1 INTRODUCTION

Cholangiocarcinoma (CCA), including intrahepatic and extrahepatic cholangiocarcinoma, is a kind of epithelial malignant tumor originating in the biliary tract and is a highly invasive biliary tract tumor with poor survival.¹ Overall, CCA constitutes 3% of all gastrointestinal tumors, representing the second most frequently diagnosed primary liver cancer following hepatocellular carcinoma (HCC)² and the overall incidence of CCA has been increasing in countries around the world,^{3, 4} and the 5-year overall survival (OS) rate of CAA was <5%. Surgery and liver transplantation may be potential treatments for patients with CCA. The current study does not recommend liver transplantation as a first choice because of its technical difficulty and high cost. Despite the changes in surgical methods and techniques, the resection rate of patients with cholangiocarcinoma is still very low. Postoperative recurrence and metastasis are the main causes of poor prognosis. In addition, the discovery of effective blood prognostic signatures for recurrence and survival in CCA patients after hepatectomy remains an unmet need.

Recent studies have shown that inflammation is considered to be a hallmark feature of cancer development and progression, and that proinflammatory cytokines and chemokines in the cancer microenvironment lead to the survival and proliferation of tumor cells, metastasis, angiogenesis, and destruction of adaptive immunity, thereby affecting survival and prognosis.⁵ On the one hand, chronic inflammation triggers the local accumulation of monocytes, platelets, and neutrophils, secreting cytokines and inflammatory factors to induce tumor angiogenesis and metastasis. On the other hand, the increase of monocytes and lymphocytes will enhance the ability to resist tumor invasion.⁶ The impact of inflammatory-based scores on cancer therapy outcomes has been extensively studied. Previous studies found that there was a prognostic value for inflammation biomarkers such as the radio among neutrophils, lymphocytes, C-reactive protein (CRP), and platelet.^7-9 Previous studies have shown that intrahepatic cholangiocarcinoma patients with a low level of NLR (neutrophil to lymphocyte ratio) and PLR (platelet to lymphocyte ratio) had longer OS than those with high levels, suggesting that low levels of NLR and PLR were associated with poorer prognosis.⁹ Lymphocyte to monocyte ratio (LMR), a relatively new inflammatory-related score and a biomarker associated with translational inflammatory response, has been exhibited to have potential prognostic value in patients with kinds of tumors, such as lymphoma,¹⁰ colorectal cancer,¹¹ and lung cancer.¹² Known risk factors for CCA, such as liver fluke, liver calculus, and primary sclerosing cholangitis, can cause chronic inflammation of the liver, suggesting that inflammation is closely related to the carcinogenesis of CCA.^{13, 14} Furthermore, Studies showed that preoperative LMR is a very significant indicator that can be used to indicate whether patients with cholangiocarcinoma are resectable, and dynamic changes of LMR can accurately predict early recurrence in patients with advanced curable cholangiocarcinoma.¹⁵ However, the prognostic and survival significance of preoperative LMR for patients with postoperative CCA has not been reported in detail.

In this study, we aimed to assess the prognostic value of preoperative LMR in patients with CCA following hepatic resection and analyze risk factors for the overall survival of CCA patients with hepatic resection via using receiver-operating characteristic (ROC) curve and establishing Cox regression analysis model.

2 MATERIALS AND METHODS

3 RESULTS

3.1 Kaplan–Meier analysis curves for OS and ROC curve analysis of the LMR in cholangiocarcinoma patients

In this part, we respectively analyzed the OSof 145 cholangiocarcinoma patients who were newly treated and underwent surgical treatment at the Fudan University Shanghai Cancer Center from August 2008 to October 2020. As shown in Figure 1A, we used the Kaplan–Meier method to analyze the OS of 145 cholangiocarcinoma patients. Furthermore, we used the ROC curve to explore the diagnosis value and the optimum truncation value of preoperative LMR. The results as shown in Figure 1B, indicated that the optimal cut-off value of LMR was 3.225, and the area under the curve (AUC) was 0.599 (95% CI 0.498–0.700), p = 0.056. We so divided these patients into two groups according to the optimal cut-off of LMR: the low LMR group (LMR≤3.225) and high LMR group (LMR >3.225). The results of OS via the Kaplan–Meier analysis in the two groups showed that there was a longer OS in cholangiocarcinoma patients of the high LMR (LMR >3.225) group than the low LMR group (LMR≤3.225) (long-rank chi-square = 10.425, p = 0.001) (Figure 1C). The results of the risk analysis function showed that the patients in the low LMR group exhibited a high risk, while the patients in the high LMR group exhibited a significantly low risk, and the p value was statistical significance after the comparison between the two groups (Figure 1D).

3.3 Univariate regression analysis for overall survival in cholangiocarcinoma patients

In order to further the prognostic factor for OS of cholangiocarcinoma patients, we used the Cox univariate regression model to analyze the effects of clinical characteristics on OS, and results showed that the levels of CA199 and CEA, TNM stage, nerve invasion, postoperative treatment, preoperative cholangitis, pathological type, AST, ALB, and LMR all significantly affected the OS (Table 3). As shown in Figure 2, we found that the OS of patients with elevated levels of CA199, CEA, and AST were shorter than patients with normal levels; however, there was a longer OS in the LMR high group than in the LMR low group. We also found that preoperative cholangitis, low albumin expression, staging (III-IV), and nerve invasion were obviously associated with the shortened OS, while postoperative treatment was related to the longer OS.

TABLE 3. Univariate analysis for overall survival in cholangiocarcinoma patients.

Variable	Number	Relative risk	95.0% Confidence interval		p-Value
			Inferior limit	Upper limit
Age (year)
≤55	48	1
>55	97	0.782	0.511	1.197	0.257
Sex
Male	78	1
Female	67	0.778	0.524	1.154	0.212
HBV infection
No	103	1
Yes	42	1.212	0.772	1.901	0.403
CA199 (ng/mL)
<27	35	1
≥27	110	2.053	1.214	3.471	0.007
AFP (ng/mL)
<10	126	1
≥10	19	1.399	0.793	2.470	0.246
CEA (ng/mL)
<5.2	84	1
≥5.2	61	2.630	1.731	3.996	<0.0001
TNM stage
I + II	56	1
III+IV	89	1.987	1.280	3.084	0.002
Nerve invasion
No	94	1
Yes	51	1.611	1.062	2.445	0.025
Surgical margin
No	136	1
Yes	9	1.695	0.852	3.372	0.133
Lymphatic metastasis
No	116	1
Yes	39	0.874	0.568	1.343	0.539
Bile duct infringement
No	124	1
Yes	21	0.934	0.520	1.677	0.819
Vascular invasion
No	55	1
Yes	90	1.196	0.786	1.819	0.404
Tumor diameter (cm)
≤2	22	1
>2	123	0.849	0.480	1.500	0.573
Tumor number
≤1	127	1
>1	18	1.213	0.645	2.279	0.549
Satellite focal
No	120	1
Yes	25	1.100	0.651	1.857	0.722
Preoperative treatment
No	19	1
Yes	126	1.183	0.614	2.280	0.615
Postoperative treatment
No	60	1
Yes	85	0.633	0.425	0.944	0.025
Pathological differentiation
high	10	1			0.827
moderately	72	0.926	0.423	2.026	0.847
poorly	63	0.878	0.582	1.326	0.538
Tumor location
Left	50	1			0.613
Right	80	0.975	0.494	1.923	0.941
porta hepatis	15	0.796	0.394	1.605	0.523
Preoperative cholangitis
No	95	1
Yes	50	2.361	1.550	3.595	<0.0001
Pathological type
ICC	89	1
EHCC	56	1.591	1.065	2.378	0.023
ALT
≤50	112	1
>50	33	1.322	0.837	2.089	0.232
AST
≤40	107	1
>40	38	1.780	1.161	2.729	0.008
ALB
≤35	6	1
>35	139	0.401	0.162	0.990	0.047
TBIL
≤26	132	1
>26	13	1.781	0.947	3.348	0.073
LMR
≤3.25	61	1
>3.25	84	0.520	0.347	0.779	0.002

• Abbreviations: ALB, albumin; AFP, alpha fetal protein; ALT, glutamic-pyruvic transaminase; AST, glutamic oxalacetic transaminase; CA199, carbohydrate antigen 19–9; CEA, carcino-embryonic antigen; EHCC, extrahepatic cholangiocarcinoma; ICC, intrahepatic cholangiocarcinoma; LMR, lymphocyte to monocyte ratio; TBIL, total bilirubin.
• Bold indicates significance level at p < 0.05.

3.4 Multiple regression analysis for overall survival in cholangiocarcinoma patients

To further illustrate the influencing factors for OS of cholangiocarcinoma patients, we used the Cox multiple regression model to analyze the significant index that was statistically significant in univariate regression model analysis. As shown in Table 4, we found that preoperative cholangitis (p = 0.001, HR = 2.128, 95% CI (1.371–3.301)), elevated CEA level (p < 0.0001, HR = 2.313, 95% CI (1.488–3.594)), and nerve invasion (p = 0.038, HR = 1.596, 95% CI (1.027–2.481)) were risk factors for the OS of cholangiocarcinoma patients, while the high LMR level (p = 0.003, HR = 0.517, 95% CI (0.336–0.796)) and postoperative treatment (p = 0.009, HR = 0.582, 95% CI (0.387–0.875)) were protective factors for the OS of cholangiocarcinoma patients. As shown in Figure 3, the results of multiple regression analysis also indicated that the OS of patients with elevated level CEA were shorter than patients with normal level; however, there was a longer OS in the LMR high group than in the LMR low group. We also found that nerve invasion was obviously associated with the shorter OS, while postoperative treatment was related to the longer OS.

TABLE 4. Multivariate analysis for overall survival in cholangiocarcinoma patients

Variable	Number	Relative risk	95.0% Confidence interval		P value
			Inferior limit	Upper limit
TNM stage
I + II	56	1
III+IV	89	1.109	0.692	1.779	0.666
Postoperative treatment
No	60	1
Yes	85	0.601	0.398	0.908	0.016
CA199 (ng/mL)
<27	35	1
≥27	110	1.132	0.635	2.020	0.674
CEA (ng/m)
<5.2	84	1
≥5.2	61	2.313	1.488	3.594	<0.0001
Nerve invasion
No	94	1
Yes	51	1.596	1.027	2.481	0.038
Preoperative cholangitis
No	95	1
Yes	50	2.128	1.371	3.301	0.001
AST
≤40	107	1
>40	38	1.523	0.979	2.369	0.062
ALB
≤35	6	1
>35	139	0.758	0.278	2.066	0.588
LMR
LMR ≤3.225	61	1
LMR >3.225	84	0.517	0.336	0.796	0.003

• Abbreviations: ALB, albumin; AFP, alpha fetal protein; AST, glutamic oxalacetic transaminase; CA199, carbohydrate antigen 19–9; CEA, carcino-embryonic antigen; LMR, lymphocyte to monocyte ratio.
• Bold indicates significance level at p < 0.05.

3.5 Creation and use of the nomogram

In order to further evaluate the factors that significantly influenced the overall survival of CCA patients in the multivariate regression model, we constructed a Nomogram diagram to show the risk coefficient of some indexes, including TNM stage, the preoperative level of CA199 and CEA, nerve invasion, and the level of LMR. This compiled nomogram allowed for the estimation of the risk of OS and the prediction of median survival time and survival probability for each CCA patient (Figure 4). The steps to use the Nomogram are as follows: (1) determine the patient's value for each predictor, (2) draw a line up from each predictor to the vertex reference line, (3) add the points from each predictor, (4) locate the sum at the total point reference line, (5) draw a line down from the total point line, the median survival probability and 1-, 3-, 5-year survival rates were obtained.

4 DISCUSSION

Cholangiocarcinoma (CCA) is a malignant tumor with a very poor prognosis. In many countries, the incidence, recurrence, and mortality rates of CCA remain high. Surgery is the only effective method; However, postoperative recurrence and metastasis significantly affect the prognosis of CCA patients. In addition, there is still a lack of effective markers to predict the prognosis of patients after surgery, which further provides direction for the treatment of patients after surgery. Previous studies exhibited that the OS and DFS (disease-free survival) of CCA patients with low preoperative PLR (platelet to lymphocyte ratio) were significantly shorter than those of the high PLR group, indicating that the low level of PLR was associated with poor prognosis.¹⁶ In this study, we found that a new indicator, LMR, served as a prognostic biomarker for CCA patients with hepatic resection.

More and more studies have demonstrated that inflammatory indexes in peripheral blood and proinflammatory factors and chemokines in the immune microenvironment have significant effects on the prognosis of tumor patients.^17-19 The inflammatory cells filling in tumor microenvironment are indispensable in tumor initiation, angiogenesis, invasion, and migration. Inflammatory cells and tumor cells used epithelial-to-mesenchymal transition (EMT) as a bridge to interact with each other at multiple levels and continuously, thus affecting the survival and prognosis of tumor patients.²⁰ Various markers of systemic inflammation, including cytokines, C-reactive protein, and monocyte, neutrophil, or lymphocyte count, as well as their ratios such as NLR (neutrophil/lymphocyte)²¹ and LCR (lymphocyte/C-reactive protein),²² have been investigated for their prognostic roles in certain tumor populations. Furthermore, LMR, a biomarker of tumor inflammation and host immunity, was also associated with decreased mortality in tumors. As we all know, lymphocyte was a vital biomarker of host immune status, and lymphocytopenia was associated with lower survival rates for various cancers.^{23, 24} Monocyte was considered as surrogate markers of the tumor microenvironment. Recent studies have been demonstrated that LMR could serve as an effective marker for predicting the response of treatment and prognosis in lymphoma²⁵ and tongue cancer patients.²⁶ LMR could also predict survival in patients with advanced esophageal cancer who receive concurrent chemoradiotherapy, and low levels of LMR were significantly associated with poor prognosis in patients with esophageal cancer.²⁷ In the present study, we found that LMR was also a prognostic marker that can be used to predict the overall survival of postoperative patients with CCA. The results of our study showed that the high LMR level (p = 0.023, HR = 0.609, 95% CI (0.397–0.935)) was a protective factor for the OS of CCA patients, and the OS of the low LMR group (10 months, 95% CI 6.766–13.234) was markedly lower (p = 0.001) than the high LMR group (20.23 months, 95% CI 14.706–25.754). We further employed the chi-square test to analyze the correlation between LMR and clinical characteristics, and the results showed that the level of CA199 and CEA, TNM stage, and tumor number as statistically significant factors related to the LMR of CCA patients. Meanwhile, multivariate analysis of results indicated that the preoperative level of LMR and CEA, nerve invasion, and postoperative treatment independent prognostic factors for patients with CCA. Form our results in the previous part, we found that the preoperative level of LMR and CEA both predicted the OS of CCA patients with hepatic resection, and it was of great significance to judge the prognosis of patients with CCA in advance and to positively intervene in early treatment. Whether there was nerve invasion indicated by postoperative pathology was also a key factor affecting prognosis, and timely and postoperative treatment could also prolong the overall survival. Furthermore, unlike previous studies, we firstly created a nomogram figure to show the contribution of significant variables for overall survival in multivariate regression analysis, and to predict median survival time and 1-, 3-, and 5-year survival rates for patients with cholangiocarcinoma.

The mechanism of decreased LMR in preoperative patients and poor prognosis was still unclear, of which the association of LMR with inflammation in the tumor microenvironment might be significant. A plausible explanation for this phenomenon was that the decreased LMR may reflect impaired immune function, as it was associated with lymphocytopenia and monocyte proliferation. Previous studies have shown that preoperative lymphocyte count was a good indicator of tumor prognosis and played a key role in the host's cytotoxic immune response to tumor, which was considered to reflect the general state of immune function, suggesting that lymphocyte-mediated cytotoxicity could lead to the release of cytokines and inhibit the growth and metastasis of cancer cells.²⁸ Lymphocytes could be divided into T cells, B cells, natural killer cells, etc., which played a key role in the anticancer reaction of the immune system by contributing to cancer cell apoptosis. Therefore, lymphopenia is supposed to be a surrogate signature of host immunological incompetence. Recent studies exhibited that monocyte has also been reported to be related to poor prognosis in various tumors, mainly because of tumor-associated macrophages (TAMs).²⁹ TAMs mainly came from circulating monocytes. There was increasing evidence that TAMs have major proto-oncogenic activities, including promoting metastasis, immunosuppression, and tumor angiogenesis.³⁰ Thus, it could be seen that lymphocytes and monocytes were involved in the anti-cancer immune response and the formation of cancer microenvironment, thereby influencing the prognosis of CCA patients.

However, our study is only a single-center retrospective study, and, due to a large number of patients with locally advanced stages in our included cases, our retrospective analysis of overall patient survival was lower than the data reported in previous studies. The reasons that tumor factors such as lymphatic metastasis and tumor diameter are not commonly reported as prognostic factors in this study are as follows: on the one hand, the sample may be small; on the other hand, confounding factors may interfere with the results. For example, patients with large-diameter tumors did not have lymph node metastases; Patients with small tumor diameters existed in metastasis. In addition, although there was a significant difference in OS among patients with cholangiocarcinoma in the low and low LMR group, the AUC value (AUC = 0.599, 95% CI 0.498–0.700, p = 0.056) in our study was not high and there was no convincing result. Therefore, our result was only a preliminary study, and we need further study to demonstrate the effects of LMR on the OS of patients with cholangiocarcinoma. In the future, we still need the participation of multiple centers to further expand the sample size for verifying our conclusions. At the same time, more basic experiments including cell culture and animal model construction are needed to verify our findings at multiple levels.

In summary, this study was the first retrospective analysis of the correlation between preoperative LMR and OS in CCA patients. We found that there was a longer OS in the CCA patients with high LMR and normal CEA level, no nerve invasion, and postoperative treatment. We firstly created a nomogram plot to exhibit the contribution of significant variables for OS in multivariate regression analysis, and to predict survival time and rate, thereby providing a new sight for CCA patients with hepatic resection.

AUTHOR CONTRIBUTIONS

Minghe Lv: Conceptualization (lead); data curation (lead); formal analysis (lead); investigation (lead); methodology (lead); software (lead); validation (lead); visualization (lead); writing – original draft (lead); writing – review and editing (lead). Kun Wang: Data curation (supporting); formal analysis (supporting); methodology (supporting). Zhiyuan Zhang: Formal analysis (supporting); validation (supporting). Zhen Zhang: Conceptualization (supporting); funding acquisition (equal); project administration (supporting); writing – review and editing (supporting). Juefeng Wan: Conceptualization (equal); funding acquisition (equal); project administration (lead); writing – review and editing (supporting).

FUNDING INFORMATION

This work was supported by the Shanghai Anti-Cancer Association (Grant Number: SACA-AX201902) and Shenkang—Three-year action plan clinical study (Grant Number: SHDC2020CR3082B).

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no competing interests.

ETHICAL APPROVAL STATEMENT

The study was approved by the ethics committee of Fudan University Shanghai Cancer Center and the ethical code is 2003215-1. Informed consent was obtained for all patients.