The clinical influence of the preoperative lymphocyte‐to‐monocyte ratio on the postoperative outcome of patients with early‐stage gastrointestinal cancer

Abstract Aim The lymphocyte‐to‐monocyte ratio (LMR) is useful for predicting the prognosis of patients with gastric cancer (GC) and those with colorectal cancer (CRC) undergoing surgery. The relationship between the LMR and postoperative outcome of patients with early‐stage gastrointestinal cancers such as stage I GC and CRC remains unclear. Methods We retrospectively evaluated 323 stage I GC and 152 stage I CRC patients undergoing surgery. Univariate and multivariate analyses using the Cox proportional hazards model were performed to identify the clinical characteristics associated with overall survival (OS), and the cut‐off values of these variables were determined by receiver operating characteristic analysis. The Kaplan–Meier method and log‐rank test were used for postoperative survival comparisons according to the LMR (GC: LMR < 4.2 vs ≥4.2; CRC: LMR < 3.0 vs ≥3.0). Results Univariate and multivariate analyses revealed that OS was significantly associated with the LMR (<4.2/≥4.2) (HR, 2.489; 95% CI, 1.317‐4.702; P = 0.005), as well as age (>75/≤75 years) (HR, 3.511; 95% CI, 1.881‐6.551; P < 0.001) and albumin level (≤3.5/>3.5 g/dL) (HR, 3.040; 95% CI, 1.575‐5.869; P = 0.001), in stage I GC patients. Survival analysis demonstrated a significantly poorer OS in stage I GC patients with a LMR < 4.2 compared with ≥4.2 (P < 0.001). In stage I CRC patients, despite a significant difference in OS according to the LMR (<3.0 vs ≥3.0) (P = 0.040), univariate analysis revealed no significant association between the LMR and OS. Conclusion LMR is a useful predictor of the postoperative outcome of stage I GC patients treated surgically.


| INTRODUC TI ON
Although the 5-year overall survival (OS) rate after surgery in patients with stage I gastric cancer (GC) or colorectal cancer (CRC) is >90%, some patients have poor postoperative outcomes due to recurrence or other diseases. 1,2 Several studies have revealed that a high age, elevated tumor marker levels, lymphovascular invasion, and male sex are associated with poor postoperative outcomes in patients with stage I GC or CRC. 1,2 Therefore, predicting postoperative outcomes is important for appropriate postoperative follow-up of such patients.
During the last decade, many blood-cell-based prognostic systems have been reported as useful for predicting the prognosis of GC and CRC patients. 3,4 For example, the neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio are blood-cell-based prognostic markers for cancer patients. Although the mechanism underlying how these prognostic markers is associated with the prognosis of cancer patients is still unclear, it was previously reported that these markers are associated with cancer-related inflammation and a tumor microenvironment favoring tumor progression. 5 Recently, a low peripheral blood lymphocyte-to-monocyte ratio (LMR) was reported to be significantly associated with a poor prognosis, including tumor progression and distant metastasis, in patients with GC or CRC. [6][7][8][9] Additional reports showed that the pretreatment LMR predicts the prognosis of early-stage cancer patients. [10][11][12] These findings suggest that the LMR is associated with postoperative outcomes in patients with both stage I GC and CRC.
However, the relationship between the LMR and postoperative outcome in patients with early-stage gastrointestinal cancer remains unclear. Herein, we investigated the relationship between the LMR and postoperative outcomes in both patients with stage I GC and stage I CRC using the database from a single institution. Research of the Ministry of Health, Labour and Welfare in Japan (http://www.mhlw.go.jp/seisa kunit suite /bunya /hokab unya/kenky ujigy ou/i-kenky u/index.html).

| Definition of GC tumor location
Based on the General Rules for Japanese Classification of Gastric Carcinoma (Japanese Gastric Cancer Association, 3rd English Edition), the stomach is anatomically divided into three portions (upper, middle, and lower) delineated by the lines connecting the trisected points on the lesser and greater curvatures. If the tumor involves more than one stomach portion, all involved portions are recorded in descending order of the degree of involvement, e.g., lower, middle or upper, middle, lower. 13

| Statistical analysis
Data are presented as medians with interquartile ranges. Intergroup differences were analyzed using the chi-squared test or Mann-Whitney U test, as appropriate. Clinical factors closely related to OS were identified by univariate and multivariate analyses using the Cox proportional hazards model, with calculation of the hazard ratio (HR) and 95% confidence interval (CI). The Kaplan-Meier method and log-rank test were used to compare postoperative OS according to the LMR in the GC patients (LMR < 4.2 vs ≥4.2) and CRC patients (LMR < 3.0 vs ≥3.0). All statistical analyses were performed using SPSS software (version 25.0; IBM Co., New York, NY, USA), and differences with a P-value < 0.05 were considered statistically significant.
The cut-off values of the various clinical characteristics evaluated were determined using receiver operating characteristic (ROC) analysis, defined according to the most prominent point on the ROC curve (Youden index = maximum [sensitivity -(1 -specificity)]). We also calculated the area under the ROC curve. 14 The optimal cutoff LMR for stage I GC and stage I CRC patients were 4.2 and 3.0, which had sensitivities of 66.3% and 86.9%, specificities of 70.0% and 33.3%, and areas under the ROC curve of 0.673 and 0.610, respectively ( Figure 1). Excluding serum levels of carbohydrate antigen 19-9 (U/mL), carcinoembryonic antigen (ng/mL), and C-reactive protein (CRP; mg/dL), cut-off values for other variables, such as age (years), body mass index (kg/m 2 ), maximum tumor size (cm), platelet count (×10 4 /mm 3 ), serum level of albumin (g/dL), and white blood cell count (×10 3 /mm 3 ) were also calculated using ROC analyses.

| Clinical characteristics of stage I GC patients
Of the 323 stage I GC patients (217 males and 106 females) enrolled in this study, 197 had a high LMR (≥4.2) and 126 a low LMR (<4.2). Table 1A shows the clinical characteristics of the stage I GC patients according to the LMR. There were significant differences in age, serum levels of albumin (g/dL), carcinoembryonic antigen (ng/mL) and CRP (mg/dL), sex,

| Clinical characteristics of stage I CRC patients
Of the 152 stage I CRC patients (93 males and 59 females) enrolled in this study, 130 had a high LMR (≥3.0) and 22 had a low LMR (<3.0).

| Postoperative death and recurrence in stage I GC patients
During the observation period, 50 of the GC patients died, including 14 cancer-related deaths (Table 2A,

| Postoperative death and recurrence in stage I CRC patients
During the observation period, 15 of the stage I CRC patients died, including five cancer-related deaths: one from CRC and four from other cancer types (Table 3A,B). Among the non-cancer-related deaths, two patients died of infectious disease, and one patient each died of cerebrovascular disease, liver failure, heart disease, hypoglycemia, hypoxemia, old age, hematemesis, and unknown causes.

| Survival of stage I GC patients
The median and maximum follow-up periods of the surviving patients with stage I GC were 1905 and 5844 days, respectively, with a mean OS of 2025 ± 1393 days. The Kaplan-Meier method and log-rank test revealed a significant difference in OS according to the LMR (≥4.2 vs <4.2) (Figure 2A).

| Survival of stage I CRC patients
The median and maximum follow-up periods of the surviving patients with stage I CRC were 1864 and 6009 days, respectively, with a mean OS of 2213 ± 1344 days. The Kaplan-Meier method and log-rank test revealed a significant difference in OS according to the LMR (≥3.0 vs <3.0) ( Figure 2B).

| Postoperative incidence of infectious diseases in stage I GC patients
During the observation period, 62 GC patients had incidence of infectious diseases. Among the 62 patients, 30 had pneumonia, six had cholecystitis, five had cholangitis, five had shingles, five had skin infection, three had pancreatitis, three had urinary tract infection, one had diverticulitis, one had intra-abdominal hemorrhage, one had peritonitis, one had spondylitis, and one had sinusitis, respectively.
The Kaplan-Meier method and log-rank test revealed a significant difference between the two groups according to the LMR (≥4.2 vs <4.2) in incidence of infectious diseases ( Figure 3A).

| Postoperative incidence of infectious diseases in stage I CRC patients
During the observation period, 30  to the LMR (≥3.0 vs <3.0) in incidence of infectious diseases ( Figure 3B).

| Univariate and multivariate analyses of OS in stage I GC patients
Univariate analyses conducted in the stage I GC patients revealed as-

| Univariate and multivariate analyses of OS in stage I CRC patients
Univariate analyses among the stage I CRC patients revealed that OS was not significantly associated with the LMR (<3.0/≥3.0), but

| D ISCUSS I ON
Consistent with previous studies, [10][11][12] we found that a low LMR was significantly associated with poor prognosis in patients with earlystage gastrointestinal cancers (i.e., GC and CRC) ( Figure 2). However, very few patients died of GC or CRC (Table 2A, B and Table 3A, B), indicating that the LMR is associated with other causes of death after surgery.
To emphasize the usefulness of LMR in patients with early-stage gastrointestinal cancer, we compared LMR with the conventional inflammation-based prognostic score, Glasgow prognostic score (GPS) in prognostication of such patients. Multivariate analyses revealed that GPS was not significantly associated with OS in both stage I GC and stage I CRC patients. A previous study showed that GPS was not good at prognostication of early-stage cancer patients, because most patients with early-stage cancer did not have cancer cachexia due to tumor progression. 15 These facts suggest that LMR is superior to GPS in predicting non-cancer-related death after surgery in stage I GC patients.
Our results revealed that a low LMR was significantly associated with older age, hypoalbuminemia, a high serum CRP level, and male sex in patients with early-stage gastrointestinal cancers (Table 1A,  responses are affected by aging. 16 Regarding hypoalbuminemia and high serum CRP levels, recent studies revealed that these characteristics are associated with immunosuppression and malnutrition in cancer patients. 17,18 All of these findings support that a low LMR reflects immunosuppression due to high age and malnutrition.
A previous study revealed that the LMR was significantly associated with the incidences of infectious diseases, such as pneumonia and urinary tract infections, in patients with acute ischemic stroke. 19 The authors suggested that the LMR might reflect immunosuppression induced by stroke, and in turn, the immunosuppression is the cause of infectious diseases. 19 Therefore, the LMR is useful for predicting the outcome of not only patients with cancer but also those with heart or vascular disease. 20 The preoperative LMR might be useful for prognostication in stage I GC patients, because GC is associated with postoperative weight loss. Unlike in CRC patients, gastric storage dysfunction, reduced ghrelin levels, and digestion/absorption disorders lead to postoperative weight loss in GC patients. [22][23][24] According to recent studies, being underweight is associated with increased incidences of stroke, atrial fibrosis, and impaired endothelial dysfunction, [25][26][27] as well as an increased risk of pneumonia. 28 Thus, the combination of postoperative weight loss and a low LMR might increase the risks of other diseases, leading to a worse postoperative outcome in stage I GC patients.
Recent studies showed that oral nutritional supplements significantly improved postoperative weight loss in GC patients. 29,30 In the same way, another study showed that exercise interventions prevented postoperative muscle loss in GC patients. 31 In addition, exercise interventions prevented not only incidence of cancer and cardiovascular disease but also all-causes of mortality. 32,33 Therefore, in order to prevent non-cancer-related death, both nutritional supplements and exercise interventions would be needed in GC patients with low LMR (<4.2).
There were some limitations to our study. First, this was a retrospective study conducted at a single institution. Second, the population of stage I CRC patients in this study was relatively small (n = 152). To overcome these limitations, validation of our results in multi-institutional studies with larger sample sizes is needed.
In conclusion, the present findings indicated a relationship between the preoperative LMR and the outcome of patients with early-stage gastrointestinal cancer. The novelty of the study is that LMR could predict not only primary cancer death but also non-cancer-related death due to infectious and vascular diseases. Based on these results, the LMR could be considered a factor determining both nutritional supplements and exercise interventions for such patients.

D I SCLOS U R E
Conflicts of Interest: The authors declare no conflicts of interest regarding the publication of this paper.