosis, ECOG performance status, cancer stage, adjuvant chemotherapy administration, and ever lymphopenia were risk factors for gastric cancer recurrence. However, only higher stage (stage III vs. I: HR, 70.3; 95% CI, 33.7 to 146; p < 0.001) and ever lymphopenia (HR, 1.83; 95% CI, 1.18 to 2.81; p = 0.006) were associated with shorter DFS after multivariate analysis (Table 4 and Fig. 2A). In contrast to the univariate analysis, patients who received adjuvant chemotherapy achieved longer DFS (HR, 0.36; 95% CI, 0.21 to 0.63; p < 0.001) (Table 4) compared to those who did not. Analysis of the association between ever lymphopenia and DFS in patients stratified by cancer stage revealed that ever lymphopenia was associated with recurrence not only in those with stage I gastric cancer (stage I: HR, 3.26; 95% CI, 1.04 to 10.18; p = 0.042) (Fig. 2B), but also in those with stage II–III (HR, 1.84; 95% CI, 1.16 to 2.92; p = 0.009) (Fig. 2C). The 5-year cumulative recurrence rates were 15.1% and 4.6% for ever lymphopenic and never lymphopenic patients, respectively (Fig. 2A).