Tumor necrosis and infiltrating macrophages predict survival after curative resection for cholangiocarcinoma

ABSTRACT

Background. Tumor necrosis as well as tumor-associated macrophages (TAMs) in the tumor invasive front (TIF) have been suggested to have a prognostic value in selected solid tumors, inclusive hilar cholangiocarcinoma. However, little is known regarding their influence on tumor progression and prognosis in intrahepatic cholangiocarcinoma (iCC).

Methods. We analyzed surgically resected tumor specimens of human iCC (n = 88) for distribution and localization of TAMs, as defined by expression of CD68, formation of necrosis and extent of peritumoral fibrosis. Abundance of TAMs, tumor necrosis and grade of fibrosis were assessed immunohistochemically and histologically and correlated with clinicopathological characteristics, tumor recurrence and patients' survival. Statistical analysis was performed using SPSS software.

Results. Patients with tumors characterized by low levels of TAMs in TIF or necrosis showed a significantly decreased 1-, 3- and 5-y recurrence-free survival and a significantly decreased overall survival, when compared with patients with tumors showing high levels of TAMs in TIF or no necrosis. Patients with high density of TAMs in TIF showed significantly lower incidence of tumor recurrence, as well (p < 0.05). Absence of tumor necrosis and TAMs in TIF were confirmed as independent prognostic variables in the multivariate survival analysis (all p < 0.05).

Conclusions. High levels of TAMs in TIF or absence of histologic tumor necrosis are associated with a significantly improved recurrence-free and overall survival of patients with iCC. These results suggest TAMs and necrosis as valuable prognostic markers in routine histopathologic evaluation, and might indicate more individualized therapeutic strategies.

Abbreviations

HCC

hepatocellular carcinoma

HC

hilar cholangiocarcinoma

H&E

hematoxylin and eosin

iCC

intrahepatic cholangiocarcinoma

mAb

monoclonal antibody

PDGRF

platelet-derived growth factor receptor

TBS

tris-buffered saline

TAMs

tumor-associated macrophages

TCA

tumor central area

TIF

tumor-infiltrating front

TME

tumor microenvironment

TT

tumorous tissue

UICC

Union for International Cancer Control

VEGF

vascular endothelial growth factor

Introduction

All classes of leukocytes can be detected within malignant tumors. Tumor-associated macrophages (TAMs) constitute up to 50% of this leukocyte cell population. TAMs influence tumor progression in various ways that depend on differences in tumor microenvironment (TME), histology and TAMs' polarization state.¹ TAMs can be divided schematically into two phenotypes, M1 and M2, according to their effects on tumor. M1 polarization promotes inflammation and suppresses tumor progression by producing pro-inflammatory and antitumoral cytokines. In contrast, M2 state suppresses inflammation and antitumor immune response, resulting in the promotion of tumor progression.^2,3 TME-tailored immunologic phenotype of infiltrating immune cells mediates an impaired tumor surveillance that facilitates tumor progression. It is TME, tumor-derived factors and escape mechanisms that foster tumor growth, neovascularization and metastasis. However, a constant interplay between TME, infiltrating immune cells and their polarization state plays a key role in this process.

TME can shape and reeducate the infiltrating immunologic components into a pro-tumoral phenotype. However, novel experimental research highlights a differential influence and functionality of localized tumor sites, i.e., tumor central area (TCA) and tumor invasive front (TIF), on related immunologic responses.^4,5 TAMs localized in TCA or TIF can impact tumor progression in opposite manner.⁶ High abundance of TAMs is associated with an unfavorable prognosis in hepatocellular carcinoma (HCC), esophageal, ovarian and breast cancer.^7-12 In contrast, M1-polarized TAMs suppress tumor progression in non-small cell lung cancer.¹³

There is emerging data linking hepatic macrophage heterogeneity to formation of histologic tumor necrosis and peritumoral fibrosis.^14-18 Necrosis is associated with tumor recurrence in different tumor entities. Consequently, tumor necrosis has been included in pathological classifications and prognostic indices in primary solid organ malignancies.^19,20 The formation of peritumoral fibrosis has been shown to impact patient outcome and prognosis of several inflammatory and haematopoietic disorders, as well.^21,22

We have previously shown that formation of necrosis and TAMs in TIF impact patient survival and prognosis in hilar cholangiocarcinoma (HC).²³ Moreover, TAMs association to formation of tumor necrosis is considerable in HC.²⁴ However, the importance of TAMs and tumor necrosis remains uncertain in iCC. Scarce vascularization of this tumor entity may facilitate hypoxic conditions, and thus formation of necrosis might well characterize subsets of iCC, as well.^25,26 The aim of this study was therefore to evaluate the relationship and potential association of infiltrating TAMs, tumor necrosis and peritumoral fibrosis with tumor growth, metastasis, recurrence and clinical prognosis in iCC.

Results

One-, three- and five-year overall and recurrence-free survival of our cohort was 61.0%, 27.2% and 18.8%, and 46.8%, 22.7% and 18.6%, respectively (Figs. 1A and B). Sixty (67.4%) of 88 patients died within the follow-up interval. Sixty-four patients (71.9%) developed tumor recurrence and 55 (61.8%) patients died for recurrence of iCC. Local tumor recurrence was seen in 19 (21.3%) patients, while 2 (2.2%) patients developed distant metastases without manifestation of local recurrence (Table 1). None of the patients with local recurrence had a metastatic spread of the tumor.

Figure 1.

(Continued)

Figure 1.

(Continued)

Figure 1.

Overall survival and recurrence-free survival of the patients included in the study.

Figure 1.

(Continued)

Table 1.

Clinicopathological characteristics of the patients included in the study.

	Clinicopathological characteristics
Variable	Value (%)
No. of patients	88
Gender
Male	46 (52.3%)
Female	42 (47.7%)
Patient age
≤ 60	46 (52.3%)
>60	42 (47.7%)
Tumor nodules
Solitary	30 (35.8%)
Multiple	58 (64.2%)
Tumor diameter
Small	28 (31.8%)
Large	60 (68.2%)
Bilobular hepatic tumor involvement
Yes	52 (59.1%)
No	36 (40.9%)
Pathologic R-Category
R0	48 (54.5%)
R1/R2	40 (45.5%)
Distant metastases
With	2 (2.2%)
Without	86 (97.8%)
Angionvasion
Positive	5 (5.6%)
Negative	83 (94.4%)
Lymphangiosis carcinomatosa
Positive	38 (43.8%)
Negative	50 (56.2%)
Perineural sheath infiltration
Positive	24 (28.1%)
Negative	64 (71.9%)
Histologic differentiation
Well	62 (70.1%)
Moderate/Poor	26 (29.9%)
Pathologic T stage
T1/T2	35 (39.3%)
T3/T4	53 (60.7%)
Pathologic N stage
Positive	34 (38.6%)
Negative	54 (61.4%)
Tumor recurrence
With	62 (70.8%)
Without	26 (29.2%)
Local recurrence
With	18 (20.2%)
Without	70 (79.8%)

Macrophage distribution in intrahepatic cholangiocarcinoma

TAMs showed a homogenous infiltration pattern in TCA and TIF (Fig. 2A). The majority of TAMs were located in direct contact to tumor cells. Interestingly, a high density of TAMs was observed at TIF in 42 patients (TIF score 2/high density/positive), while rest of tumor specimens (n = 34) displayed only a scarce or complete absence of infiltrating TAMs (TIF scores 0 and 1/low density/negative) (Fig. 2B).

Figure 2.

(A) Intrahepatic cholangiocarcinoma stained with mAb CD68 PG-M1 with a high abundance of TAMs in tumor central area (TCA). Original magnification: x 200. (B) Intrahepatic cholangiocarcinoma stained with mAb CD68 PG-M1 with a high abundance of TAMs in tumor-infiltrating front (TIF). Original magnification: x 200. (C, D) Histological tumor specimen with (C) necrosis and (D) fibrosis. *Vital tumor cell formations, +Tumor necrosis, #Tumor fibrosis.

Tumor-infiltrating macrophages associate with pathological parameters and tumor recurrence

Patients with iCC were divided into several groups either by “low” or “high” density of macrophages in TCAs (“TumorCD68-low group”; n = 50) and “TumorCD68-high group”; n = 38, respectively), or “low” or “high” density of TAMs in TIF (“Invasive frontCD68-low group” n = 34; and “Invasive frontCD68-high group”, n = 42, respectively).

High frequency of TAMs in TIF correlated with reduced incidence of lymphangiosis carcinomatosa (Table 2). Thirty (71.4%) patients in the Invasive frontCD68-high group did not have lymphangiosis carcinomatosa, whereas in 23 (67.6%) patients in the Invasive frontCD68-low group presence of lymphangiosis carcinomatosa was detected (p = 0.001). Thirty-four (81.0%) patients in the Invasive frontCD68-high group had a well- differentiated tumor (G1), when compared with 20 (58.8%) patients in the Invasive frontCD68-low group (p = 0.021). Similarly, 23 (81.3%) patients in the Invasive frontCD68-low group displayed an advanced T stage, when compared with 18 (43.9%) cases in the Invasive frontCD68-high group (p = 0.001).

Table 2.

Correlation of CD68-positiveTAMs in the tumor invasive front with clinicopathological characteristics of intrahepatic cholangiocarcinoma.

Clinicopathological analysis
Variable	high CD68	low CD68	p
No. of patients	42	34
Patient age, years	0.356 ≤ 60	19 (45.2%)	19 (55.9%)
>60	23 (54.8%)	15 (44.1%)
Gender			0.951
Male	17 (40.5%)	14 (41.2%)
Female	25 (59.5%)	20 (58.8%)
Tumor nodules			0.062
Solitary	9 (21.4%)	14 (41.2%)
Multiple	33 (78.6%)	20 (58.8%)
Tumor diameter			0.064
Small	17 (40.5%)	7 (20.6%)
Large	25 (59.5%)	27 (79.4%)
Bilobular hepatic tumor involvement			0.101
Yes	11 (26.2%)	15 (44.1%)
No	31 (73.8%)	19 (55.9%)
Distant metastases			0.111
With	0 (0.00%)	2 (5.9%)
Without	42 (100%)	32 (94.1%)
Angioinvasion			0.197
With	2 (4.8%)	0 (0.00%)
Without	40 (95.2%)	34 (100%)
Lymphangiosis carcinomatosa			0.001
Positive	12 (28.6%)	23 (67.6%)
Negative	30 (71.4%)	11 (32.4%)
Perineural sheath infiltration			0.581
Positive	10 (23.8%)	10 (29.4%)
Negative	32 (76.2%)	24 (70.6%)
Histological differentiation			0.021
Well/moderate	34 (82.9%)	20 (58.8%)
Poor	7 (17.1%)	14 (41.2%)
Pathological T stage			0.001
T1/T2	23 (56.1%)	6 (18.8%)
T3/T4	18 (43.9%)	26 (81.3%)
Pathological N stage			0.744
Positive	15 (35.7%)	13 (39.4%)
Negative	27 (64.3%)	20 (60.6%)
Tumor recurrence			0.004
With	22 (52.4%)	30 (88.2%)
Without	20 (47.6%)	4 (11.8%)
Local recurrence			0.002
With	3 (7.1%)	12 (35.3%)
Without	39 (92.9%)	22 (64.7%)

Furthermore, low abundance of TAMs in TIF correlated with a significantly increased overall tumor recurrence (Table 2). Thirty (88.2%) patients in Invasive frontCD68-low group displayed tumor recurrence (p = 0.001). In comparison, in the Invasive frontCD68-high group in only 22 (52.4%) cases a recurrent disease was apparent. Moreover, patients with high levels of TAMs in TIF showed a significantly reduced incidence of local tumor recurrence, as well. Thirty-nine (92.9%) patients in the Invasive frontCD68-high group showed a complete absence of local tumor recurrence, whereas 22 (63.7%) in the Invasive frontCD68-low group did not show local tumor recurrence (p = 0.002).

Tumor necrosis associates with angioinvasion and formation of multiple tumor nodules

Tumor samples of 31 patients (34.8%) displayed no tumor necrosis (Necrosisnegative group), in 57 cases (64.7%), necrosis was evident (Necrosispositive group) (Fig. 2C). In the Necrosismild group, there were 34 (59.6%) patients. In comparison, the Necrosissevere group comprised 23 (40.4%) patients.

In the Necrosispositive group 4 (12.9%) patients displayed angioinvasion, whereas in the Necrosisnegative group in only one case angioinvasion was evident (p = 0.031) (Table 3). In the Necrosissevere group, 12 (52.2%) of 23 patients showed a formation of multiple tumor nodules, compared with only 9 (26.5%) cases in Necrosismild group (p = 0.048) (Table 3). The formation of histologic tumor necrosis showed a distinct trend to positively correlate with presence of TAMs in TCA or TIF, although without reaching statistical significance (p = 0.109 and p = 0.251, respectively) (Table 3). Patients with mild formation of histologic tumor necrosis showed a trend to reduced incidence of lymph node metastases, overall and local tumor recurrence (p = 0.078, p = 0.138 and p = 0.153, respectively; data not shown).

Table 3.

Correlation of histologic tumor necrosis with clinicopathological characteristics of intrahepatic cholangiocarcinoma.

Clinicopathological analysis
Variable	Necrosis positive	Necrosis negative	p
No. of patients	57	31
Patient age, years			0.325
≤ 60	23 (40.4%)	14 (45.2%)
>60	34 (59.6%)	17 (54.8%)
Gender			0.662
Male	17 (40.5%)	14 (41.2%)
Female	25 (59.5%)	20 (58.8%)
Tumor nodules			0.288
Solitary	25 (43.9%)	10 (32.3%)
Multiple	32 (56.1%)	21 (67.7%)
Tumor diameter			0.586
Small	17 (29.8%)	11 (35.5%)
Large	40 (70.2%)	20 (64.5%)
Bilobular hepatic tumor involvement			0.180
Yes	36 (63.2%)	15 (48.4%)
No	21 (36.8%)	16 (51.6%)
Distant metastases			0.658
With	1 (1.8%)	1 (3.2%)
Without	56 (98.2%)	30 (96.8%)
Angioinvasion			0.031
With	1 (1.8%)	4 (12.9%)
Without	56 (98.2%)	27 (87.1%)
Lymphangiosis carcinomatosa			0.093
Positive	29 (50.9%)	10 (32.3%)
Negative	28 (49.1%)	21 (67.7%)
Perineural sheath infiltration			0.371
Positive	18 (31.6%)	7 (22.6%)
Negative	39 (68.4%)	24 (77.4%)
Histological differentiation			0.308
Well/moderate	37 (67.0%)	23 (76.7%)
Poor	20 (34.0%)	8 (23.3%)
Pathological T stage			0.103
T1/T2	19 (33.3%)	16 (51.7%)
T3/T4	38 (66.7%)	15 (48.3%)
Pathological N stage			0.0.85
Positive	26 (45.6%)	9 (26.7%)
Negative	31 (54.4%)	22 (73.3%)
Tumor recurrence			0.368
With	42 (73.7%)	20 (64.5%)
Without	15 (26.3%)	11 (35.5%)
Local recurrence		0.195
With	14 (24.6%)	4 (12.9%)
Without	43 (75.4%)	27 (87.1%)
CD68-positive TAMs in TCA			0.109
High	29 (50.9%)	21 (69.0%)
Low	28 (49.1%)	10 (31.0%)
CD68-positive TAMs in TIF			0.251
High	25 (50.0%)	10 (36.0%)
Low	25 (50.0%)	16 (64.0%)
Variable	Mild necrosis	Severe necrosis	p
No. of patients	34	23
Multiple tumor nodules			0.048
With	9 (26.5%)	12 (52.2%)
Without	25 (73.5%)	11 (47.8%)

Peritumoral fibrosis associates with angioinvasion and lymph node metastasis

Thirty-two specimens (36.3%) showed mild fibrosis. Strong formation of peritumoral fibrosis was apparent in 56 (63.6%) patients (Fig. 2D).

In the Fibrosisstrong group, 55 (98.9%) patients displayed absence of angioinvasion, whereas in the Fibrosismild group, only in 4 (12.8%) patients no angioinvasion was detected (p = 0.033). Interestingly, patients with strong formation of peritumoral fibrosis showed a significantly reduced incidence of lymph node involvement, as well (p = 0.037). Thirty-seven (69.8%) cases in the Fibrosisstrong group showed a complete absence of lymph node metastases, whereas 14 (46.7%) patients in the Fibrosismild group did not have positive lymph nodes.

Prognostic significance of CD68-positive macrophages and tumor necrosis in intrahepatic cholangiocarcinoma

In the next step, we analyzed whether TAMs and tumor necrosis in addition to other clinicopathological parameters predict survival after resection for iCC. In the univariate analysis T-stage, multiple tumor nodules, histologic grading, nodal status, perineural sheet infiltration, R-category, overall tumor recurrence, TAMs in TIF and histologic tumor necrosis were associated with a statistically significant impact on patient survival after resection (Figs. 1C–K).

Figure 1.

(Continued)

As related to recurrence-free survival, in univariate analysis formation of multiple tumor nodules, bilobular hepatic tumor involvement, R category, histologic grading, T-stage, nodal status and TAMs in TIF were associated with a statistically significant impact (Figs. 1J–P).

Figure 1.

(Continued)

In multivariate analysis in regard of overall survival TAMs in TIF, histologic tumor necrosis, patient age, histologic grading and tumor recurrence were identified as independent prognostic factors for survival (all p < 0.05) (Table 4). When looking at recurrence-free survival, in multivariate analysis bilobular hepatic tumor involvement, perineural sheet infiltration and histologic grading exerted a significant impact on survival (all p < 0.05; Table 4). Other well-established common prognostic factors, e.g., lymph node involvement and local tumor recurrence did not exert prognostic significance in the current work.

Table 4.

Multivariate analysis of prognostic factors in patients with intrahepatic cholangiocarcinoma.

Multivariate analysis (Overall Survival)
Variable	Category	Odds ratio	p	Confidence interval
Histologic differentiation	well or moderate/ poor	2.266	0.040	1.038–4.946
T status	T1,T2/T3,T4	1.688	0.236	0.710–4.009
Overall tumor recurrence	negative/positive	0.011	0.0001	0.001–0.093
Local tumor recurrence	negative/positive	0.707	0.401	0.315–1.588
CD68-positive TAMs in the tumor invasive front	high/low	2.752	0.016	1.207–6.275
Tumor necrosis	negative/positive	0.244	0.002	0.100–0.596
Multivariate analysis (Recurrence-free survival)
Variable	Category	Odds ratio	p	Confidence interval
Tumor nodules	solitary/multiple	0.301	0.001	0.149–0.606
Perineural sheet infiltration	negative/positive	0.370	0.008	0.178–0.772
Histologic differentiation	well or moderate/ poor	2.123	0.047	1.010–4.460
Lymph node Involvement	negative/positive	0.542	0.066	0.282–1.042
CD68-positive TAMs in the tumor invasive front	high/low	0.558	0.094	0.282–1.104

Influence of TAMs invasion, tumor necrosis and peritumoral fibrosis on overall and recurrence-free survival in intrahepatic cholangiocarcinoma

Based on the identification of TAMs in TIF and tumor necrosis as independent prognostic factors in both the univariate and multivariate analysis, we investigated their effect on patients' survival in iCC. The presence of TAMs in TIF affected patients' survival after resection for iCC. Univariate analysis revealed a significantly better survival in the Invasive frontCD68-high group (p = 0.003). The overall survival rates in the Invasive frontCD68-high group were 74.9%, 47.5% and 41.8% at 1-, 3- and 5-y following liver resection in comparison with 60.2%, 15.1% at 1-, 3-y following resection in the Invasive frontCD68-low group (p = 0.008) (Fig. 1J). None of the patients in the Invasive frontCD68-low group could reach 5-y survival after surgery. Moreover, a high density of CD68-positive cells in TIF correlated with a significantly improved recurrence-free survival, as well. 1-, 3- and 5-y recurrence-free survival rates in the Invasive frontCD68-high group were 57.9%, 41.6% and 41.6% compared with 38.8%, 8.7% in the Invasive frontCD68-low group, respectively (Fig. 1P). Again, none of the patients in the Invasive frontCD68-low group could reach 5-y recurrence-free survival after surgery. The differences were statistically significant (p = 0.002).

Figure 1.

(Continued)

Formation of histologic tumor necrosis affected patients' survival, as well. Univariate analysis revealed significantly better results for patients with absence of histologic tumor necrosis (p = 0.037). Kaplan–Meier method revealed that patients in the Necrosisnegative group have a significantly improved overall survival, when compared with Necrosispositive group (Fig. 1K) (p = 0.034).

Figure 1.

(Continued)

Overall and recurrence-free survival of patients with iCC were significantly better in patients with high levels of TAMs in the area of TIF and absence of tumor necrosis when compared with those with a high density of TAMs and presence of histologic necrosis in tumor. Formation of severe histologic tumor necrosis showed a trend to a reduced overall and recurrence-free survival, as well (p = 0.135 and p = 0.198, respectively; data not shown). Interestingly, low TAMs invasion in TCA and formation of strong peritumoral fibrosis showed distinct trends toward an improved overall and recurrence-free survival, though without reaching statistical significance (p = 0.076, p = 0.173 and p = 0.260, p = 0.321, respectively) (Figs. 1Q and R).

Figure 1.

(Continued)

Discussion

Analyzing immunoreactivity for TAMs, formation of tumor necrosis and peritumoral fibrosis in tumor samples from patients who underwent liver resection for iCC, we were able to show that formation of mild peritumoral fibrosis is a risk factor for lymph node involvement, low abundance of CD68 in TIF and histologic tumor necrosis are risk factors for tumor recurrence and formation of multiple tumor nodules, serve as independent prognostic factors for survival, and translate into poor overall as well as recurrence-free survival.

The significance of tumor-related immunologic responses with regard to disease progression and patient prognosis has long been recognized. In the field of solid organ transplantation, novel immunologic therapeutic and diagnostic tools have gained importance and been successful translated from bench to bedside.^27,28 However, in clinical oncology only marginal practical approach on clinical daily basis has been implemented in recent years.^29-31 Overwhelming experimental and clinical evidence suggests a key role of innate and adaptive immune system in tumor progression.^32-34 Forasmuch, novel diagnostic and therapeutic strategies are urgently needed to tailor a personalized immunologic approach in clinical management of underlying oncological disease.

The tumor mass and microenvironment represent a multifaceted playground, where different cell types, including cancer cells, fibroblasts, endothelial, stromal and immune-competent cells are entangled into a constant interaction. Cholangiocarcinoma exhibits a desmoplastic nature, associated with a massive presence of TAMs. Cholangiocarcinoma-associated macrophages represent a poorly defined but very intriguing immune subset. Indeed, TAMs density significantly impacts prognosis, survival and development of metastasis in cholangiocarcinoma patients, suggesting a major role of macrophages in progression of this tumor entity. Additionally, TAMs polarization is influenced by both macrophage localization and TME signals. Hence, related to their localization in specific tumor areas, macrophages can exert opposing effects, either promoting or inhibiting tumor progression.^2,35,36

In the present work, we hypothesized that based on TAMs localization in TCA or TIF a different impact on tumor progression and patient outcome may result. Our findings demonstrate that in a highly aggressive tumor entity like iCC, high TAMs frequencies in TIF associate with improved patient outcome. In contrast, we found that a high TAMs density in TCA reveals a strong tendency to a deteriorated survival. These results are similar to reports regarding colon, gastric, and endometrial cancers, which showed that higher densities of tumor-infiltrating macrophages in tumor-peripheral areas correlate with a favorable prognosis.^37-39 Moreover, similar to our results, a high number of TAMs in tumor-central area were correlated with a worse prognosis in endometrial cancer.³⁹

A possible explanation for these findings is that macrophage function may be altered by hypoxia and formation of necrosis, i.e., through increased expression of hypoxia-inducible factors like (HIF) 1-α and HIF2-α.^40,41 Oxygen concentration and gradients vary between central and peripheral areas of solid tumors, affecting TAMs polarization state depending on their location. In particular, both HIF-2α accumulation in immune cells and hypoxia activate molecular mechanisms fostering M2 polarization, and thus promotion of tumor progression.^41,42

In the present work, we furthermore propose histologic tumor necrosis as a simple diagnostic tool prognosticating patients' outcome after surgery for iCC. To date, there is only a scarce amount of data regarding the association between tumor necrosis and various clinicopathological characteristics in solid tumors. Recently published data suggest a prognostic value of histological tumor necrosis in solid organ malignant disease.^19,20,43-45 Our group was the first to report on the prognostic value of histologic tumor necrosis in patients with HC.²⁴ In the current work, tumor necrosis was associated with survival and was identified as independent prognosticator for iCC, as well.

Hepatic macrophages are central in the pathogenesis of chronic liver injury, but are crucially involved in maintaining homeostasis, as well.¹⁷ Tumor-promoting properties of TAMs may reflect, at least to some extent, normal features of resident macrophages rather than functions induced by TME.⁴⁶ In most cases formation of tumor necrosis is related to creation of hypoxic conditions in aggressive and fast-growing tumors. The cause for necrosis formation in iCC and its mechanistic link to poor clinical outcome may be related to immunologic factors, such as infiltrating TAMs, because larger tumors did not associate with presence of necrosis in the current work. We found an outlined tendency to formation of histologic tumor necrosis and high TAMs densities in TCA and TIF, respectively. Moreover, these results are in correspondence with our previous findings linking macrophage density to formation of histologic tumor necrosis in patients with HC, and with novel experimental results highlighting a fundamental role of macrophages in creation of fibrosis and necrosis via death receptor, chemokine and fibrocystin/polyductin-dependent mechanisms.^{17,23,24,47,48}

In conclusion, in the current study, we demonstrate for the first time the significance of tumor necrosis and its association with survival and patients' outcome after liver resection for iCC. Moreover, we show that presence of CD68-positive TAMs in TIF correlates with reduced tumor recurrence and serves as an independent prognostic factor for survival, as well. Thus, TAMs and formation of histological tumor necrosis may be crucially involved in the progression of this tumor entity, and their utilization as a diagnostic tool may deliver new approaches in the management of this cancer. Moreover, the current results confirm our findings in HC, which may implicate similar therapy regimes for both tumor entities.

Methods

Patients and tumor samples

A total of 88 patients who underwent major hepatectomy between January 1996 and December 2002 for iCC were included in the study. ICC was confirmed histopathologically and classified according to the Union for International Cancer Control (UICC) classification. Written-informed consent was obtained from all patients. This study was approved by the ethics committee of Charité – Universitätsmedizin Berlin (ID: 111 EA1/318/15).

In all included patients, liver surgery was performed in formally curative intent. None of the patients received neoadjuvant therapy before operation. None of the patients died in the postoperative course. Histological evaluation of all specimens was performed by an independent pathologist, without any knowledge of prognosis or clinicopathological variables.

Formalin-fixed, paraffin-embedded tumor samples with a representative sample of the tumor were retrieved from the files of the Institute of Pathology. Histological diagnosis of the primary tumor stage and nodal status were determined by hematoxylin and eosin (H&E) stained sections. The clinicopathological characteristics of the study population are depicted in Table 1.

Immunohistochemistry

All used protocols in regard to immunohistochemistry, histology, quantification of cell infiltrates, necrosis and peritumoral fibrosis have been described previously.^23,24

In detail, formalin-fixed and paraffin-embedded tumor sections (5-µm thick) were dewaxed and rehydrated. Antigen retrieval was performed by heating the slides in 10 mM Tris buffer with 1 mM EDTA (pH 9) in a streamer for 20 min. Endogenous peroxidase activity was inhibited with 3% H₂O₂ for 5 min. After washing with Tris buffered saline (TBS) with tween, the endogenous biotin was suppressed by sequential incubations with 0.1% avidin and 0.01% biotin (Dako, Glostrup, Denmark) for 10 min each at room temperature. Additional non-specific binding sites were blocked with 3% skimmed milk powder for 30 min at room temperature. Tissue sections were incubated with the monoclonal mouse antibody anti-human CD68 Clone PG-M1 (1:50, Dako, Glostrup, Denmark) for 30 min at room temperature. The universal LSAB+ system-HRP (Dako, Glostrup, Denmark) and the DAB+ liquid substrate chromogen system (Dako, Glostrup, Denmark) was applied for the visualization of the antibody reaction. The biotinylated anti-rabbit, anti-mouse and anti-goat immunoglobulin and the streptavidin-conjugated horseradish peroxidase of the universal LSAB+ system was incubated consecutively for 20 min each. Sections were counterstained with hematoxylin. Specificity controls were performed without primary antibody.

Quantification of CD68 density

All specimens were evaluated by three independent researchers (GA, CD and IP), and an independent pathologist (KS), without any knowledge of prognosis or clinicopathological variables. TAMs were defined by their expression of CD68. For quantification of infiltrating TAMs, the whole tumor area was thoroughly investigated for presence of CD68-positive cells. The evaluation of TAMs density was performed as percentage of CD68 stained/positive cells related to total tumor cell amount.

The density of TAMs in TCA was semi-quantitatively classified into the following categories: 0, negative; 1, 1–25%; 2, 26–75%; and 3, >75%. For statistical analysis, TCA scores 0 and 1 were categorized as low density, and TCA scores 2 and 3 as high density of TAMs. The density of TAMs in TIF were semi-quantitatively classified into the following major categories: 0, negative; 1, 1–25% and 2, >25%. For statistical analysis, TIF scores 0 and 1 were categorized as low density or negative, TIF score 2 as high density (or positive).

Occurrence of tumor necrosis

The degree of necrosis was classified according to the amount of necrotic cells/areals of the slice in relation to vital tissue. The following degrees of necrosis were established: none (0%), 5%, 10%, 15%, 20%, 25%, 30%, 35% and 40%. The presence of histological tumor necrosis in the studied specimens was classified into two categories: 1, negative (< 5%) or 2, positive (> 5% necrosis). Patients were divided into two groups according to their “positive” or “negative” necrosis scores. In a second step, a further discrimination of the “positive” group into mild (Necrosismild group; 5–20% necrosis) and severe necrosis (Necrosissevere group; >20%) was undertaken.

Occurrence of peritumoral fibrosis

Degree of peritumoral fibrosis in the examined tumor specimens was classified into three categories: mild, moderate and high.

Patients were divided into two groups either by “mild” or “strong (moderate/high)” fibrosis scores (Fibrosismild group or Fibrosisstrong group).

Statistical analysis

Survival analysis, univariate analysis and Kaplan–Meier curves were generated with assistance of the SPSS software program (Version 23.0.0.0 / Year 2015). Comparison of categorical and continuous variables was performed using the Chi²-test and the Wilcoxon-test, respectively. Survival data were compared with the log rank-test. Variables with a significant influence on survival in the univariate analysis were entered into a cox regression analysis. A difference was considered significant for p < 0.05.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The work presented in this paper was made possible by funding from the Berlin Institute of Health (BIH). Georgi Atanasov is participant of the BIH Charité Clinician Scientist Program funded by the Charité - Universitätsmedizin Berlin and the Berlin Institute of Health.