Macrotrabecular-massive hepatocellular carcinoma: A distinctive histological subtype with clinical relevance

Liver cancer ranks as the fifth most frequent cancer worldwide and the second most common cause of cancer-related death.1, 2 Hepatocellular carcinoma (HCC) is the most frequent primary liver tumor, and its main risk factors are hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, alcohol consumption, and metabolic syndrome.2, 3 Clinical outcome of HCC remains poor: the majority of patients present with advanced disease, and curative therapies, such as surgical resection and radiofrequency ablation (RFA), are hampered by high rates of tumor recurrence (50%-70%).3

Over the last decade, the emergence of high-throughput molecular technologies has allowed us to define the main molecular mechanisms involved in HCC development and progression. HCC appears as a highly heterogeneous entity, composed of distinct transcriptomic subgroups with various genetic alterations.4-9

A high degree of heterogeneity can also be observed at the histological level. Indeed, although neoplastic cells most often mimic the trabeculae of normal liver, numerous cytological variants and architectural patterns are recognized by the World Health Organization.10 Tumor differentiation is also assessed according to a three-grade system combining architectural and nuclear features.10, 11 Several histological subtypes, which feature distinctive and recognizable morphological features, have also been reported, such as the scirrhous subtype, characterized by a fibrous, hyaline stroma, and the lymphoepithelioma-like subtype, defined by a dense intratumor lymphocytic infiltration.12-14 Although some histological features have been shown to influence recurrence, their independent prognostic value, compared to microvascular invasion and satellite nodules, is not powerful enough to be included in therapeutic algorithms.15

We have recently shown that HCC morphology is tightly associated to specific underlying molecular alterations16 and noticeably identified a histological subtype, designated as “macrotrabecular-massive” (MTM-HCC), associated with an aggressive phenotype and particular biological (high alpha-fetoprotein [AFP] serum levels) and molecular (G3 transcriptomic subgroup, TP53 mutations, and FGF19 amplifications) features.16

Consistent with its microscopic appearance, which consists of thick trabeculae coated by endothelial cells and surrounded by vascular spaces, MTM-HCC was also characterized by high expression of angiopoietin 2 (Ang2) and vascular endothelial growth factor A (VEGFA), two key regulators of neoangiogenesis and vascular remodeling.16-18

In order to assess the clinical relevance of this particular histological HCC subtype, we aimed in the present work to validate its independent relationship with patient outcome in an independent series of patients treated by surgical resection and to determine its prognostic value when identified in biopsy samples of patients treated by RFA.

Materials and Methods

Results

PATIENTS TREATED BY SURGICAL RESECTION

Among the 237 patients, 81% were male. The main risk factors were HCV infection (32%), HBV infection (25%), and alcohol intake (23%) (Table 1). Mixed etiologies were observed in 13% of the patients. Disease stage was very early/early (BCLC 0/A) and intermediate/advanced (BCLC B/C) in 85% and 15% of the cases, respectively.

Table 1. Clinical, Biological, and Pathological Features of the Patients and the Tumors According to MTM-HCC Subtype (Surgical Series)

Variables	Available Data (n)	n (%)	MTM-HCC n (%)	Non-MTM-HCC n (%)	P
Clinical and biological features
Age >60 years	237	133 (56)	17/38 (45)	116/199 (59)	0.17
Male gender	237	193 (81)	27/38 (71)	165/199 (83)	0.12
Alcohol	237	52 (23)	3/35 (8)	49/193 (25)	0.05
HCV	228	73 (32)	12/35 (34)	61/193 (32)	0.91
HBV	228	58 (25)	15/35 (43)	43/193 (22)	0.02
NASH	228	41 (18)	2/35 (6)	39/193 (20)	0.05
Other/undetermined	228	33 (15)	6/35 (17)	27/193 (14)	0.62
BCLC B-C	237	35 (15)	15/38 (39)	20/199 (10)	<0.001
AFP serum level >100 ng/mL	191	55 (29)	19/26 (73)	36/165 (22)	<0.001
Pathological features of tumors and nontumoral liver
Tumor size >50 mm	237	108 (45)	25/38 (66)	83/199 (42)	0.01
Satellite nodules	237	61 (26)	18/38 (47)	43/199 (22)	0.001
Macrovascular invasion	237	31 (13)	15/38 (39)	16/199 (8)	<0.001
Microvascular invasion	237	111 (47)	34/38 (89)	77/199 (39)	<0.001
Poor differentiation	237	36 (15)	13/38 (34)	23/199 (11)	<0.001
Cirrhosis	231	94 (39)	18/36 (50)	76/195 (39)	0.29

• Abbreviation: NASH, nonalcoholic steatohepatitis.

Elevated preoperative AFP serum levels (>100 ng/mL) were detected in 29% of the cases (Table 1). As expected in a series of patients with HCC treated by surgical resection, a high rate of noncirrhotic livers was observed (F0-F3 61%) (Table 1).

Tumors were large (>50 mm, 45%), with satellite nodules, macrovascular invasion, and microvascular invasion in 26%, 13%, and 47% of the cases, respectively.

Thirty-eight cases of MTM-HCC were identified, representing 16% of the whole series (Fig. 1). Interobserver agreement between the three pathologists for the diagnosis of MTM-HCC was good (Fleiss kappa 0.74).

Regarding clinical and biological features, MTM-HCC was associated with HBV infection (P = 0.02), BCLC B-C disease stages (P < 0.001), and elevated AFP serum levels (P < 0.001). At the pathological level, MTM-HCCs were more frequently large (>50 mm, P = 0.01), with satellite nodules (P = 0.001), macrovascular invasion (P < 0.001), and microvascular invasion (P < 0.001) (Table 1).

Ninety-one recurrences were observed, including 61 early recurrences (67%). Variables associated with early tumor recurrence in univariate analysis were HBV infection (hazard ratio [HR], 1.86; P = 0.02), BCLC B-C stages (HR, 3.93; P < 0.001), AFP serum levels >100 ng/mL (HR, 2.53; P = 0.001), satellite nodules (HR, 2.89; P < 0.001), microvascular invasion (HR, 3.60; P < 0.001), poor tumor differentiation (HR, 2.25; P = 0.008), and MTM-HCC subtype (HR, 5.81; P < 0.001) (Table 2). Multivariate analysis showed that MTM-HCC was an independent predictor of both early (P = 0.006) and overall (P < 0.001) recurrence. MTM-HCC subtype retained a prognostic value even after stratification according to common clinical, biological, and pathological features of aggressiveness (Table 2 and Fig. 2). Indeed, patients with MTM-HCC and small tumors (n = 13), without satellite nodules (n = 20), classified BCLC 0/A (n = 23), or without vascular invasion (n = 4) had significantly lower event-free survival compared to patients with other histological subtypes (Fig. 2). The other variable independently associated with overall recurrence was satellite nodules (P = 0.03) (Table 2).

Table 2. Univariate and Multivariate Analyses of Clinical, Biological, and Pathological Features for Early and Overall Recurrence (Surgical Series)

	Early Tumor Recurrence				Overall Tumor Recurrence
	Univariate Analysis		Multivariate Analysis		Univariate Analysis		Multivariate Analysis
Variables	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
Clinical and biological features
Age >60 years	0.59 (0.36-0.98)	0.04	0.93 (0.44-1.96)	0.86	0.89 (0.59-1.35)	0.59
Male gender	1.23 (0.62-2.43)	0.54			0.89 (0.53-1.50)	0.67
BCLC stage B-C	3.93 (2.26-6.84)	<0.001	2.27 (0.82-6.23)	0.11	2.71 (1.65-4.47)	<0.001	1.70 (0.89-3.21)	0.10
AFP serum level >100 ng/mL	2.53 (1.42-4.51)	0.001	1.12 (0.57-2.17)	0.74	1.51 (0.91-2.52)	0.11
Alcohol	0.57 (0.28-1.17)	0.12			0.61 (0.34-1.10)	0.09	0.62 (0.34-1.14)	0.13
HCV	0.63 (0.35-1.16)	0.13			0.90 (0.57-1.42)	0.65
HBV	1.86 (1.09-3.15)	0.02	1.50 (0.71-3.15)	0.28	1.23 (0.77-1.95)	0.38
NASH	0.37 (0.15-0.92)	0.03	0.30 (0.09-1.00	0.05	0.58 (0.31-1.10)	0.09	0.63 (0.33-1.20)	0.16
Other/undetermined	1.39 (0.71-2.76)	0.33			1.31 (0.73-2.38)	0.36
Histological and gross features of the tumors
Tumor size >50 mm	1.51 (0.91-2.50)	0.10			1.26 (0.83-1.90)	0.27
Satellite nodules	2.89 (1.72-4.85)	<0.001	1.30 (0.55-3.07)	0.54	2.45 (1.56-3.83)	<0.001	1.89 (1.07-3.31)	0.03
Microvascular invasion	3.60 (2.07-6.24)	<0.001	1.94 (0.90-4.17)	0.08	2.00 (1.32-3.03)	<0.001	1.14 (0.70-1.87)	0.60
Poor differentiation	2.25 (1.22-4.17)	0.008	0.68 (0.28-1.65)	0.39	1.45 (0.83-2.53)	0.19
Macrotrabecular-massive subtype	5.81 (3.46-9.76)	<0.001	3.03 (1.38-6.65)	0.006	3.83 (2.40-6.11)	<0.001	2.76 (1.63-4.67)	<0.001
Cirrhosis	1.20 (0.72-2.01)	0.48			1.24 (0.81-1.88)	0.32

• Abbreviations: CI, confidence interval; NASH, nonalcoholic steatohepatitis.

PATIENTS TREATED BY RFA

As observed in the surgical series, a strong male predominance was observed among the 284 patients (75%), with HCV infection and alcohol intake being the most frequent risk factors of liver disease (Table 3). Preoperative AFP was elevated in 13% of the patients. Tumor size at imaging was >30 mm in 26% of the cases.

Table 3. Clinical, Biological, and Pathological Features of the Patients and the Tumors According to MTM-HCC Subtype (RFA Series)

Variables	Available Data (n)	n (%)	MTM-HCC n (%)	Non-MTM-HCC n (%)	P
Clinical and biological features
Age >60 years	284	203 (71)	18/24 (75)	185/260 (71)	0.87
Male gender	284	213 (75)	20/24 (83)	193/260 (74)	0.46
Alcohol	284	129 (45)	12/24 (50)	117/260 (45)	0.80
HCV	284	135 (47)	12/24 (50)	123/260 (47)	0.97
HBV	284	23 (8)	1/24 (4)	22/260 (8)	0.70
NASH	284	52 (18)	6/24 (25)	46/260 (18)	0.54
Other/undetermined	284	8 (3)	1/24 (4)	7/260 (3)	0.67
AFP serum level >100 ng/mL	262	36 (13)	8/23 (35)	28/211 (13)	0.005
Pathological features
Tumor size >30 mm	284	74 (26)	11/24 (46)	63/260 (24)	0.04
Poor differentiation	284	46 (16)	7/24 (33)	39/260 (15)	0.13

• Abbreviation: NASH, nonalcoholic steatohepatitis.

Pathological review of biopsy samples identified 24 cases (8.5%) of MTM-HCC, with an excellent interobserver agreement (Cohen kappa, 0.82). MTM-HCC subtype samples were easily identified at low-power field examination of biopsies due to the characteristic dyscohesive aspect of trabeculae (Fig. 3).

Cases classified as MTM-HCC were associated with high AFP serum levels (P = 0.005) and high tumor size (>30 mm, P = 0.04) (Table 3).

One-hundred and ninety-four recurrences were observed, including 135 (69%) early recurrences. MTM-HCC subtype was an independent predictive factor of both early and overall recurrence in multivariate analysis (HR, 2.37; P = 0.002; and HR, 2.19; P = 0.001, respectively) (Table 4). No other clinical, biological, or histological factor was related to recurrence.

Table 4. Univariate and Multivariate Analyses of Clinical, Biological, and Pathological Features for Early and Overall Recurrence (RFA Series)

	Early Tumor Recurrence				Overall Tumor Recurrence
	Univariate Analysis		Multivariate Analysis		Univariate Analysis		Multivariate Analysis
Variables	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
Clinical and biological features
Age >60 years	1.01 (0.69-1.49)	0.94			1.26 (0.90-1.77)	0.17
Male gender	1.20 (0.79-1.82)	0.39			1.21 (0.86-1.71)	0.21
AFP serum level >100 ng/mL	1.55 (0.98-2.49)	0.06	1.32 (0.80-2.17)	0.27	1.24 (0.82-1.88)	0.30
Alcohol	0.82 (0.59-1.16)	0.27			0.86 (0.65-1.15)	0.32
HCV	1.04 (0.74-1.45)	0.83			1.08 (0.81-1.43)	0.61
HBV	0.98 (0.53-1.81)	0.94			0.97 (0.58-1.62)	0.91
NASH	1.25 (0.82-1.89)	0.29			1.05 (0.72-1.51)	0.81
Other/undetermined	0.61 (0.19-1.91)	0.40			0.51 (0.21-1.25)	0.13
Histological and gross features of tumors
Tumor size >30 mm	1.26 (0.87-1.81)	0.22			1.21 (0.89-1.65)	0.22
Poor differentiation	1.45 (0.94-2.22)	0.09	1.33 (0.84-2.11)	0.22	1.43 (0.99-2.07)	0.06	1.38 (0.95-1.99)	0.09
Macrotrabecular-massive subtype	2.40 (1.44-4.00)	<0.001	2.37 (1.36-4.13)	0.002	2.23 (1.38-3.60)	<0.001	2.19 (1.35-3.54)	0.001

• Abbreviations: CI, confidence interval; NASH, nonalcoholic steatohepatitis.

Discussion

The present work demonstrates that the macrotrabecular-massive subtype of HCC is robustly recognized by pathologists on resected and biopsied samples. We further show, in two large series of HCC patients treated by resection or RFA, that the identification of this histological entity has a major impact on clinical outcome, exceeding known predictors such as size, vascular invasion, satellite nodules, BCLC stage, and AFP levels. This subtype, observed in 16% of surgical samples and 8.5% of biopsy samples, independently predicted early and overall tumor recurrence in both investigated series.

The definition of this particular subtype of HCC, which represents 12% of our overall series of tumors, relies on morphological features; and overall good and excellent interobserver agreements were observed among the different pathologists involved in the study. Therefore, this tumor variant may be reliably diagnosed in both surgical and small biopsy samples without specific training.

We were able to confirm the association of MTM-HCC with various adverse biological and pathological parameters (AFP serum level, tumor size, vascular invasion, and satellite nodules) and to show that this subtype was a key determinant of patient prognosis, with a strong and independent predictive value for early and overall tumor recurrence in both investigated series, even after stratification according to main classical HCC prognostic indicators. Interestingly, the relative risk of relapse associated with MTM-HCC appeared much higher than that associated with classical HCC outcome predictors, such as AFP serum levels, vascular invasion, or satellite nodules. Of note, although this variant is more prevalent in intermediate or advanced HCC, it still represents a significant fraction (10%) of tumors classified as very early or early (BCLC 0/A).

The aggressiveness of this variant may be related to molecular determinants. We indeed previously showed that this subtype was characterized by an activation of neoangiogenesis, with both Ang2 and VEGFA overexpression.16 Initially identified as an antagonist of Ang1, which promotes vascular stabilization by recruiting pericytes and smooth muscle cells to growing blood vessels, Ang2 is known to promote neoangiogenesis in cooperation with VEGFA and is an indicator of adverse outcome in various solid tumors.22 Ang2 promotes vascular sprouting and destabilizes blood vessels by disrupting interactions between endothelial cells and periendothelial cells, which results in increased sensitivity to VEGFA and thus favors neoangiogenesis activation.18, 23

It has also been shown that secretion of Ang2 by various primary solid tumors is capable of inducing metastasis.24 Indeed, by loosening endothelial cell junctions and increasing the vascular permeability of distant organs, Ang2 allows subsequent extravasation of circulating tumor cells and is considered a key regulator of the premetastatic niche.24-26 Satellite nodules are neoplastic nodules of HCC located <2 cm away from the main tumor and are usually considered early intrahepatic metastases,11 and we may thus hypothesize that Ang2 produced by neoplastic cells may disseminate in the directly adjacent liver parenchyma and contribute to the increased frequency of satellite nodules observed in MTM-HCC.

The underlying oncogenic pathways associated with this particular HCC subtype may also open therapeutic perspectives, and further studies will have to determine if MTM-HCC may be sensitive to potent inhibitors of neoangiogenesis. Noticeably, bispecific anti-Ang2 and anti-VEGFA antibodies have been reported to induce strong angiogenesis inhibition and tumor necrosis in various preclinical models of human malignancies.23, 27-29 Interestingly, Schmittnaegel et al. also demonstrated that blocking angiogenesis using these types of antibodies enhanced antitumor immunity. Indeed, the normalization and maturation of blood vessels induced by the treatment increased extravasation and perivascular accumulation of antitumor cluster of differentiation 8+, interferon-γ-producing cytotoxic T lymphocytes.28

The role of tumor biopsy for the management of patients with HCC is one of the most active debates in the liver cancer community.30-33 European Association for the Study of the Liver and American Association for the Study of Liver Diseases guidelines currently do not require histological confirmation due to the reported excellent diagnostic performance of imaging procedures.34, 35 Moreover, although numerous histological tumor variants have been reported over the last decades, they did not demonstrate robust clinical impact and have thus aroused no or very little interest for physicians involved in the clinical care of patients with HCC. We provide in the present study evidence that MTM-HCC can be accurately diagnosed in biopsy samples and believe that the identification of this very aggressive subtype may lead to a reconsideration of the role of biopsy for HCC. Indeed, the identification of this variant during pretherapeutic workup has strong prognostic implications, and patients with this particular HCC subtype may benefit from adjuvant therapies and/or upfront registration on liver transplant waiting lists after resection or RFA. However, further studies are needed to assess the prognostic value of this subtype on recurrence after transplantation. In the same line, the activation of angiogenesis in this subset of tumors may also result in distinctive imaging features, and radiologic–pathologic correlations will have to determine if MTM-HCC may be diagnosed or at least suspected by computed tomography or magnetic resonance imaging.

In conclusion, we have demonstrated the value of the MTM-HCC subtype in predicting tumor recurrence in two large series of patients treated by distinct curative therapeutic modalities. We additionally show that this histological variant can reliably be identified on biopsy samples. Representing 12% of our overall series of patients treated by resection or RFA, we believe that MTM-HCC should be recognized as a distinct group of tumors that may require more personalized and aggressive treatments than conventional HCC.

Author names in bold designate shared co-first authorship.