Clinicopathologic features and prognosis of steatohepatitic hepatocellular carcinoma based on varying cutoffs of tumoral steatohepatitic changes

Abstract

Objectives

Steatohepatitic hepatocellular carcinoma (SH-HCC) is currently recognized as a distinct histologic subtype of HCC. The prognosis and specific criteria for determining the amount of steatohepatitis required to define SH-HCC are still unclear.

Methods

After excluding all recognized HCC subtypes from 505 HCC cases (2010-2019), the remaining cases were categorized as conventional HCC (CV-HCC) (n = 223). The cases classified as SH-HCC (n = 171) were further divided into groups based on the percentage of steatohepatitis: 5% or more, 30% or more, and 50% or more.

Results

Hepatitis C virus infection was the predominant underlying liver disease in both the CV-HCC and SH-HCC groups. Metabolic dysfunction-associated steatotic liver disease (formerly nonalcoholic fatty liver disease) was more prevalent in all cases of SH-HCC with different steatohepatitic cutoffs than in cases of CV-HCC. There were no differences in the stage of fibrosis of the background liver between the CV-HCC and SH-HCC groups. SH-HCC with different cutoffs exhibited a notable increase in the presence of glycogenated nuclei, Mallory-Denk bodies, and hyaline globules in tumor cells. Survival analysis did not reveal substantial differences in overall survival between the CV-HCC and SH-HCC groups and among patients with SH-HCC with different steatohepatitis cutoffs.

Conclusions

The degree of intratumoral steatohepatitis in patients with SH-HCC does not appear to be a notable prognostic factor. The presence of steatohepatitis in the tumor is better recognized as 1 of the histopathologic patterns of HCC.

Key Points.

• There was a notable increase in the degree of steatosis in SH-HCCs compared with CV-HCC but no difference in fibrosis stage in the background liver.

• There were no notable differences in overall survival between patients with CV-HCC and SH-HCC nor among patients with SH-HCC at different steatohepatitic cutoffs.

• The presence of steatohepatitis in the tumor is better recognized as 1 of the histopathologic patterns of HCC rather than a distinct subtype.

Introduction

Hepatocellular carcinoma (HCC) is a heterogeneous tumor with various etiopathogenesis and histopathologic morphologies and subtypes.¹ Recognition of distinct HCC subtypes is essential for improving our understanding of disease pathogenesis and tailoring treatment strategies. One such subtype recognized by the 2019 World Health Organization (WHO) classification of digestive system tumors,² steatohepatitic HCC (SH-HCC), is characterized by the presence of “steatohepatitis-like” features, including steatosis (≥5%), ballooning of tumor cells, intratumoral inflammation, Mallory-Denk bodies, and fibrosis.^3,4 The prevalence of SH-HCC among all HCCs ranges from 8.7% to 59.1%.^3,5-7 The wide range in prevalence reported in different studies may reflect variations in patient populations and underlying etiologies. More importantly, it may be attributed to the different thresholds used to define SH-HCC. The pathologic classification of SH-HCC, particularly the threshold for the proportion of steatohepatitis required for classification, remains a topic of debate, with thresholds ranging from 5% to 50%.^4,7-9 This ambiguity has prompted further investigation into whether establishing a specific threshold is clinically useful and necessary.

In addition, SH-HCC has been linked to metabolic dysfunction-associated steatotic liver disease (MASLD—formerly nonalcoholic fatty liver disease) and alcohol-associated liver disease (ALD) in patients with or without chronic viral hepatitis. This discovery highlights the intricate interplay between multiple risk factors and liver carcinogenesis.^3,10 Similar to other HCC subtypes, the proper classification of SH-HCC, whether it is a specific subtype or just a histologic pattern, should be determined based on its clinicopathologic significance and underlying molecular changes. Despite the growing interest in SH-HCC, its histopathologic characteristics, especially those associated with prognosis, remain poorly understood. Although studies have shown no clinically significant prognostic impact,^4-6 the prognostic significance and histopathologic features warrant further investigation on a larger scale. The aim of this study was to further clarify the distinctive clinical features, association with MASLD and other etiologies, outcome, and correlation with certain proportions of steatohepatitic features in SH-HCC.

Methods

Case Selection

Cases of HCC diagnosed between 2010 and 2019 were searched from our pathology database, and 505 HCC cases were identified (217 resections, 288 biopsies). Clinical and laboratory data, including age, sex, underlying liver diseases (ALD, MASLD, hepatitis B virus [HBV] infection, hepatitis C virus [HCV] infection, other risk factors [eg, autoimmune hepatitis, primary biliary cholangitis, hemochromatosis], and unknown etiology), and overall survival (the interval between the date of diagnosis and death) were retrieved from electronic health records.

Histopathologic Examination

Pathology reports and slides for all cases were reviewed. Based on the WHO classification, HCC cases were categorized into specific subtypes (steatohepatitic, clear cell, macrotrabecular massive, scirrhous, chromophobe, fibrolamellar, neutrophil rich, and lymphocyte rich), and the remaining cases were classified as conventional HCC (CV-HCC).² Steatohepatitic HCC was diagnosed when a tumor consisted of steatosis (≥5%); ballooning of tumor cells; and varying combinations of intratumoral inflammation, fibrosis, and Mallory-Denk bodies FIGURE 1, regardless of whether these steatohepatitic features were focal or diffusely distributed in the tumor.^3,4,9 Steatohepatitic HCC cases were further divided into 3 groups based on different cutoff thresholds for steatohepatitis (≥5%, ≥30%, and ≥50%) in the tumor. The tumor size, histologic grade, presence or absence of vascular invasion, glycogenated nuclei of tumor cells, Mallory-Denk bodies, cytoplasmic hyaline globules, intratumoral granuloma, and the degree of intratumoral lymphocytic and neutrophilic inflammation were also evaluated. In addition, to investigate whether HNF1A inactivation is involved in the development of SH-HCC, immunohistochemical staining for liver fatty acid-binding protein (L-FABP) was performed on 10 representative SH-HCC cases with 50% or more steatohepatitic features in the tumor.

FIGURE 1.

Representative steatohepatitic hepatocellular carcinoma showing steatosis, tumor cell ballooning, inflammation, Mallory-Denk bodies, and pericellular fibrosis (H&E, original magnification ×200).

Histopathology of the background liver was assessed as previously described.¹¹ In brief, for patients lacking separate biopsy specimens from the background liver taken concurrently with or within 6 months of the HCC diagnosis, the background liver was assessed using the same biopsy specimens as the HCC, provided there was sufficient background liver tissue. In resection cases, the background liver was evaluated on sections taken as far from the tumor as possible. Care was taken to avoid interpreting changes due to a mass effect. When necessary, correlation with imaging findings to confirm the stage of fibrosis was pursued. The fibrosis stage of the background liver was evaluated using the Batts and Ludwig system.¹² The degree of steatosis was assessed and graded as less than 5%, 5% to 33%, 33% to 66%, and more than 66%.¹¹

Statistical Analysis

Categorical data were compared using the χ² or Fisher exact test; continuous variables were assessed using the t test or Wilcoxon rank sum test. Survival analysis was conducted using Kaplan-Meier curves. Data analysis was performed using GraphPad Prism 8 software (GraphPad Software, LLC). P < .05 was considered statistically significant.

Results

Patient Characteristics

Among the 505 HCC cases, 171 (33.9%, with 92 resections) were identified as SH-HCCs when a 5% cutoff threshold was applied TABLE 1. Compared with the CV-HCC group (n = 223, with 70 resections), SH-HCC was identified with the 5% cutoff were more commonly in women (P = .009), but there was no statistically significant (P = .94) difference in patient age. Regarding underlying liver diseases, HCV was the most common etiology in both the SH-HCC (80/171 [47%]) and CV-HCC (137/223 [61%]) groups, followed by MASLD (SH-HCC, 41/171 [24%] vs CV-HCC, 21/223 [9%]) and ALD (SH-HCC, 18/171 [11%] vs CV-HCC, 18/223 [8%]). Hepatitis C virus was more prevalent in the CV-HCC group (P = .004), whereas MASLD was more prevalent in the SH-HCC group (P < .001) TABLE 1.

TABLE 1.

Clinical Characteristics of CV-HCC vs SH-HCC at Different Steatohepatitis Cutoffs

Characteristics	CV-HCC (n = 223)	SH-HCC (5% cutoff) (n = 171)	P value	SH-HCC (30% cutoff) (n = 107)	P value	SH-HCC (50% cutoff) (n = 83)	P value
Age, mean (SD), y	64.3 (9)	64.4 (9)	.94	64.5 (9)	.87	63.6 (9)	.58
Sex (M:F)	191:32	128:43	.001**	85:22	.16	67:16	.29
Background liver disease, n/N (%)	HCV: 137/223 (61) ALD: 18/223 (8) MASLD: 21/223 (9) HBV: 10/223 (4) Other: 16/223 (7) Unknown: 21/223 (9)	HCV: 80/171 (47) ALD: 18/171 (11) MASLD: 41/171 (24) HBV: 3/171 (2) Other: 14/171 (8) Unknown: 23/171 (13)	.001*** (vs CV-HCC)	HCV: 47/107 (44) ALD: 15/107 (14) MASLD: 27/107 (25) HBV: 1/107 (1) Other: 10/107 (9) Unknown: 15/107 (14)	<.001*** (vs CV-HCC)	HCV: 35/83 (42) ALD: 12/83 (14) MASLD: 23/83 (28) HBV: 1/83 (1) Other: 8/83 (10) Unknown: 4/83 (5)	<.001*** (vs CV-HCC)

ALD, alcoholic liver disease; CV-HCC, conventional hepatocellular carcinoma; F, female; HCV, hepatitis C virus; HBV, hepatitis B virus; M, male; MASLD, metabolic dysfunction-associated steatotic liver disease; SH-HCC, steatohepatitic hepatocellular carcinoma.

*P < .05; **P < .01; ***P < .001.

We then investigated SH-HCC using higher cutoff thresholds for steatohepatitis: 30% (107/505 [21%], with 58 resections) and 50% (83/505 [16%], with 45 resections) TABLE 1. Similar to SH-HCC identified with the 5% cutoff, there was no statistically significant difference in age among patients with SH-HCC at the 30% (P = .12) or 50% (P = .58) cutoffs compared with patients with CV-HCC. As the cutoff threshold increased, sex differences between patients with SH-HCC using the 30% (P = .16) or 50% (P = .13) cutoff and patients with CV-HCC were no longer statistically significant. Regarding underlying liver diseases, HCV remained more prevalent in patients with CV-HCC (61%) than in patients with SH-HCC using the 30% (47/107 [44%], P = .003) or 50% (35/83 [42%], P = .003) cutoff; MASLD was still more prevalent in patients with SH-HCC and either a 30% (27/107 [25%], P < .001) or 50% (23/83 [28%], P < .001) cutoff than in patients with CV-HCC (21/223 [9%]).

We further compared whether there were differences in the clinical characteristics among the SH-HCC subgroups, defined by different cutoff thresholds (5%-29%, 30%-49%, and ≥50%) TABLE 2. There were no statistically significant differences in patient age (P = .79), sex (P = .49), or underlying liver diseases (P = .79) among the 3 SH-HCC subgroups.

TABLE 2.

Clinical Characteristics of SH-HCC at Different Percentages of Intratumoral Steatohepatitis

Characteristics	SH-HCC (5%-29%) (n = 64)	SH-HCC (30%-49%) (n = 24)	SH-HCC (≥50%) (n = 83)	P value
Age, mean (SD), y	64.9 (9)	65.1 (9)	63.6 (9)	.79
Sex (M:F)	43:21	18:6	67:16	.17
Background liver disease, n/N (%)	HCV: 33/64 (52) ALD: 3/64 (5) MASLD: 14/64 (22) HBV: 2/64 (3) Other: 4/64 (6) Unknown: 8/64 (13)	HCV: 12/24 (50) ALD: 3/24 (13) MASLD: 4/24 (17) Other: 2/24 (8) Unknown: 3/24 (13)	HCV: 35/83 (42) ALD: 12/83 (14) MASLD: 23/83 (28) HBV: 1/83 (1) Other: 8/83 (10) Unknown: 4/83 (5)	.45

ALD, alcoholic liver disease; F, female; HBV, hepatitis B virus; HCV, hepatitis C virus; M, male; MASLD, metabolic dysfunction-associated steatotic liver disease; SH-HCC, steatohepatitic hepatocellular carcinoma.

Histopathologic Characteristics

Using a 5% steatohepatitis cutoff to define SH-HCC, tumors were found to have a smaller size (mean, 3.31 cm) than in CV-HCC cases (mean, 4.22 cm) (P = .007), but when defining SH-HCC with a 30% or 50% cutoff, there was no statistically significant difference in tumor size compared with CV-HCC (P = .07 and P = .13, respectively). Regarding the frequency of vascular invasion, there were no statistically significant differences between SH-HCC and CV-HCC, regardless of whether a 5%, 30%, or 50% cutoff was used. Although SH-HCC defined by a 5% cutoff showed a higher frequency of moderately differentiated (grade 2) tumors (P = .005), there were no statistically significant differences in tumor histologic grade when using a 30% or 50% cutoff compared with CV-HCC (P = .10 and P = .11, respectively). Morphologic features commonly associated with non-neoplastic livers with steatohepatitis, such as Mallory-Denk bodies and glycogenated nuclei of tumor cells, were more frequently identified in SH-HCC, irrespective of the cutoff used (5%, 30%, or 50%), than in CV-HCC. Similarly, cytoplasmic hyaline globules, which have been described in HCC,¹³ was more frequently seen in SH-HCC, regardless of the cutoff used (5%, 30%, or 50%), than in CV-HCC. Intratumoral inflammatory infiltrates, including lymphocytes and neutrophils, were more commonly seen in SH-HCC than in CV-HCC, regardless of the cutoff used (5%, 30%, or 50%). In addition, SH-HCC exhibited a statistically significantly higher prevalence of intratumoral granulomas when defined by a 30% (P = .04) or 50% (P = .02) cutoff, but this difference was not present when a 5% cutoff used compared with CV-HCC. Immunohistochemical staining for L-FABP showed preserved expression in all 10 SH-HCC cases in the 50% cutoff group TABLE 3.

TABLE 3.

Pathologic Characteristics of CV-HCC vs SH-HCC at Different Steatohepatitis Cutoffs

Characteristics	CV-HCC (n = 223)	SH-HCC (5% cutoff) (n = 171)	P value	SH-HCC (30% cutoff) (n = 107)	P value	SH-HCC (50% cutoff) (n = 83)	P value
Size, mean (SD), cm	4.22 (2.84)	3.31 (2.28)	.007***	3.47 (2.22)	.065	3.55 (2.38)	.13
Vascular invasion, n/N (%)	18/70 (26)	29/92 (32)	.42	17/58 (29)	.65	13/45 (29)	.71
Histologic grade, n/N (%)	Grade 1: 33/222 (15) Grade 2: 156/222 (70) Grade 3: 24/222 (11) Grade 4: 9/222 (4)	Grade 1: 12/171 (7) Grade 2: 144/171 (84) Grade 3: 14/171 (9) Grade 4: 1/171 (1)	.005***	Grade 1: 10/107 (9) Grade 2: 88/107 (82) Grade 3: 8/107 (7) Grade 4: 1/107 (1)	.10	Grade 1: 8/83 (10) Grade 2: 68/83 (82) Grade 3: 7/83 (8) Grade 4: 0 (0)	.11
Glycogenated nuclei in tumor cells, n/N (%)	56/223 (25)	90/171 (53)	<.001***	63/107 (59)	<.001***	47/83 (57)	<.001***
Mallory-Denk bodies, n/N (%)	8/223(4)	66/171 (39)	<.001***	56/107 (52)	<.001***	47/83 (57)	<.001***
Presence of hyaline globules, n/N (%)	19/223 (9)	29/171 (17)	.013*	21/107 (20)	.006***	18/83 (22)	.003**
Intratumoral granulomas, n/N (%)	2/223 (0.9)	6/171 (3.5)	.082	5/107 (5)	.04*	5/83 (6)	.02*
Intratumoral lymphocytes, n/N (%)
Focal (<3 foci/20×)	72/223 (32)	59/171 (35)	<.001***	36/107 (34)	—	28/83 (34)	<.001***
Moderate (<6 foci/20×)	36/223 (15)	60/171 (35)		42/107 (39)		34/83 (41)
Marked (>6 foci/20×)	17/223 (8)	16/171 (9)		14/107 (13)		12/83 (15)
Intratumoral neutrophils, n/N (%)
Focal (<3 foci/20×)	15/223 (7)	21/171 (12)	<.05*	15/107 (14)	—	13/83 (16)	.005***
Moderate (<6 foci/20×)	0	2/171 (1)		2/107 (2)		2/83 (2)
Marked (>6 foci/20×)	5/223 (2)	1/171 (1)		0		0
Background liver fibrosis, n/N (%)	Stage 0: 5/223 (2) Stage 1: 8/223 (4) Stage 2: 18/223 (8) Stage 3: 17/223 (8) Stage 4: 175/223 (78)	Stage 0: 9/171 (5) Stage 1: 8/171 (5) Stage 2: 15/171 (9) Stage 3: 9/171 (5) Stage 4: 130/171 (76)	.45	Stage 0: 4/107 (4) Stage 1: 4/107 (4) Stage 2: 10/107 (9) Stage 3: 5/107 (5) Stage 4: 84/107 (79)	.79	Stage 0: 3/83 (4) Stage 1: 4/83 (5) Stage 2: 8/83 (10) Stage 3: 4/83 (5) Stage 4: 64/83 (77)	.82
Background liver steatosis, n/N (%)	<5%: 183/223 (82) 5%-33%: 31/223 (14) 33%-66%: 9/223 (4) >66%: 0 (0)	<5%: 113/171 (66) 5%-33%: 37/171 (22) 33%-66%: 14/171 (8) >66%: 5/171 (3)	.001**	<5%: 64/107 (60) 5%-33%: 30/107 (28) 33%-66%: 7/107 (7) >66%: 4/107 (4)	<.001***	<5%: 48/81 (59) 5%-33%: 26/81 (32) 33%-67%: 5/81 (6) >67%: 2 /81(2)	<.001***

CV-HCC, conventional hepatocellular carcinoma; SH-HCC, steatohepatitic hepatocellular carcinoma.

*P < .05; **P < .01; ***P < .001.

Regarding background liver steatosis TABLE 3, there was a statistically significant increase in the degree of steatosis in all 3 SH-HCC groups, regardless of whether a 5%, 30%, or 50% cutoff was used to define SH-HCC compared with CV-HCC. The stages of background liver fibrosis, however, did not differ significantly between CV-HCC and SH-HCC, regardless of the cutoff used.

We further compared the above histopathologic characteristics among the 3 SH-HCC subgroups using the cutoff thresholds of 5% to 29%, 30% to 49%, or 50% or higher TABLE 4. There were no statistically significant differences among the 3 subgroups in tumor size (P = .55), vascular invasion (P = .4041), histologic grade (P = .17), presence of glycogenated nuclei of tumor cells (P = .73) and cytoplasmic hyaline globules (P = .28), degree of intratumoral neutrophils (P = .47), or background liver fibrosis stages (P = .91). The presence of intratumoral lymphocytes (P = .02), intratumoral granulomas (P = .03), Mallory-Denk bodies (P < .001), and background liver steatosis (P = .04), however, was more frequent in the SH-HCC subgroups with 30% to 49% or 50% or higher cutoffs.

TABLE 4.

Pathologic Characteristics of SH-HCC at Different Percentages of Intratumoral Steatohepatitis

Characteristics	SH-HCC (5%-29%) (n = 64)	SH-HCC (30%-49%) (n = 24)	SH-HCC (≥50%) (n = 83)	P value
Size, mean (SD), cm	3.07 (2.29)	3.17 (1.50)	3.55 (2.38)	.55
Vascular invasion, n/N (%)	12/34 (35)	4/13 (31)	13/45 (29)	.40
Histologic grade, No. (%)	Grade 1: 2 (3) Grade 2: 56 (88) Grade 3: 6 (9) Grade 4: 0 (0)	Grade 1: 2 (8) Grade 2: 20 (83) Grade 3: 1 (4) Grade 4: 1 (4)	Grade 1: 8 (10) Grade 2: 68 (82) Grade 3: 7 (8) Grade 4: 0 (0)	.17
Glycogenated nuclei in tumor cells, n/N (%)	27/64 (42)	16/24 (67)	47/83 (57)	.73
Mallory-Denk bodies, n/N (%)	10/64 (16)	9/24 (38)	47/83 (57)	<.001***
Presence of hyaline globules, n/N (%)	8/64 (13)	3/24 (13)	18/83 (22)	.28
Intratumoral granulomas, n/N (%)	1/64 (2)	0 (0)	5/83 (6)	.03*
Intratumoral lymphocytes, n/N (%)
Focal (<3 foci/20×)	23/64 (36)	8/24 (33)	28/83 (34)	.02*
Moderate (<6 foci/20×)	18/64 (28)	8/24 (33)	34/83 (41)
Marked (>6 foci/20×)	2/64 (3)	2/24 (8)	12/83 (15)
Intratumoral neutrophils, n/N (%)
Focal (<3 foci/20×)	6/64 (9)	2/24 (8)	13/83 (16)	.47
Moderate (<6 foci/20×)	0	0	2/83 (2)
Marked (>6 foci/20×)	1/64 (2)	0	0
Background liver fibrosis, n/N (%)	Stage 0: 5/64 (8) Stage 1: 4/64 (6) Stage 2: 5/64 (8) Stage 3: 4/64 (6) Stage 4: 46/64 (72)	Stage 0: 1/24 (4) Stage 1: 0 (0) Stage 2: 2/24 (8) Stage 3: 1/24 (4) Stage 4: 20/24 (83)	Stage 0: 3/83 (4) Stage 1: 4/83 (5) Stage 2: 8/83 (10) Stage 3: 4/83 (5) Stage 4: 64/83 (77)	.91
Background liver Steatosis, n/N (%)	<5%: 49/64 (77) 5%-33%: 7/64 (11) 33%-66%: 7/64 (11) >66%: 1/64 (2)	<5%: 16/24 (67) 5%-33%: 4/24 (17) 33%-66%: 2/24 (8) >66%: 2/24 (8)	<5%: 48/81 (59) 5%-33%: 26/81 (32) 33%-67%: 5/81 (6) >67%: 2 /81(2)	.04*

SH-HCC, steatohepatitic hepatocellular carcinoma.

*P <.05; **P <.01; ***P <.001.

Survival Analysis

Survival analysis was conducted for patients with CV-HCC and SH-HCC using different cutoff thresholds. The median follow-up times were 75.53 months for the CV-HCC group and 90.20 months for the SH-HCC group. Overall survival was similar between patients with CV-HCC and SH-HCC using 5% (P = .11), 30% (P = .19), or 50% (P = .31) cutoff thresholds FIGURE 2A. Furthermore, no statistically significant difference in overall survival was observed among patients within the 5% to 29%, 30% to 49%, and 50% or higher SH-HCC subgroups (P = .8080) FIGURE 2B.

FIGURE 2.

Kaplan-Meier curves for overall survival analysis. A, Overall survival for CV-HCC vs SH-HCC at 3 cutoffs. B, Overall survival for SH-HCC at 3 cutoffs based on the percentages of steatohepatitis in the tumor. C, Overall survival in biopsy cases for CV-HCC vs SH-HCC at 3 cutoffs. D, Overall survival in resection cases for CV-HCC vs SH-HCC at 3 cutoffs. E, Overall survival in biopsy cases for SH-HCC at 3 cutoffs based on the percentages of steatohepatitis in the tumor. F, Overall survival in resection cases for SH-HCC at 3 cutoffs based on the percentages of steatohepatitis in the tumor. CV-HCC, conventional hepatocellular carcinoma; SH-HCC, steatohepatitic hepatocellular carcinoma.

Given that a significant number of cases in the study were from biopsy specimens and there were different outcomes between patients who underwent resections and patients who underwent biopsy procedures, we conducted a separate analysis to compare patient survival between biopsy and resection cases. As illustrated in FIGURES 2C and 2D, no statistically significant differences in overall survival were observed between patients with CV-HCC and patients with SH-HCC using 5%, 30%, or 50% cutoff thresholds in both biopsy and resection cases. Furthermore, the analysis revealed no statistically significant differences in overall survival among patients with CV-HCC and patients with HCC containing 5% to 29%, 30% to 49%, and 50% or more steatohepatitic components in both biopsy and resection cases FIGURES 2E and 2F.

Discussion

Hepatocellular carcinoma is the most common form of primary liver cancer, with cirrhosis due to chronic liver disease from viral infection (HBV/HCV) and MASLD remaining the strongest risk factors for HCC development. Salomao et al³ were the first to identify a subset of HCV-related HCCs showing features of steatohepatitis—namely, SH-HCC, which was associated with concurrent MASLD/metabolic dysfunction-associated steatohepatitis (MASH—formerly nonalcoholic steatohepatitis). Subsequent study by the same group demonstrated that most SH-HCC cases were associated with underlying steatohepatitis, with or without viral hepatitis.⁴ Although there is a higher frequency of MASLD in patients with SH-HCC, studies from different demographic regions have shown that SH-HCCs can also occur in association with ALD, pure viral hepatitis without underlying MASLD/ALD, or even in a noncirrhotic/normal liver background without any underlying liver disease.^6,9,10 In the current study, our data showed that although MASLD and ALD were important underlying liver diseases in SH-HCC cases, HCV infection remained the most common risk factor for SH-HCC and CV-HCC. Furthermore, among the SH-HCC cases using different steatohepatitis cutoffs, the cutoff thresholds did not show statistically significant etiologic differences. We also noted that a subset of background livers showed noncirrhotic/normal histology without underlying liver disease. Reflecting the higher frequency of MASLD among patients with SH-HCC, our results demonstrated a correlation between increased steatosis in the background liver and a higher percentage of steatohepatitic components in the tumor. Given the prevalent background liver steatosis in SH-HCC and the challenges in assessing background liver steatohepatitis, particularly in advanced-stage liver disease or livers affected by concurrent MASLD with other liver diseases, Shibahara et al⁷ have proposed that tumor histology may serve as a more accurate indicator of the metabolic background than the background liver tissue itself. It has been proposed that the lipotoxic effects of high levels of fatty acid oxidation on hepatocytes contribute to MASLD-associated HCC,¹⁴ but it is suggested that in nonsteatotic chronic liver disease–associated SH-HCC, an inflammation-induced metaplastic phenomenon rather than lipotoxicity is the primary driver.¹⁵ The exact mechanism underlying SH-HCC carcinogenesis, however, remains unclear. Therefore, the heterogeneity of associated etiologies may explain the differences in molecular profiles in SH-HCC.^10,15-17 More research is needed to enhance our understanding of the pathobiology and interactions among these co-existing diseases and to develop therapeutic strategies to improve clinical outcomes.¹⁸

When SH-HCC was first described by Salomao et al³ in 2010, the authors used a cutoff of 5% of steatohepatitis in the tumor to define SH-HCC. They reported that 59.1% (13/22) of HCC cases from 2003 to 2005 and 22.5% (9/40) of HCC cases from 2007 to 2008 were classified as SH-HCC, with an average of 35.5% (22/62) being SH-HCCs. A subsequent report from the same group, using a 50% cutoff for steatohepatitis, found that 13.4% of HCCs were classified as SH-HCC during the period from 2007 to mid-2010.⁴ Following these studies, SH-HCC cases were further investigated and subsequently recognized by the WHO classification of digestive system tumors as 1 of the HCC subtypes.² Regarding the definition of SH-HCC, although all studies agree that the presence of histopathologic features of steatohepatitis is required for diagnosis, different studies have variably used different cutoff thresholds, such as greater than 5%,^{3,6,7,10,15,19-21} greater than 30%,^10,22 and greater than 50%^4-6,9,10 of the steatohepatitis component in the tumor. Furthermore, the number of steatohepatitic components used to define SH-HCC varies from 3 or more^4,5,10, 4 or more^5,7,9,19, to all 5.^7,22. Of note, some studies did not specify the number of steatohepatitic components required.^3,6,15,19,20 The reported prevalence of SH-HCC among HCCs ranges from 8.7% to 59.1%.^3,5-7 This variability in prevalence rates is likely due to the use of different cutoff thresholds for the percentage of steatohepatitis and the number of steatohepatitic components used to define SH-HCC. As with other HCC subtypes, it is important to establish universally accepted criteria for the proper classification of SH-HCC to better define its clinicopathologic correlates.

To further assess clinicopathologic significance, we divided SH-HCCs into different groups and subgroups, using various cutoffs for comparison with conventional HCCs. It is well known that different HCC subtypes have varying clinicopathologic significance and prognoses. In our current study, we excluded all other HCC subtypes and focused on comparing SH-HCC with CV-HCC, which was different from some previous studies that included other HCC subtypes compared with CV-HCC.^4,6,7 When a 5%, 30%, or 50% cutoff was used, the prevalence rates of SH-HCC among HCC cases were 33.9%, 21.2%, and 16.4%, respectively, which are consistent with previously reported ranges. More importantly, there were no statistically significant differences in overall survival between patients with CV-HCC and patients with SH-HCC using the 3 cutoffs nor among patients in the SH-HCC subgroups with different amounts of steatohepatitis. Our findings, along with those from previous studies,^4-7,19 indicate that HCC with steatohepatitic features does not represent a clinically significant prognostic factor, and the different cutoff thresholds do not provide meaningful clinical prognostic value.

Features of typical steatohepatitis, including macrovesicular steatosis, ballooning of malignant hepatocytes, Mallory-Denk bodies, intratumoral fibrosis, and inflammation, are the criteria for diagnosing of SH-HCC. Initial studies by Salomao et al showed that SH-HCC cases exhibited steatosis and ballooning of tumor cells, but not all cases had intratumoral inflammation, Mallory-Denk bodies, or fibrosis. Although many cases showed all 5 steatohepatitic features, considering the varying degrees of combination of each feature, the authors defined SH-HCC based on the presence of at least 3 of these features, without specifying a minimal amount of each feature for diagnosis.^3,4 Because the presence of at least 5% steatosis is consistently included as 1 of the criteria, the other components, although commonly present, can appear in varying combinations. Subsequent studies have used different cutoffs for these 5 steatohepatitic features, such as 3 or more, 4 or more, or all 5 features, to diagnose SH-HCC. In the present study, we followed the initial definition by Salomao et al to define SH-HCC when a tumor with steatosis (≥5%) and ballooning of tumor cells, along with varying combinations of intratumoral inflammation, fibrosis, and Mallory-Denk bodies. This definition aligns with the diagnosis of steatohepatitis in non-neoplastic livers, where the presence of steatosis, inflammation, and ballooning of hepatocytes, with or without Mallory-Denk bodies or fibrosis, is often sufficient for diagnosis.²³ Notably, as the percentage of steatohepatitic features in the tumor increased, all 5 steatohepatitic features and intratumoral granulomas were more likely to be present, and the background livers were more likely to show a higher degree of steatosis. One of the most promising recent treatments for HCC is immunotherapy with immune checkpoint inhibitors,^24,25 but studies have shown that immune checkpoint inhibitors are less effective against MASH-related HCC than in HCCs with viral etiologies.²⁶ The histopathologic features of HCCs related to MASH vs viral hepatitis were not specified in these studies, leaving uncertainty as to whether HCCs with SH-HCC morphologies are responsive or tolerant to immune checkpoint inhibitors in different background liver diseases. Further research is needed to better understand the complex interactions between HCCs with steatohepatitic features and their response to immune therapies.

Hepatocellular carcinoma can be classified into 2 major molecular categories: the proliferation class and the nonproliferation class.^16,17 The proliferation class, often associated with an HBV-related etiology, is characterized by aggressive tumors with poorly differentiated histology, frequent vascular invasion, enrichment in TP53 variations, and amplification of FGF19 and CCND1. In contrast, the nonproliferation class includes a heterogeneous group of HCCs related to MASH, ASH, and HCV infection. These tumors are generally less aggressive, with well-differentiated to moderately differentiated histology and less frequent vascular invasion. Steatohepatitic HCC has been categorized under the nonproliferation class and often exhibits activation of the interleukin 6/JAK-STAT pathway, indicated by positive immunohistochemical expression of C-reactive protein (CRP). However, CRP is widely expressed in conventional HCCs,^6,27 making it an unsuitable specific marker for SH-HCC. Studies have shown that loss of L-FABP expression can occur in HCC, and a subset is associated with HNF1A inactivation, similar to findings in HNF1A-inactivated hepatocellular adenomas.²⁸ In our study, L-FABP immunohistochemical staining in 10 SH-HCC cases did not reveal loss of expression, indicating that HNF1A inactivation is unlikely to be a common event in SH-HCC development. These findings suggest that the current available molecular observations are not yet sufficient to definitively characterize SH-HCC as a distinct subtype.¹

In summary, this study provides a comprehensive analysis of clinicopathologic features of SH-HCC categorized by varying degrees of steatohepatitis in the tumor. Our findings suggest that there are no clinically significant differences in clinicopathologic characteristics and overall survival among patients with SH-HCC at different steatohepatitis cutoffs compared with patients with CV-HCC. The degree of intratumoral steatohepatitis does not appear to be a clinically significant prognostic factor. Even the presence of 5% steatohepatitis in the tumor justifies a diagnosis of SH-HCC, which is better recognized as 1 of the histopathologic patterns of HCC rather than a distinct subtype.

Conflict of interest disclosure: None of the authors has any relationships with or financial interests in any commercial companies pertaining to this article.

Funding

This project was supported in part by the Yale Liver Center award NIH P30 DK034989 Morphology core.

This article is available for CME credit.Go to academic.oup.com/ajcp/pages/journal_cme to see the latest articles. The complete catalog of journal CME courses can be found at store.ascp.org.