Histologic Assessment of Intratumoral Lymphoplasmacytic Infiltration Is Useful in Predicting Prognosis of Patients with Hepatocellular Carcinoma

In the present study, we investigated the clinicopathologic significance of intratumoral lymphoplasmacytic infiltration in a large cohort of patients with solitary hepatocellular carcinoma (HCC). Based on examination of hematoxylin and eosin-stained sections, significant infiltration was defined as dense lymphoplasmacytic infiltration, either multifocal or diffuse, in 2 or more fields under low-power magnification. Of 544 cases, 216 (39.7%) were positive for significant infiltration (HCC-LI group), while 328 (60.3%) were negative (HCC-NLI group). There were no significant between-group differences in patient age, sex, or background etiology. The lower incidence of Child-Pugh stage B (P = 0.001) and lower level of indocyanine green retention rate at 15 minutes (P < 0.001) in the HCC-LI group indicated better liver function in this group. Histologically, tumors were significantly smaller in size in the HCC-LI group than in the HCC-NLI group (P < 0.001). In addition, prominent neutrophilic infiltration, interstitial fibrosis and tumor steatosis were significantly more frequent (P < 0.001) in the HCC-LI group, while tumor necrosis was significantly less frequent (P = 0.008). Kaplan-Meier analyses revealed that overall and recurrence-free survival were significantly better in the HCC-LI group (P < 0.001). Multivariate Cox regression analysis showed that intratumoral lymphoplasmacytic infiltration was independently prognostic of both overall and recurrence-free survival (P < 0.001), with absence of infiltration showing high Cox-hazard ratios for poor prognosis. In conclusion, intratumoral lymphoplasmacytic infiltration, as determined by assessment of hematoxylin and eosin-stained slides, was significantly associated with the clinical and pathologic features of HCC and has profound prognostic importance.


Introduction
Hepatocellular carcinoma (HCC) is a major cancer worldwide, [1] and its incidence is expected to increase. [2] Despite recent advances in resection and ablation techniques, the recurrence rate after initial treatment remains high and the prognosis of patients with HCC is generally unfavorable. [3,4] In addition, chemotherapy for recurrent HCC after surgical resection is relatively ineffective. [5] To manage treatment-refractory cancers, including HCC, therapies are being designed to specifically target the immune system rather than cancer cells. [6] Immune cells have both antitumorigenic and protumorigenic activities, with the latter predominating in patients with noneliminated, clinically detectable tumors. [7] The roles of inflammatory cell types in tumor immunity have been assessed. For example, cytotoxic T cells exert antitumor activity, whereas tumor-infiltrating macrophages and regulatory T cells repress such activity. [8] Immune-targeted therapies reinforce antitumor response or prevent inhibitory activity. The degree of immune cell infiltration has prognostic value in many types of cancer. [6] Tumor infiltrative lymphocytes (TILs) have generally been associated with favorable prognosis. [9] More specifically, different types of infiltrative lymphocytes have different prognostic values. [8] Several studies have analyzed the prognostic significance of immune cell types in patients with HCC. For example, infiltration of CD8-positive T cells has been associated with good prognosis, [10] whereas infiltration of Foxp3-positive T cells has been associated with poor prognosis. [11,12] However, the immunohistochemical procedures required for these analyses may be difficult to implement in routine diagnostic practice. In addition, considering that TILs have cell-type-dependent roles, it is important to determine the general roles and significance of TILs. To our knowledge, however, few studies of HCC have focused on the prognostic significance of TILs from a broad point of view. Wada et al. [13] revealed that massive lymphocyte infiltration was associated with better prognosis in patients with small HCC after surgical resection. Unitt et al. [14] showed that mild to moderate lymphocytic infiltration correlated with better prognosis after liver transplantation for HCC. These studies, however, did not provide sufficient evidence on the clinicopathologic significance of TILs in patients with HCCs treated by surgical resection, the primary curative treatment for solitary HCC of any size. [15] In addition to lymphocyte infiltration, accompanied plasma cell infiltration has been reported in several types of cancer, including HCC [13,16]. Though the role of infiltrative plasma cells remains controversial, a recent article described their characteristics, in particular their role in producing IgG4 [17].
In this study, we assessed intratumoral lymphoplasmacytic infiltration in more than 500 solitary HCCs by examining hematoxylin and eosin-stained (H&E) slides to precisely determine the clinicopathologic significance of intratumoral lymphoplasmacytic infiltration in HCC.

Patients
This retrospective analysis included consecutive patients with HCC who underwent initial surgical resection at the University of Tokyo Hospital from January 1, 1995 to December 31, 2011. Patients who underwent preoperative transarterial therapy or portal embolization were included only if a sufficient portion of the tumor remained viable. To precisely evaluate tumor prognosis, only patients with solitary HCCs, with or without intrahepatic metastasis, were included, whereas patients with multicentric HCCs [18] and those undergoing resection for liver transplantation were excluded.

Clinical data
Evaluation of patient records was used to determine baseline demographic and clinical data, including patient age, sex, and serum data immediately before surgery; status of hepatitis virus infection; presence or absence of diabetes mellitus; and history of heavy alcohol consumption ( 80 g/day). All patients were regularly screened for HCC recurrence by monitoring plasma tumor markers, ultrasonography, and computed tomography. Recurrence was defined as the appearance of a new lesion with radiological features compatible with HCC. Overall survival was defined as the interval between the date of surgery and the date of death, and recurrencefree survival was defined as the interval between the date of surgery and the date of recurrence.

Histologic review
Pathology reports and all tissue slides from all patients were reviewed. Tumor size, histologic grade, presence or absence of microvascular invasion, bile duct involvement, and intrahepatic or lymph node metastasis were re-evaluated. Other factors recorded included the presence or absence of steatosis in 5% of tumor cells, prominent interstitial fibrosis at least in one lowpower field, prominent neutrophilic infiltration, and tumor necrosis. In most of these patients, these histologic factors had been evaluated in our previous study. [19] Significant lymphoplasmacytic infiltration was defined in this study as dense multifocal or diffuse lymphoplasmacytic infiltration in two or more fields under low power magnification (×4 objective). Perinecrotic areas were excluded from the assessment of inflammatory cells. HCCs with and without significant lymphoplasmacytic infiltration were assigned, respectively, to the HCC with lymphoplasmacytic infiltration (HCC-LI) group and the HCC with no lymphoplasmacytic infiltration (HCC-NLI) group.
The fibrosis stages of background liver in patients with and without viral hepatitis were evaluated according to the METAVIR system [20] and the Nonalcoholic Steatohepatitis Clinical Research Network (NASH-CRN) scoring system, [21] respectively. Advanced fibrosis was defined as stages 3 and 4 in both the METAVIR system and NASH-CRN scoring systems. The presence or absence of steatosis (> 5%) was also recorded.

Statistical Analysis
All statistical analyses were performed using JMP Pro 11 (SAS Institute Inc., Cary, NC, USA). Differences were considered significant at P < 0.05. Categorical data were analyzed using the two-sided Fisher's exact test. Quantitative variables were compared using Student's t test or the Wilcoxon rank sum test, as appropriate. The Kaplan-Meier method and log-rank test were used to analyze survival. Multivariate Cox proportional hazards regression models were used to control for confounding variables. Multivariate regression analysis was performed using all clinicopathologic variables. For subanalysis (treatment-naïve vs. pretreatment patients), only clinicopathologic variables identified as significant on univariate analyses are included due to limitations regarding data size. The "liver cirrhosis" variable was excluded from multivariate analysis because all patients with cirrhosis were included in the group with advanced fibrosis.

Ethics Statement
The study protocol was approved by the Medical Research Center Ethics Committee of the University of Tokyo. Clinical samples were collected under the University of Tokyo institutional guidelines for the study of human tissues, with all patients providing written informed consent. and C infections were noted in 116 (21.3%) and 289 (53.1%) patients, respectively, and 143 (26.3%) were negative for hepatitis virus infection. Diabetes mellitus was documented in 139 of 542 patients (25.6%), and 97 of 537 patients (18.1%) had a history of heavy alcohol consumption.

Clinical characteristics of HCC with lymphoplasmacytic infiltration
Based on our criteria, 216 (39.7%) of the 544 patients were assigned to the HCC-LI group and 328 (60.3%) to the HCC-NLI group. Lymphoplasmacytic infiltration was noted at the tumor periphery, in the interstitium, or within the deep parenchyma of the tumor (Fig 1A, S1 Digital Data. Case 1). Of 216 HCC-LIs, multifocal and diffuse patterns were observed in 162 cases (75.0%) and 54 cases (25.0%), respectively. In particular, three cases showed diffuse lymphoplasmacytic infiltration throughout the tumor, corresponding to lymphoepithelioma-like carcinoma ( Fig 1B, S1 Digital Data. Case 2).
The clinical features of the HCC-LI and HCC-NLI groups are summarized in Table 1. Patient age, sex, status of hepatitis virus infection, incidence of diabetes mellitus, and history of heavy alcohol consumption did not differ significantly between the two groups. Of four serum tumor markers, the mean level of protein induced by vitamin K absence or antagonist 2 (PIVKA-2) was significantly lower in the HCC-LI group than in the HCC-NLI group (P = 0.022). The percent of patients with Child-Pugh stage B liver function (6.5% vs. 15.5%, P = 0.001) and the mean indocyanine green retention rate at 15 minutes (ICGR15) (13.3% vs.15.7%, P < 0.001) were significantly lower in the HCC-LI group than in the HCC-NLI group. Logistic regression analyses indicated that the absence of intratumoral lymphoplasmacytic infiltration was independently predictive of Child-Pugh Stage B and a higher ICGR15 (Table 2). Additionally, the mean serum aspartate aminotransferase (AST) concentration was significantly lower in the HCC-LI group than in the HCC-NLI group (P < 0.001). Significantly fewer patients in the HCC-LI group received preoperative treatment (24.5% vs. 39.3%, P < 0.001).

Pathologic Features of HCC with lymphoplasmacytic infiltration
The pathologic features of HCC-LIs and HCC-NLIs are summarized in Table 3. Tumors in the HCC-LI group were significantly smaller in size than those in the HCC-NLI group (43 ± 29  mm vs. 58 ± 42 mm, P < 0.001), but there were no between-group differences in histologic grade, rates of microvascular invasion, bile duct invasion, or intrahepatic metastasis. Prominent neutrophilic infiltration (12.5% vs. 3.7%, P < 0.001), interstitial fibrosis (76.4% vs. 33.5%, P < 0.001) and tumor steatosis (30.6% vs. 17.7%, P < 0.001) were significantly more frequent in the HCC-LI group than in the HCC-NLI group, whereas tumor necrosis was significantly less frequent (49.1% vs. 60.7%, P = 0.008). The rates of steatosis, advanced fibrosis, and liver cirrhosis in background liver did not differ significantly between the two groups.

Prognosis
The median follow-up period for the 544 patients was 52.7 months (range 0.8-168.5 months). Their cumulative overall and recurrence-free survival rates at five years were 42.7% and 19.0%, respectively. Kaplan-Meier analyses showed significantly longer overall and recurrence-free survival of patients in the HCC-LI group than in the HCC-NLI group (P < 0.001 each by logrank tests) (Fig 2). Multivariate Cox regression analyses showed that intratumoral lymphoplasmacytic infiltration, microvascular invasion, intrahepatic metastasis, tumor steatosis, and advanced fibrosis of background liver were independently prognostic of overall survival, whereas intratumoral lymphoplasmacytic infiltration, tumor size, microvascular invasion, intrahepatic metastasis, and advanced fibrosis of background liver were independently prognostic of recurrence-free survival (Table 4).

Discussion
This study showed that lymphoplasmacytic infiltration in HCC, defined by our simple criteria, was significantly associated with several clinical and pathologic features. Most importantly, this infiltration was independently prognostic for overall and recurrence-free survival, with absence of infiltration showing high hazard ratios for poor prognosis. The mean serum level of PIVKA2 was lower in the HCC-LI group than in the HCC-NIL group. The previously demonstrated correlation of this tumor marker with tumor size, [22] confirmed in the present study (data not shown), indicated that the lower level of PIVKA2 in   the HCC-LI group was largely attributable to the significantly smaller sizes of these tumors. We speculated that the notably better liver function in patients in the HCC-LI group was also related to the smaller sizes of these tumors (larger volume of residual liver). Logistic multivariate analysis, however, showed that intratumoral lymphoplasmacytic infiltration was an independent negative predictor of Child-Pugh stage B and high ICGR15, indicating that liver function was directly associated with the extent of inflammatory infiltration. Presumably, patients with good liver function tend to retain anti-tumor immunity, whereas those with impaired liver function are prone to lose that reactivity. This hypothesis is in agreement with the finding that patients with advanced liver disease are frequently in an immunocompromised state. [23] The lower incidence of preoperative treatment in the HCC-LI group was likely due to the smaller tumor sizes in these patients, because tumor size and estimated residual liver volume have been identified as major determinants of preoperative treatment in our clinical practice. The higher incidence of tumor necrosis in the HCC-NLI group probably resulted from frequent preoperative treatment.
Pathologically, significantly smaller tumor sizes in the HCC-LI group may reflect inflammatory responses that suppressed tumor growth. Patients in both the HCC-LI and HCC-NLI groups, at least those infected with hepatitis virus, had been regularly and similarly screened for tumor development, suggesting that tumors with inflammatory infiltration grew more slowly. Alternatively, tumors may acquire anti-inflammatory (anti-immune) properties when they grow larger. Further basic studies are needed to determine the mechanism by which inflammatory infiltration influences tumor growth.
The cause of the higher frequency of tumor steatosis in the HCC-LI group may be multifactorial. It may be related to the smaller-size tumors in this group, because small, early-stage HCCs frequently manifest fatty changes. [18] The causal relationship between intracellular lipid accumulation and inflammatory reaction, as observed in patients with non-alcoholic fatty liver disease, [24] suggests that tumor steatosis may have induced inflammatory reaction. A recent study demonstrating that inflammatory cells induced hepatocellular steatosis in a mouse model of non-alcoholic steatohepatitis [25] indicates another possibility.
Frequent interstitial fibrosis of HCC-LIs is quite plausible, as hepatic fibrosis generally results from the immune reactive process. [26] Hepatic stellate cells are thought to play crucial roles in fibrogenesis. [27] These cells also play an important role in fibrosis in HCC, interacting with tumor cells and inflammatory cells. [28] The present study, analyzing a large cohort of patients with HCC, showed that intratumoral lymphoplasmacytic infiltration was an independent prognostic factor for survival in patients with HCC. This study confirmed the prognostic significance of histologic features previously shown to be prognostic in patients with HCC, including tumor size, histologic grade, microvascular invasion of tumors, and advanced fibrosis of background liver. Of particular note, absence of infiltration showed the second highest hazard ratio for overall survival, after advanced fibrosis of background liver, and the second highest hazard ratio for recurrence-free survival, after intrahepatic metastasis. Additional subgroup analyses of treatment-naïve vs. pretreatment patients (S1-S5 Tables and S1 Fig) showed that HCC-LI cases had better preserved liver function than treatment-naïve cases, while of the pretreatment patients, HCC-LI showed a stronger association with neutrophil infiltration or steatosis. Remarkably, in both the treatment-naïve and pretreatment groups, HCC-LI was associated with better prognosis in terms of both recurrence-free and overall survival as determined using multivariate Cox proportional hazards regression models. These findings indicate that intratumoral lymphoplasmacytic infiltration should be assessed in routine diagnostic practice.
The favorable prognosis of patients with HCC-LI presumably had at least two causes. First, robust antitumor immunity should have inhibited the growth of residual tumor that was not apparent at the time of surgical resection. This mechanism seems responsible for the favorable prognosis of tumors with inflammatory infiltration, which have been shown in cancers of various organs. [9,29] Additionally, specific conditions associated with hepatocarcinogenesis should be entertained; most patients in this study had advanced chronic liver diseases, and thus were cancer prone. Considering that the rates of liver cirrhosis and advanced fibrosis were similar in the HCC-LI and HCC-NLI groups, patients with HCC-LI, with supposedly healthy immune function, probably exhibited effective antitumorigenic immune reactions to newly developing lesions.
Much remains to be elucidated about the basic mechanisms of tumor immunity. Even the roles of inflammatory cell types have not been sufficiently clarified. In this study, we found that plasma cell infiltration was not infrequently prominent in HCC-LIs, and we therefore described the infiltration as "lymphoplasmacytic infiltration." Compared with T cells, the roles of B cells in tumor immunity are largely unknown. Although numerous studies have shown that B cells promote rather than inhibit tumor development, recent studies have revealed these cells' antitumorigenic roles. [30,31] Molecular signatures differentiating HCCs with and without infiltration are also of interest. Genetic aspects of tumors may determine the extent of anti-tumor inflammatory reactions. For example, microsatellite instability is strongly associated with lymphocytic infiltration in colorectal cancers. [32] Interestingly, Tan et al. [33] demonstrated that HCCs with marked immune cell infiltrate were associated with the S1 molecular subclass, defined by a transcriptome metaanalysis of global HCC populations. Further studies are needed to determine the detailed relationships between the genetic characteristics of HCC and immune cell reactions.
Determining the optimal definition of lymphoplasmacytic infiltration was one of the most challenging issues in this study. We reviewed several definitions, and adopted a definition that was the most straightforward and reproducible. Two pathologists independently evaluated 82 recent three-year cases, and in 69 (84.1%) of them they were in agreement regarding the presence of lymphocytic infiltration. We believe that this classification scheme is currently reasonable and useful, but further research should evaluate more objective definitions that can be used in daily diagnosis.

Conclusions
Intratumoral lymphoplasmacytic infiltration was associated with favorable liver function, as shown by lower incidence of Child-Pugh stage B and lower ICGR15 levels, and with several pathologic features, including smaller tumor size, more frequent interstitial fibrosis, and steatosis of the tumor. Most importantly, infiltration was an independent prognostic factor for overall and recurrence-free survival, with absence of infiltration showing high Cox-hazard ratios for poor prognosis.