Preoperative lymphocyte to monocyte ratio and prognostic nutritional index were correlated with clinical outcomes in patients with operable colon cancer


 Background: Systemic inflammation plays an important part in tumorigenesis and progression. The predictive values of the preoperative lymphocyte to monocyte ratio (LMR) and prognostic nutritional index (PNI) in colon cancer remained unclear. Methods: A total of 308 patients with colon cancer undergoing radical resection were enrolled and analyzed. The receiver operating curves were applied to identify the thresholds for these biomarkers. Kaplan-Meier method and multivariate analysis were used to identify independent prognostic factors. Results: The univariate analysis showed that elevated LMR and PNI were significantly correlated with better overall survival and progression-free survival. The multivariate analysis showed that LMR and PNI were the independent prognostic factors for overall survival. Conclusions: Preoperative LMR and PNI could serve as useful prognostic factor in patients with colon cancer undergoing radical resection.


Introduction
With increasing incidence and mortality, colon cancer represented one of the most frequent cancer and the major cause of cancer-related mortality in the world [1,2]. Even though the great progress in surgical techniques and adjuvant chemotherapy in recent decades, the 5-year overall survival (OS) of colon cancer was still dissatisfactory in that a large proportion of patients with colon cancer would develop recurrent or metastatic diseases. Some parameters, such as the clinical stage, tumor differentiation, neural invasion, vascular invasion, could predict the prognosis of colon cancer [3]. Neveretheless, the prognosis of patients with the same parameter differed from individul to individual. Meanwhile, it was expensive, time consuming and inconvenient to detect these parameters. There was growing interest in the discovery of prognostic biomarkers contributing to optimize colon cancer management decision-making and improve clinical outcomes.
Systemic in ammation has been proved to play an important part in the tumorigenesis and tumor progression because of in ammatory mediators having the ability of accelerating vascular permeability, facilitating tumor cell in ltration via blood and lymphatic vessels [4][5][6]. Some biomarkers of peripheral blood examination, such as platelet to lymphocyte ratio (PLR) [7,8], neutrophil to lymphocyte ratio (NLR) [9][10][11][12], the Glasgow prognostic score (GPS) [13], prognostic nutritional index (PNI) [14], have been shown to re ect the degree of systemic in ammation. This theory has led to widely investigate peripheral blood examination in the hope of developing reliable, reproducible and low-cost prognostic factor. Among these biomarkers, PNI has been reported to be an independent prognostic factor in breast cancer, non-small cell lung cancer [14,15]. Peripheral lymphocyte to monocyte ratio (LMR), a new parameter of prognosis, has been drawing more and more attention [16][17][18]. Up to now, several studies have demonstrated that LMR and PNI were correlated with clinical outcomes in various malignancies, rare studies focused on colon cancer.
Proof for the predictive values of LMR and PNI in patients with colon cancer undergoing radical resection remained limited.
Given all this, the present study was designed to investigate the predictive values of preoperative LMR and PNI in patients with operable colon cancer. These results would facilitate further clinical research on systemic in ammation and optimize the treatment for colon cancer.

Patients
In the present retrospective study, all patients undergoing radical hemicolectomy were recruited at the department of gastrointestinal surgery, West China Hospital between February 2009 and March 2013. The study complied with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of West China Hospital of Sichuan University. The inclusion criteria included: (1) all patients with histologically con rmed colon cancer; (2) age 19-94 years; (3) without distant metastasis; (4) adequate bone marrow function (leukocyte count ranging from 4×10 9 /L to 10×10 9 /L); (5) survival time beyond three months. The exclusion criteria were as follows: (1) patients with in ammatory disease; (2)received preoperative therapy (chemotherapy or radiotherapy), or underwent emergency resection for perforation/obstruction; (3) patients with recurrent or multifocal lesions; (4) missing preoperative blood examination; (5) with other active malignancy; (6) follow up period not exceeding three months.

Peripheral blood parameters and patients' characteristics
Peripheral blood parameters including absolute lymphocyte count, absolute monocyte count and serum albumin level, were obtained within one week before resection. The LMR was calculated as peripheral absolute lymphocyte count divided by peripheral absolute monocyte count. The PNI was calculated as 0.05×peripheral absolute lymphocyte count (10 9 /L) + serum albumin levels (g/L). Patients' clinicopathological features, such as age, gender, the TNM stage, tumor differentiation, etc, were retrieved from the medical database. Follow-up visit was regularly performed until October 2017 or death. The de nition of the scoring system were as follow: patient with LMR≥3 and PNI≥44.2 was score 0, patient with LMR<3 or PNI<44.2 was score 1, patient with LMR<3 and PNI<44.2 was score 2.

Statistical analysis
The follow-up period was de ned as our previous study [19]. Progression-free survival (PFS) was calculated from the date of resection to the date of metastasis or recurrence. OS was from the date of surgery to the date of last follow-up or death by any causes. We performed the receiver operating curve (ROC) analysis to count the area under ROC curve (AUC) and then identify the optimal cut-off values of the LMR and PNI. The average was presented as mean ± standard deviation or median (quartile range). The relationships among the LMR, PNI and clinicopathological features were analyzed by the chisquare test or Fisher exact test. The Kaplan-Meier method with a log-rank test was employed to perform univariate analysis. Independent prognostic factors for OS and PFS were identi ed by using multivariate Cox-regression analysis with a forward conditional method. SPSS 19.0 software (SPSS, Inc., Chicago, IL, USA) was employed to perform all statistical analyses. We considered thevalue of two-tailed P<0.05 as statistically signi cant.

Results
A cohort of 364 patients with colon cancer were screened in the present study. Fifty-six of them were excluded because of distant metastasis (n = 39), multifocal lesions (n = 6), perioperative mortality (n = 1), received preoperative therapy (n = 3), loss to follow up (n = 7). After exclusion, 308 patients with stage I-III colon cancer were enrolled in the present study. Details about baseline patient characteristics and clinicopathological features were shown in Table 1. Of these, there were 173 males, and 135 females, 127 patients were under 60 years old, 181 patients were over 60 years old. The median age at the time of diagnosis was 62 years (range 19-94). The median follow-up period from resection was 49 months (range 4-69). The median value of PNI was 45.9 (42.0-49.8); LMR was 4.08 (3.16-5.77). During the follow-up period, 46 (14.9%) out of 308 patients developed metastatic disease, 40 (13.0%) were dead.
ROC analysis was performed to identify the optimal cutoff values of the LMR and PNI. The AUC in LMR and PNI for OS was 0.640 (P = 0.004), 0.651 (P = 0.002), respectively ( Figure 1). The AUC in LMR and PNI for PFS was 0.595 (P = 0.040), 0.604 (P = 0.025), respectively ( Figure 2). By using ROC analysis, the Youden indexes of LMR and PNI were 3.0 and 44.2 for OS, respectively.
We also performed subgroup analysis based on clinical stage to evaluate the predictive values of LMR, PNI and the scoring system. Among patients with stage I-II colon cancer, signi cant differences were observed in all subgroups ( Figure 4A-F). Among patients with stage III colon cancer, signi cant differences were observed in LMR for OS group and in the scoring system for OS and PFS groups ( Figure 5A-F).
The relationship among OS and clinicopathological features were shown in Table 2. In the univariate analysis, tumor location, clinical stage, lymph node involvement, differentiation, diabetes mellitus, LMR, PNI and the scoring system signi cantly correlated with the OS. Due to clinical stage deriving from distant metastasis, lymph node involvement and tumor depth, only clinical stage entered into the multivariate analysis. Multivariate analysis indicated that tumor location, differentiation, clinical stage, diabetes mellitus, LMR and PNI were independent prognostic factors for OS. The relationship among PFS and clinicopathological features were shown in Table 3. Tumor differentiation, lymph node involvement, clinical stage, diabetes mellitus, LMR, PNI and the scoring system were signi cantly correlated with the PFS in univariate analysis. Multivariate analysis showed that only tumor differentiation, clinical stage were the independent prognostic factors for PFS.

Discussion
In the nineteenth century, Rudolph Virchow rstly hypothesized the relationship among systemic in ammation and tumorigenesis. Until recently, cumulative evidences have demonstrated that systemic in ammation correlated with tumorigenesis and tumor progression because in ammatory cells and mediators could facilitate tumor proliferation, angiogenesis and metastasis [4,20]. Several systemic in ammatory parameters, such as NLR, PLR, GPS, could be used to predict clinical outcomes in various malignancies. LMR and PNI, as the new parameters, have earned great intereset recently [21][22][23]. The results of these studies concerning the relationship among LMR, PNI and clinical outcomes remained inconsistent. Therefore, we examined the cohort of patients with stage I-III colon cancer to identify the predictive values of LMR and PNI.
In the present study, we have demonstrated that elevated LMR and PNI were signi cantly correlated with favorable outcomes in operable colon cancer. LMR and PNI were independent predictors of OS among patients with colon cancer undergoing curative hemicolectomy, which were in consistent with previous studies [18,23]. Some other clinicopathological features, such as tumor location, differentiation, lymph node involvement, clinical stage, diabetes mellitus, still remained signi cance with OS.
Previous study by Li et al.
[10] demonstrated that elevated LMR was signi cantly correlated with favorable OS and PFS for operable colorectal cancer. In patients who received adjuvant chemotherapy, LMR was independent prognostic factor for OS and PFS. In patients without adjuvant chemotherapy, LMR lost their signi cances with OS and PFS. The study by Stotz et al. [24] examined a cohort of stage II-III colon cancer to identify the predictive value of LMR, which have shown that LMR was signi cantly correlated with OS and PFS. For subgroup analysis, there was signi cant relationship in patients with stage III colon cancer, no signi cant relationship in stage II colon cancer. Another study by Peng et al. [25] was designed to investigate the PNI in stage III colon cancer, which demonstrated that elevated PNI was positively correlated with elevated LMR. Compared with the elevated PNI group, clinical outcomes were dissatisfactory in the low PNI group among the entire study population. There were prominent relationship among LMR and clinical outcomes in stage IIIC colon cancer, but no prominent relationship in stage IIIA-B colon cancer. Another study also investigated the systemic in ammation in patients undergoing potentially curative hemicolectomy for stage I-III colon cancer, and demonstrated that right-sided tumor location was correlated with an elevated SIR [26].
The underlying mechanism of the relationship among the systemic in ammation parameters and clinical outcomes remained elusive. Several hypotheses could contribute to explain the relationship. Lymphocytes and monocytes, the major components of peripheral leukocyte population, could be recruited and migrated from peripheral circulation blood to tumor microenvironment, thereby representing the status of the systemic in ammation and host immune response to tumor burden. Elevated LMR indicated that lymphocytosis and monocytopenia. Previous studies have elucidated that lymphocytes were the essential component of adaptive and innate immunity, contributing to induce cytotoxic cell death and inhibit tumor progression and migration [27][28][29]. It was hypothesized that peripheral absolute lymphocyte count could re ect the degree of host immune response to tumor burden [30]. Lymphocytopenia marked as the increase in CD8+ suppressor lymphocytes and the decrease in CD4+ helper lymphocytes [31], was considered to be responsible for the weak, insu cient immunologic reaction against tumor, thereby representing the unfavorable prognosis in various malignancies. Lymphocytosis could lead to upregulate the immune response against tumor, thereby indicating the favorable prognosis. Peripheral circulating monocytes were recruited to the tumor microenvironment and differentiated into tumor-associated macrophages (TAMs) which have been classi ed as M1 or M2 phenotypes. The majority of TAMs have an M2 protumoral phenotype, contributing to promote tumor progression, angiogenesis by releasing some cytokines and chemokines [32]. Increased peripheral monocyte counts could re ect the increased levels of TAMs and lead to upregulate the immunosupression, thereby indicating unfavorable prognosis. Our previous studies also demonstrated that peripheral lymphocytosis and monocytopenia predicted favorable outcomes in colon cancer [19,33].
Serum albumin produced in the liver was the largest quantity of blood plasma protein, playing an important part in physiological function, such as maintaining the colloid osmotic pressure, transporter of various substances, detoxi er and antioxidant. Several studies demonstrated that malnutrition indicated an increased risk of postoperative morbidity and mortality. The main characteristics of malnutrition were hypoalbuminemia and lymphocytopenia. Thus, serum albumin was one of the most important indicator for nutritional status, which has been proved to be an independent prognostic indicator in various malignancies, such as gastric cancer [ It must be acknowledged that the present study have several limitations. First, our study was a retrospective single-center cohort study with a small number of patients. Second, other medical conditions, such as infection, ischemia, postoperative adjuvant chemotherapy, which might potentially affect the clinical outcomes or peripheral leukocyte count, were not evaluated in this study. Therefore, selection bias could not be entirely excluded in the study. Third, the cutoff values of LMR and PNI were calculated by the method of Youden index. Our cutoff values were inconsistent with previous study. Overall, further larger prospective studies should be conducted to verify the results. The strength of the study included that the study only recruited operable colon cancer and we comprehensively compared the LMR and PNI.

Conclusions
The present study demonstrated that preoperative PNI and LMR were independent prognostic factors for OS in patients with stage I-III colon cancer undergoing curative hemicolectomy. Patients with elevated LMR and PNI predicted favorable clinical outcomes. On the basis of LMR and PNI, patients with high risk was required to receive intensive postoperative treatment for preventing recurrence or metastasis.   Figure 1 The receiver operating curve analysis of preoperative LMR and PNI according to overall survival.

Figures
Page 15/18 Figure 2 The receiver operating curve analysis of preoperative LMR and PNI according to progression-free survival. Kaplan-Meier curves of LMR, PNI and the scoring system for overall survival and progression-free survival in patients with stage I-III colon cancer. Note: A and B are the survival curves of LMR for overall survival and progression-free survival; C and D are the survival curves of PNI for overall survival and progression-free survival; E and F are the survival curves of the scoring system for overall survival and progression-free survival. Kaplan-Meier curves of LMR, PNI and the scoring system for overall survival and progression-free survival in patients with stage I-II colon cancer. Note: A and B are the survival curves of LMR for overall survival and progression-free survival; C and D are the survival curves of PNI for overall survival and progression-free survival; E and F are the survival curves of the scoring system for overall survival and progression-free survival. Kaplan-Meier curves of LMR, PNI and the scoring system for overall survival and progression-free survival in patients with stage III colon cancer. Note: A and B are the survival curves of LMR for overall survival and progression-free survival; C and D are the survival curves of PNI for overall survival and progression-free survival; E and F are the survival curves of the scoring system for overall survival and progression-free survival.