Pretreatment Liver Injury Predicts Poor Prognosis of DLBCL Patients

Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous group of lymphoma, with different clinical manifestation and prognosis. The International Prognostic Index (IPI), an index designed during the prerituximab era for aggressive lymphoma, showed variable values in the prediction of patient clinical outcomes. The aim of this study was to analyze the prognostic value and causes of pretreatment liver injury in 363 de novo DLBCL patients in our institution. Pretreatment liver impairment, commonly detected in lymphoma patients, showed significant association with poor outcomes and increased serum inflammatory cytokines in DLBCL patients but had no relation to hepatitis B virus replication nor lymphomatous hepatic infiltration. Multivariate analysis revealed that liver dysfunction, advanced Ann Arbor stage, and elevated lactate dehydrogenase (LDH) were independent adverse prognostic factors of both PFS and OS. Accordingly, a new liver-IPI prognostic model was designed by adding liver injury as an important factor in determining IPI score. Based on Kaplan-Meier curves for PFS and OS, the liver-IPI showed better stratification in DLBCL patients than either the IPI or the revised IPI in survival prediction.

Cytokines are documented to be closely associated with both inflammation and immune modulation while playing a key role in the development of liver damage in a variety of liver disease such as chronic hepatitis B virus (HBV) infection, alcoholic liver injury, nonalcoholic fatty liver disease, and drug-induced liver injury [5][6][7][8]. It is generally believed that cytokines are deregulated in many kinds of haematological disorders [9,10], while elevation of interleukin-(IL-) 6, IL-10, tumor necrosis factor-(TNF-) α, IL-8, and IL-2 receptor (IL-2R) was demonstrated valuable in the prediction of unfavorable prognosis in lymphoma [11][12][13][14].
The aim of the present study was to determine the role of liver inflammation, reflected by the cytokines and serum transaminase activities, gamma-glutamyltranspeptidase (γ-GT), and alkaline phosphatase (ALP) in the prediction of outcome in DLBCL patients.

Patients.
We conducted a single-center retrospective case-control study on de novo DLBCL patients. A total of 363 patients were included, with the histological classification confirmed according to the World Health Organization (WHO) 2008 criteria [15]. Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-GT, and ALP were used as markers of liver injury as recommended by the regulatory authorities [16]. Serum cytokine tests (including IL-1β, IL-2R, IL-6, IL-8, IL-10, and TNF-α) were systematically assessed before chemotherapy. Cytokines were detected in the serum of 15 healthy volunteers as controls. Clinical characteristics of the 363 patients are shown in Table 1. Patients with abnormal liver function, defined as elevation in any of the following four indexes: ALT, AST, γ-GT, or ALP, were recruited into the liver dysfunction group; then, a propensity score matching method was used to create the matched control group [17]. Patients were matched at a ratio of 1 : 1 using the nearest neighbor method with a caliber of 0.10. All the patients and volunteers gave their informed consent, following the regulations of the Shanghai Jiao Tong University School of Medicine Institutional Review Boards, in accordance with the Declaration of Helsinki.

Response Criteria.
The treatment response was evaluated according to the WHO response criteria [18]. Complete response (CR) was defined as no evidence of residual disease, partial response (PR) as having at least a 50% reduction in tumor burden from the onset of treatment, and no response as having less than a 50% reduction in tumor burden or disease progression. Assessment of the treatment response was evaluated by a follow-up clinical, radiological, or laboratory study, as determined by the clinician, as described previously [19,20].

Statistical Analysis.
Baseline characteristics of patients were analyzed using Student's t-tests for continuous variables, χ 2 tests for categorical data, and Mann-Whitney U test for the serum level of cytokines. Overall survival (OS) time was measured from the date of diagnosis to the date of death or to the last follow-up. Progression-free survival (PFS) was calculated from the date when the treatment began to the date when the disease progression was recognized or the date of the last follow-up as described previously [19,20]. Survival functions were estimated using the Kaplan-Meier method and compared by the log-rank test. Univariate hazard estimates were generated with unadjusted Cox proportional hazards. Multivariate survival analysis was performed using a Cox regression model in which significant variables in the univariate analysis were included. p < 0 05 was considered statistically significant. All statistical analyses were carried out using Statistical Package for the Social Sciences (SPSS) 22.0 software (SPSS Inc., Chicago, IL, USA).  Table 1). Patients with liver dysfunction had no relationship with HBV replication or lymphomatous hepatic infiltration but were significantly associated with advanced Ann Arbor stage (p < 0 001), poor performance status (p < 0 001), increased LDH level (p < 0 001), high IPI score (p < 0 001), presence of B symptoms (p = 0 002), and low CR rate (p = 0 004, Table 1). Since cytokines in the serum had been reported to be associated with liver inflammation and dysfunction [5][6][7][8], patients with liver dysfunction had significantly higher level of IL-2R, IL-6, IL-10, and TNF-α in the serum, when compared with those without liver dysfunction (p < 0 001, Table 1).

Liver Dysfunction Was Associated with Poor Outcome and High Serum Cytokine Levels in Matched Case-Control
Analysis in DLBCL. To avoid the unfavorable impact of advanced disease stage and high IPI score on the outcome of the patients, 87 of 276 patients without liver dysfunction were selected as case controls using 1 : 1 matching on propensity scores for sex, age, IPI score, and lymphomatous hepatic infiltration, with a caliber of 0.10. Clinical characteristics of the 174 patients selected by propensity score matching are shown in Supplement Table 1S available online at https://doi.org/10.1155/2017/7960907. After matching, elevated LDH level was still observed in patients with liver dysfunction (Table 1S).
With a median follow-up of 11.7 months in both groups, patients in the liver dysfunction group showed significantly poorer outcomes than those in the matched control group (liver dysfunction group: 2-year PFS 58.5% and 2-year OS 65.2%; matched control group: 2-year PFS 74.0% and 2year OS 84.9%, p = 0 019 and 0.001, resp.; Figure 1(a)).

The New Prognostic Model Liver-IPI Was Developed in
Our DLBCL Cohort. Since liver dysfunction is an independent prognostic factor for both PFS and OS, it was combined with the IPI to design a new prognostic model, named as the liver-IPI. In the liver-IPI model, elevation of ALT, AST, γ-GT, or ALP was scored as 1 point, combined with IPI 5 scores to reach a total score of 6. Three risk groups were formed: low-risk (0-1 scores), intermediate-risk (2-3 scores), and high-risk (4-6 scores). The liver-IPI showed better stratification of patients than either the IPI or the R-IPI in OS and PFS, since significant differences were found between low-and intermediate-risk groups (PFS (p < 0 001) and OS (p = 0 016); Figure 3(c)), as well as in intermediateversus high-risk groups (p < 0 001 for both PFS and OS; Figure 3(c)). However, according to the IPI, no significant difference of OS and PFS was found between the lowintermediate-risk group and high-intermediate-risk group (p = 0 251 and p = 0 443, resp.; Figure 3(a)). Similarly, no difference of PFS was found between high-intermediateand high-risk groups (p = 0 058; Figure 3(a)). For the R-IPI, there was no statistic difference of OS between the very good and good groups (p = 0 114; Figure 3(b)).

Discussion
To our knowledge, this is the first report showing that pretreatment liver dysfunction was associated with poor prognosis in patients with DLBCL. Elevated serum transaminase activities, γ-GT, and ALP were significantly associated with extended lymphoma disease (advanced Ann Arbor stage, elevated LDH level) and alteration of the host status (poor performance status and presence of B symptoms). Meanwhile, it is also revealed that impaired liver function is not directly caused by HBV replication or lymphomatous hepatic infiltration. Of note, in the liver dysfunction group, significant poor treatment outcome with shorter PFS and OS was observed, particularly in those patients of high-intermediate and high risk. Furthermore, in multivariate Cox regression analysis, pretreatment liver function impairment was an independent unfavorable prognostic factor, which fully demonstrated the prognostic value of liver injury on DLBCL. Therefore, a new prognostic model based on liver function and IPI score, liver-IPI, was designed. The liver-IPI showed a better stratification of different outcomes in patients than the IPI and R-IPI.
In the liver dysfunction group, patients had significantly higher level of IL-2R, IL-6, IL-10, and TNF-α, when compared with those in the normal liver function group. Accumulating data has shown that an imbalance in cytokine production is critically involved in the development of liver damage in a variety of liver diseases. TNF-α, a central regulator of inflammatory and immune responses, is secreted by activated monocytes, macrophages, and T lymphocytes [21,22]. Increased TNF-α production not only contributes to chronic alcoholic liver injury [23] but also influences the nonalcoholic fatty liver disease process [7]. Soluble IL-2R (sIL-2R) is the soluble form of IL-2R, which is expressed on the cell membrane of lymphocytes and plays an important role in their activation and proliferation [24]. It is released from activated T-cells mainly due to the cleavage by proteinase matrix metalloproteinase-9 produced by inflammation-related cells [25]. The level of sIL-2R reflects the extent of inflammation [26] and correlate with fibrosis stages in patients with chronic HBV infection [5]. Increased IL-6 and IL-10, two major inflammatory cytokines, are reported in ethanol-induced hepatocellular damage and concanavalin A-induced liver injury [27]. In vivo, cytokines usually form a network to augment the inflammation and liver impairment. As a mechanism of action, following the induction of IL-6, IL-8, and IL-10 secretion, TNF-α could activate the nuclear factor-kappa B pathway and enhance the adhesion molecule expression, which in turn results in adherence of neutrophils and monocytes to the endothelium. Accumulation and activation of inflammatory cells further generate ROS and NO and induce liver damage [8,[28][29][30][31]. These mechanisms partially explained the phenomena that pretreatment liver injury was associated with high level of cytokines and poor outcome of patients, without being related to the HBV replication and lymphomatous hepatic infiltration.
Univariate analysis revealed that elevated serum cytokines IL-2R, IL-6, IL-10, and TNF-α correlated with the decreased OS and PFS rate. Accumulating researches have pointed out that in lymphoma patients, TNF-α accumulation is associated with lymphoma progression [32] and serum sIL-2R is a predictor of poor outcome in DLBCL patients [13,33]. IL-6 and IL-10 belong to T-helper type 2 cell cytokines, contributing to inhibition of host's immune system and induction of tumor progression [34,35]. Several studies showed that increased levels of serum IL-6 and IL-10 indicated a poor therapeutic response rate and short survival time in DLBCL [11,12,[36][37][38].

Conclusion
Pretreatment liver injury was an independent poor prognostic factor in newly diagnosed DLBCL patients, correlating with increased serum levels of liver dysfunction-associated   cytokines IL-2R, IL-6, IL-10, and TNF-α. In addition, liver-IPI, based on liver function and IPI score, had a satisfactory prognostic value in the risk stratification of DLBCL.

Conflicts of Interest
The authors declare no conflict of interest.