Hepatocellular carcinoma (HCC) is one of the most common malignancies in the world and the third leading cause of tumor death1,2,3,4. Although hepatectomy has achieved significant improvements, the prognosis of patients with liver cancer is far from satisfactory, due to distant metastasis and tumor recurrence5,6,7,8. Recently, the use of inflammatory and molecular biomarkers to improve the prognosis of HCC has received increasing attention. However, there is still a lack of reliable, low-cost tumor markers that can be easily applied to clinic to predict prognosis.

There is more and more consensus that inflammation is associated with the development of malignant tumors and that persistent systemic inflammatory responses is involved with poor prognosis in many cancers9,10,11,12,13. One widely studied inflammatory markers originated from full blood count are the lymphocyte-to-monocyte ratio (LMR), the neutrophil-to-lymphocyte ratio (NLR) and the platelet- to lymphocyte ratio (PLR), which have been identified as prognostic biomarker in HCC7,14,15,16.

However, there are few studies to compare the ability of the LMR, NLR and PLR to predict long-term survival among patients with HCC undergoing liver resection with curative intent. Although Yang YT et al. have shown that lymphocyte to monocyte ratio and neutrophil to lymphocyte ratio are superior inflammation-based predictors of recurrence in patients with hepatocellular carcinoma after hepatic resection, they did not conduct survival analysis of over survival(OS)16.

In light of these, the present study aimed to compare the relative prognostic value of the preoperative LMR, NLR and PLR in disease-free survival(DFS) and OS in HCC patients with an attempt to clarify the optimal use of these markers.

Materials and Methods

This research was approved by the Ethics Committee of the First People’s Hospital of Changde and the Tumor Hospital of Guangxi Medical University, and written informed consent was obtained from patients prior to surgery. All treatments were performed in accordance with relevant guidelines and regulations.


All HCC patients underwent hepatectomy with curative intent at the First People’s Hospital of Changde City and the Affiliated Tumor Hospital of Guangxi Medical University between April 2004 and April 2012 were included in our study. The inclusion criteria were: no treatment for HCC before hepatectomy; initial hepatectomy with curative intent performed at the authors’ center; no other simultaneous malignancies; no coexistent hematologic disorders and no preoperative fever, so that the preoperative platelet count, neutrophil, lymphocyte and monocyte reflected normal baseline values; and no renal, cerebral, or cardiopulmonary dysfunction before hepatectomy.

Diagnosis and definitions

Diagnose of HCC was based on the results of postoperative pathology. Hepatectomy with curative intent is defined as complete resection of all visible tumor and no residual tumor cells at surgical margins17. Preoperative NLR, LMR and PLR were determined within 7 d before surgery.

Follow-up visits

All patients were followed up one month after liver resection and followed by a three-month interval in the first year, and then every 3–6 months in subsequent years, as appropriate. At each follow-up, serum AFP assay, ultrasound, serum biochemistry, chest X-ray and abdominal CT or MRI were performed.

Statistical analysis

Statistical analysis was carried out with SPSS 19.0 (IBM, USA). Intergroup differences in continuous variables were compared by t-test or Mann–Whitney U test, as appropriate. While intergroup differences in categorical data were analyzed by two-sided χ2 test, Mann–Whitney U test or Fisher’s exact test, as appropriate. The area under the receiver operating characteristic (ROC) curve was calculated to measure the discriminatory power of the LMR, NLR and PLR as predictors of overall survival(OS). And an optimal cutoff value of 4.01, 2.78 and 99.5 corresponded to the maximum joint sensitivity and specificity on the ROC plot for LMR, NLR and PLR. Survival analysis was conducted to compare the over survival(OS) and disease free survival(DFS) rates using Kaplan-Meier survival curves with log-rank tests and Cox proportional hazard regression analyses, and p < 0.05 was considered statistically significant.

Data sharing statement

Technical appendix, statistical code, and dataset are available from the corresponding author at Participants gave informed consent for data sharing.


Baseline characteristics of all patients

Of 854 consecutive patients who underwent hepatic resection with curative intent from April 2004 to April 2012 inclusive, 145 patients (169%) experienced other treatments before hepatectomy; 20 patients (23%) had other malignant tumors simultaneously; 25 patients (29%) had preoperative fever and coexistent hematologic disorders, and 12 patients (14%) had cardiopulmonary, cerebral or renal dysfunction before hepatectomy. After exclusion, 652 patients (763%) were enrolled in our study.

Our study were consisted of 566 males and 86 females. Details of baseline characteristics of all HCC patients and separately in the low and high LMR groups can be seen in Table 1. ROC analysis found the optimal cutpoint for the LMR, NLR, and PLR were 4.01, 2.78 and 99.5, respectively (Fig. 1). The area under the ROC curve (AUC) of the LMR (AUC:0.627) for predicting long-term survival was greater than that of the NLR (AUC:0.600) and the PLR(AUC:0.520).

Table 1 Characteristics of HCC patients treated by resection.
Figure 1
figure 1

Receiver operating curve (ROC) for the preoperative LMR, NLR and PLR in predicting long-term survival with an optimal cutoff value of 4.01 (sensitivity:62.1 percent,specifiity:59.2 percent, AUC:0.627) for LMR, 2.78 (sensitivity:30.9 percent,specifiity:81.6 percent, AUC:0.600) for NLR, and 99.5 (sensitivity:47.9 percent, specifiity:57.9 percent, AUC:0.520) for PLR. AUC: (LMR, 0.656 vs. PLR, 0.600 P < 0.001).

A lower LMR was significantly associated with higher values of AFP, tumor number, tumor size, PLT and AST (all P < 0.05). On the other hand, A lower LMR was significantly associated with lower values of albumin (P < 0.05). The lower LMR was also strongly associated with both higher NLR (P < 0.001) and higher PLR (P < 0.001). Especially, high NLR was more likely in the low LMR group (40.5%) compared to both the overall cohort (25.6%) and those with high LMR (9.3%).

Cox analyses of survival associated with LMR

Univariate analysis and cox proportional hazards regression were used to identify factors associated with OS and DFS and the detailed results were shown in Table 2.

Table 2 Cox proportional hazards regression to identify predictors of overall and disease free survival in HCC patients treated by resection.

Univariate analysis identified the following factors significantly associated with poor OS: LMR < 4.01, AFP ≥ 400 ng/mL, BCLC stage B or C, multiple tumors, tumor size ≥ 5 cm, incomplete tumor capsule, AST ≥ 80 U/L, NLR ≥ 2.78 and PLR ≥ 99.5. LMR < 4.01, AFP ≥ 400 ng/mL, BCLC stage B or C, multiple tumors, tumor size ≥ 5 cm, incomplete tumor capsule, NLR ≥ 2.78 and PLR ≥ 99.5 were also found to be significantly associated with poor DFS (Table 2, Fig. 2).

Figure 2
figure 2

Kaplan-Meier survival curves showing overall survival (A) and disease-free survival (B) of hepatocellular carcinoma patients with high or low LMR.

Multivariate analysis (Table 2) identified the following prognostic independent predictors of poor OS: LMR < 4.01, BCLC stage B or C, multiple tumors, tumor size ≥ 5 cm and incomplete tumor capsule. LMR < 4.01, AFP ≥ 400 ng/mL, B multiple tumors, tumor size ≥ 5 cm, incomplete tumor capsule and NLR ≥ 2.78 were also found to be significantly associated with poor DFS.

Subgroup analyses associated with LMR

To clarify the subgroups of patients influenced by preoperative LMR, we grouped the patients according to HBV infection and liver cirrhosis, we discovered LMR was significantly associated with poor OS and DFS for patients with or without HBV infection (all P < 0.01; Fig. 3). LMR was also significantly associated with poor OS and DFS for patients with liver cirrhosis (both P < 0.001; Fig. 4B,D) but not for patients without cirrhosis (P = 0.466; Fig. 4C).

Figure 3
figure 3

Overall survival (A,B) and disease-free survival (C,D) of HCC patients stratified by LMR according to (A,C) negative HBsAg and (B,D) positive HBsAg.

Figure 4
figure 4

Overall survival (A,B) and disease-free survival (C,D) of HCC patients stratified by LMR according to (A,C) non-cirrhosis and (B,D) cirrhosis.


Our research implies that the preoperative LMR is an independent indicator of poor OS and DFS in HCC patients suffered curative resection. Furthermore, in the subgroup of patients positive for HBsAg or with cirrhosis, we found that the LMR was also significantly associated with OS and DFS.

We have also defined optimal cutpoints for LMR, NLR and PLR in the ROC analysis, with values of 4.01, 2.78 and 99.5, respectively. And we have found the area under the ROC curve (AUC) of the LMR (AUC:0.627) for predicting long-term survival was greater than that of the NLR (AUC:0.600) and the PLR(AUC:0.520). Furthermore, we have also found the LMR is a better indicator for long-term survival compared to NLR and PLR, both publicly known independent indicators of OS and DFS7,14,18,19. Before this research, studies had usually reported the NLR is superior to the PLR16,20,21. However, when studied with LMR, we found both NLR and PLR were not independent indicator for OS. What’s more, the cutpoints applied in my study for NLR and PLR of 2.78 and 99.5, respectively were consistent with previous researches19,22.

The mechanism of why decreased LMR should indicate poor survival remains unclear, but some researches have given some explanations. As far as we know, as early as the 1970s, these markers of lymphocyte components have been fully confirmed in the study in which lymphocytopenia were observed in patients with advanced tumor23. This has been one of the foundations of other established inflammatory markers, such as PLR and NLR. Although the correlation between monocytes and prognosis have been evaluated recently, only preliminary hypotheses have been made to explain why monocytes may indicate prognostic information. Here we put forward several possible mechanisms. First, previous study have found circulating monocytes may promote tumor growth and help tumor cells escape immune surveillance24,25. Second, derived from circulating monocytes, tumor-associated macrophages (TAMs) have been reported to be able to infiltrate into the HCC matrix, exerting activity including promotion of proliferation, metastasis, angiogenesis and immunosuppression26,27,28,29.

Although offering some new findings, this study has several limitations that need to be addressed. On one hand, since the value of LMR is dynamic during follow-up, we use only baseline values to predict future results, so we may miss a lot of information. On the other hand, due to the nature of the retrospective study, the potential choice bias exists.

In summary, our study suggests that the LMR is superior to the NLR and PLR, both established independent indicators of OS. However, a larger prospective study are needed to prove our researches.