Factors Associated with Outcomes and Response to Therapy in Patients with Infiltrative Hepatocellular Carcinoma

Abstract

Background & Aims

Infiltrative hepatocellular carcinoma (iHCC) is characterized by its indistinct borders and lack of a typical pattern of contrast enhancement. There are few published data on iHCC. We assessed outcomes, effects of treatment, and prognostic factors in a large cohort of patients with iHCC.

Methods

We analyzed data from 155 patients (median age 60 years; 79% male; median level of α-fetoprotein [AFP] 347 ng/mL; median model for end-stage liver disease score [MELD] 13) with iHCC, based on contrast-enhanced computed tomography or magnetic resonance imaging, from 2002 to 2010 at the University of California, San Francisco Medical Center. All imaging study results were independently reviewed by 2 investigators.

Results

Most of the patients had tumors of Barcelona Clinic Liver Cancer stage C (70%) or D (22%). The median maximum tumor diameter was 11.3 cm; 41% of lesions were hypovascular, 82% had macrovascular invasion, and 52% had extra-hepatic metastases. Median survival was 4.0 months, and rates of survival at 6 and 12 months were 30% and 10%, respectively. On multivariate analysis, predictors of 6-month mortality were Child-Pugh class B or C cirrhosis; lack of tumor-directed therapy with chemoembolization (TACE), radiofrequency ablation, or sorafenib; AFP level >1000 ng/mL; female sex; MELD score; and maximum tumor diameter. The percentages of patients surviving 6 and 12 months were 17% and 2% for those that received no therapy (n=109), 73% and 36% for those that received sorafenib (n=11), and 45% and 17% for those that that received TACE (n=18) (all P values <.01).

Conclusions

iHCC is a radiographically distinct and advanced form of HCC with a poor prognosis. Therapy with TACE or sorafenib appears to prolong survival and requires further investigation.

INTRODUCTION

Worldwide, hepatocellular carcinoma (HCC) is the fifth most common cancer diagnosed in men and the seventh most common in women¹. Overall, it is the third leading cause of cancer-related death². Each year, there are over 600,000 new cases of HCC, with 20,000 cases diagnosed in the United States. Major risk factors for HCC include viral hepatitis, specifically hepatitis B and C, as well as alcoholic and non-alcoholic fatty liver disease^{3, 4}. In part due to the rising prevalence of hepatitis C, the incidence of HCC in the United States over the past 2 decades has nearly tripled¹.

HCC can present with different morphological subtypes including ”nodular,” ”multifocal,” and ”infiltrative/diffuse”^5-9. This gross classification of HCC is primarily based on radiologic characteristics. Nodular HCC is most common and typically presents with early arterial hyperenhancement, circumscribed borders, and a washout pattern on portal venous phase delayed images¹⁰. In contrast, infiltrative hepatocellular carcinoma (iHCC) is not recognized as a focal tumor but instead permeates throughout the liver. Radiographically, iHCC has indistinct borders, a lack of typical enhancement pattern seen in nodular HCC, and often macrovascular invasion. Its permeative appearance often blends into the background of a cirrhotic liver, causing difficulty with visualization on both ultrasound and cross-sectional imaging^{6, 11, 12}. While iHCC makes up approximately 8-18% of all HCC cases^{5, 13, 14}, there is a paucity of published data on its typical epidemiologic, radiologic, and pathologic characteristics^11-14. In addition, the small sample size of these studies precludes adequate evaluations of patient outcomes, prognostic factors that might impact survival, and the potential value of treatment of iHCC.

The primary aim of the present study was to assess the outcomes, effect of treatment, and factors predicting prognosis in a large cohort of patients with iHCC. We also sought to further evaluate the demographic, clinical and radiologic characteristics of iHCC.

PATIENTS AND METHODS

Study Design and Patient Population

This was a retrospective cohort study of patients aged 18 years and older with a new radiographic diagnosis of iHCC between January 2002 and December 2010. Because ultrasound is inadequate at detecting iHCC¹², only contrast-enhanced computed tomography (CT) and magnetic resonance imaging (MRI) scans were included. We searched the radiology database at our institution to identify cases of iHCC. Our search terms included a combination of “infiltrative”, “diffuse”, or “ill-defined,” and “HCC,” “hepatoma,” or “hepatocellular carcinoma.” Using these terms, approximately 1100 radiology reports from 600 patients were identified. Initial review of the radiographic reports led to identification of 177 “probable” cases of iHCC. Two of the investigators performed an independent review of cross-sectional imaging studies in all patients, and clear agreement on the diagnosis of iHCC was established in 161 patients. Six patients who were lost to follow-up were then excluded. The final cohort consisted of 155 patients.

The variables collected included demographic data (age, sex, gender), laboratory data (alpha-fetoprotein (AFP), Model for End-Stage Liver Disease (MELD) score), liver-related factors (etiology of liver disease, Child-Pugh class and Barcelona Clinic Liver Cancer [BCLC] stage), pathologic and radiographic data. Among patients who had biopsy of the tumor, histologic grade was based on the modified Edmondson criteria¹⁵ (grade 1: well-differentiated; grade 2: moderately differentiated; and grade 3: poorly differentiated). All laboratory and pathologic data were collected within 30 days of the diagnosis of iHCC. Radiographic characteristics included tumor size, portal or hepatic venous tumor thrombosis, extent of tumor thrombus, biliary ductal dilatation, contrast enhancement on arterial and portal venous phases of imaging, and presence and location of intra- and extrahepatic metastases. iHCC treatment data including the type, date, and duration of therapies were also obtained. For patients who received transarterial chemoembolization (TACE), pre-embolization angiographic images were assessed for the presence of tumor enhancement and for the degree of ethiodized oil uptake immediately following chemoembolization. Date of death was determined by searching the Social Security Death Index.

Statistical Analysis

Traditional descriptive statistics were used to calculate prevalence, means, and medians. Kaplan-Meier survival probabilities were determined as well as survival stratified by treatment type (sorafenib alone versus TACE alone versus no treatment). Time zero was defined as date of iHCC diagnosis. Cox proportional hazards models were used to assess for predictors of deathwithin 6 months and also predictors of death within 12 months. Univariate analyses included the following: age, gender, race, etiology of liver disease, type of iHCC treatment, AFP level, MELD score, Child-Pugh class, BCLC stage, tumor size, presence or absence of arterial phase hyperenhancement, MRI T2 signal characteristics (hyper-, iso-, or hypointense), presence of macrovascular invasion, extra- or intrahepatic metastases, and biliary ductal dilatation. The multivariate model included variables with the strongest plausible biological association with iHCC outcomes, or with a p-value <0.05 on univariate analysis. A forward selection of covariates was used to create the final multivariate model. Retention in the final model was defined using a p-value of <0.05. P-values < 0.05 were considered statistically significant in all analyses. Univariate and multivariate analyses were reported as hazard ratios (HR) and 95% confidence intervals (CI). Statistical analyses were performed using Stata (Version 12, Statacorp, College Station, TX).

RESULTS

Patient Characteristics

Of the 155 patients, 122 (78.7%) were men, the median age was 60 years, and 91 (58.7%) were Caucasian. Hepatitis C virus (HCV) was the most common etiology (56.8%), followed by hepatitis B (HBV) (16.1%), alcoholic liver disease (12.9%), and non-alcoholic fatty liver disease (NAFLD)(9.7%). The median AFP level was 347 ng/ml; 17.4% of patients had an AFP less than 20 ng/ml, and 42.3% had an AFP greater than 1000 ng/ml. At the time of diagnosis, the median MELD score was 13, the median Child-Pugh-Turcotte (CPT) was 10 with 31.3% of patients classified as Child’s class A (CPT 5-6), 46.0% Child’s class B (CPT 7-9), and 22.7% Child’s class C (CPT ≥ 10)¹⁶. The BCLC stage¹⁷ was A or B in only 6.0%, stage D in 22.5%, and BCLC stage C in the majority of patients (71.5%) (Table 1).

Table 1.

Baseline Characteristics of patients with iHCC (n=155)

Median Age (yrs)	60 (range 22-89)
Sex (M/F)	122/33
Median MELD Score	13 (range 6-38)
Race/Ethnicity		Number (%)
	Caucasian	91 (58.7%)
	Hispanic	28 (18.1%)
	Asian	27 (17.4%)
	African American	9 (5.8%)
Etiology of Liver Disease
	HCV	88 (56.8%)
	HBV	25 (16.1%)
	Alcoholic Liver Disease	20 (12.9%)
	Non-Alcoholic Fatty Liver Disease	15 (9.7%)
	HBV+HCV	2 (1.3%)
	Other	5 (3.2%)
AFP level (n=144)
	0-20	25 (17.4%)
	20-400	51 (35.4%)
	400-1000	7 (4.9%)
	1000+	61 (42.3%)
Child Pugh Class (n=150)
	A	47 (31.3%)
	B	69 (46.0%)
	C	34 (22.7%)
Barcelona Clinic Liver Cancer (BCLC) stage (n=151)
	A/B	9 (6.0%)
	C	108 (71.5%)
	D	34 (22.5%)

Regarding tumor presentation, 95 patients (61.3%) were diagnosed with iHCC at the time of referral to our institution. The other 60 patients had been diagnosed with non-infiltrative HCC prior to referral, of which 32 (53.3%) had a single lesion and 28 (46.7%) had multifocal HCC.

With regards to tumor-directed treatment for these 60 patients, 23 (38.3%) had previously received TACE, 9 (15%) had been treated with RFA, 5 (8.3%) had received both TACE and RFA, and 4 (6.7%) had undergone a partial hepatectomy or wedge resection.

Radiographic Characteristics

In all iHCC cases, margins of the mass were poorly demarcated. The median iHCC diameter was 11.3cm, and 95 patients (61.3%) had lesions larger than 10cm. The majority of cases (81.9%) had radiographic evidence of macrovascular invasion at the time of iHCC diagnosis. All patients with macrovascular invasion had some degree of portal vein tumor thrombosis (PVTT); 48.4% had main portal vein involvement, 28.4% involved either the right and/or left lobar portal vein, and 5.2% had segmental portal vein invasion. In addition to PVTT, 11.6% of patients had tumor invasion of the hepatic vein. Overall, 5.2% of patients had tumor extending into the main hepatic vein, 3.9% into the inferior vena cava, and 2.6% of iHCC cases extended into the right atrium. While nodular HCC usually exhibits early arterial hyperenhancement, only 91 (58.7%) iHCC lesions displayed such enhancement characteristics. Of these 91 lesions, all but two demonstrated wash-out on the portal-venous imaging phase. Regarding metastatic disease at the time of iHCC diagnosis, 49.7% of patients had intrahepatic satellite lesions and 48.4% had extrahepatic spread. The most common metastatic sites were the lymph nodes (34.2%) and lungs (20.6%). Intrahepatic biliary ductal dilatation was relatively uncommon, found in only 25.8% of cases. For the 30 patients who had a liver MRI, 22 tumors (73.3%) were relatively homogeneous and mildly hyperintense on T2-weighted images, and the remainder were isointense to surrounding liver parenchyma (Table 2).

Table 2.

Radiographic and Pathologic Characteristics of iHCC Patients at Time of Diagnosis (n=155)

Median Maximum Tumor Diameter (cm)	11.3 (range 4.2-25.0)
Vascular Invasion		Number (%)
	Any	127 (81.9%)
	Portal Vein	127 (81.9%)
	Main	75 (48.4%)
	Lobar	44 (28.4%)
	Segmental	8 (5.2%)
	Hepatic Vein	18 (11.6%)
	Hepatic Vein Only	8 (5.2%)
	Inferior Vena Cava	6 (3.9%)
	Right Atrium	4 (2.6%)
Arterial Hyperenhancement
	Present	91 (58.7%)
	Absent	64 (41.3%)
Extrahepatic Metastases
	Any	75 (48.4%)
	Lymph Nodes	53 (34.2%)
	Lung	32 (20.6%)
	Peritoneal	11 (7.1%)
	Bone	9 (5.8%)
	Adrenal	6 (3.9%)
Intrahepatic Metastases
	Present	77 (49.7%)
	Absent	78 (50.3%)
Biliary Duct Dilation
	Any	40 (25.8%)
	Unilobar	23 (14.8%)
	Segmental	9 (5.8%)
	Bilobar	8 (5.2%)
MRI T2 Signal Appearance (n=30)
	Hyperintense	22 (73.3%)
	Isointense	8 (26.7%)
Cytologic Differentiation (n=27)
	Poor	14 (51.9%)
	Moderate	11 (40.7%)
	Well	2 (7.4%)

Pathologic Characteristics

Of the 155 patients, 55 had a biopsy of the tumor within 30 days of iHCC diagnosis. Twenty-seven of the 55 were able to be characterized using the modified Edmonson classification¹⁵; 14 (51.9%) were poorly differentiated, 11 (40.7%) were moderately differentiated, and 2 (7.4%) were well differentiated (Table 2). Four of the 55 patients (7.3%) who were biopsied had features of both HCC and cholangiocarcinoma. Two of these four patients had poorly differentiated tumors with AFP levels of 147 and 2009, one had moderate differentiation with a normal AFP of 6.3, and one had unknown cellular differentiation and an AFP of 1408. None of these four patients had biliary ductal dilatation.

Treatment Modalities

Only 46 (29.7%) individuals received tumor-directed therapy after iHCC diagnosis. TACE was performed with an emulsion of 25 mg doxorubicin hydrochloride, 10 mg mitomycin C, 50 mg cisplatin, and ethiodized oil followed by embolization to stasis with a slurry of gelatin sponge. The administered doses of chemotherapy agents were adjusted in patients with liver or renal dysfunction, leukopenia, and thrombocytopenia. There were 18 patients (11.6%) who received TACE alone, 11 patients (7.1%) who received sorafenib alone for at least 30 days after the drug became available in 2008, and 4 patients (2.6%) who received RFA. Three of these four patients received RFA alone and one patient received RFA and a single TACE treatment. An additional 4 patients (2.6%) received sorafenib and one or more TACE treatments (Table 3). Overall, a total of 23 patients received 36 TACE treatments. On review of angiographic images, 17 TACE patients (73.9%) had mildly hypervascular tumors with diffuse and patchy retention of ethiodized oil, while 6 patients (26.1%) had strongly hypervascular lesions with dense and avid lipiodol retention following TACE (Figure 1).

Table 3.

Tumor Based Treatments After Diagnosis of iHCC (n=155)

Type of Therapy Received	Number (%)
None	109 (70.3%)
TACE alone	18 (11.6%)
RFA alone	3 (1.9%)
RFA + TACE	1 (0.6%)
Sorafenib alone	11 (7.1%)
Sorafenib + TACE	4 (2.6%)
Non-Sorafenib Chemotherapy	7 (4.5%)
Liver Transplant	1 (0.6%)
Resection	1 (0.6%)

Fig. 1. Typical appearance of an infiltrative hepatocellular carcinoma (iHCC).

(A). Coronal cross sectional image demonstrating large, ill-defined iHCC in the right hepatic lobe (B). On angiogram, this lesion demonstrates minimal hypervascularity (C). Following transarterial chemoembolization (TACE), the angiogram demonstrates only patchy and diffuse ethiodized oil uptake

One patient with a 6 cm iHCC invading the main portal vein but without extrahepatic metastases was treated with surgical resection. This patient had a preoperative MELD score of 7, CPT score of 5 and pathology showing a poorly differentiated mixed cholangio-hepatocellular carcinoma. He died from decompensated liver disease with spontaneous bacterial peritonitis and hepatorenal syndrome two months after resection. A second patient with a 4.7cm iHCC without evidence of vascular invasion, satellite lesions, or extrahepatic metastases on imaging underwent an orthotopic liver transplantation one month after iHCC diagnosis. This patient had a preoperative MELD score of 18, CPT score of 9, and explant revealing moderately differentiated, irregularly encapsulated HCC. A large tumor thrombus was noted in the main portal vein. He died 4.7 years after transplant at the age of 60 from an unknown cause.

Survival Analysis

As of September 2011, 148 of the 155 patients had died (95.5%). The median overall survival for iHCC was 4.0 months (range 0.1-55.5 months). Three month, 6 month, and 1 year survivals were 63%, 30%, and 8%, respectively. Patients who were treated with TACE alone (n=18) or sorafenib alone (n=11) had a significantly better survival (log-rank p=0.0002) than for those who did not receive tumor-directed therapy (n=109) (Figures 2a and 2b). There was no significant difference in survival between the sorafenib group and those treated with TACE (log-rank p=0.267). There were too few patients in the RFA group to include in these survival analyses. Median survival for the patients treated with RFA (with or without TACE), sorafenib alone, and TACE alone was 9.2 months, 7.5 months, and 6.0 months, respectively whereas survival for patients who received no therapy was only 3.0 months. Patients who were treated with both TACE and sorafenib were not included in these subgroup comparisons (n=4), and their median survival was 21.6 months (range 13.1-30.3 months). Stratified by BCLC stage, median survival was 14.7 months for stage A/B (n=6), 6.0 months for stage C (n=108), and 1.9 months for stage D (n=34).

Fig. 2.

Kaplan-Meier Survival Curve for infiltrative hepatocellular carcinoma (iHCC) patients treated with (A). Transarterial chemoembolization (TACE) versus no therapy and (B). Sorafenib versus no therapy

Predictors of Death

On univariate analysis, the following covariates were predictive of death within 6 months from time of iHCC diagnosis: Child-Pugh class C or B (versus A), not receiving sorafenib or TACE/RFA, AFP>1000 ng/ml, female gender, MELD (per point), tumor size (per cm), BCLC stage D (versus A/B), arterial hypovascularity, and biliary duct dilation. On multivariate analysis, the strongest predictors of death within 6 months were Child-Pugh class C versus A(HR 6.12, 95% CI 2.74-13.76) and not receiving tumor-directed therapy with sorafenib, TACE, or RFA (HR 2.79, 95% CI 1.60-4.88). Additional covariates predictive of death on multivariate analysis included Child’s class B (versus A), AFP>1000 ng/ml, female gender, MELD (per point), and tumor size (per cm) (Table 4). Due to colinearity, Child’s class and BCLC stage were not included in the same multivariate model; Child’s class was retained in the final model over BCLC stage because the majority of the patients had advanced BCLC stage C or D. We also performed univariate analysis of factors predictive of death within 12 months, and these factors were identical to those predictive of death within 6 months (data not shown).

Table 4.

Univariate and Multivariate Predictors of Death Within 6 Months

Predictor	Univariate HR (95% CI), p-value	Multivariate HR (95% CI), p-value
Child’s C vs A	7.72 (4.40-13.47), <0.0001	6.12 (2.74-13.76), <0.0001
Child’s B vs A	1.99 (1.20-3.34), 0.0078	2.61 (1.36-4.80), 0.004
No Treatment vs TACE, RFA, or Sorafenib	3.55 (2.11-5.99), 0.0001	2.79 (1.60-4.88), 0.0003
AFP >1000 ng/ml	1.76 (1.19-2.61), 0.0049	2.47 (1.56-3.84), <0.0001
Female Gender	1.74 (1.13-2.70), 0.011	1.73 (1.03-2.90), 0.039
MELD (per point)	1.12 (1.08-1.15), <0.0001	1.07 (1.03-1.12), 0.002
Tumor Size (per cm)	1.04 (1.00-1.08), 0.04	1.05 (1.01-1.10), 0.015
BCLC Stage D vs A/B	9.60 (3.28-26.84), <0.0001	*
Arterial Hypovascularity	1.93 (1.31-2.82), 0.0008	1.33 (0.86-2.10), 0.21
Biliary Ductal Dilatation	1.90 (1.26-2.88), 0.0022	0.90 (0.53-1.52), 0.70

^*Due to colinearity, Child’s class and BCLC stage were not included in the same multivariate model; Child’s class was retained in the final model over BCLC stage because the majority of the patients had advanced BCLC stage C or D

DISCUSSION

In this study involving a large cohort of patients with iHCC, we characterized the clinical and radiologic features, and evaluated patient outcomes, treatment effects, and prognostic factors associated with survival. Two of the investigators performed an independent review of cross-sectional imaging studies in all patients to confirm the initial diagnosis of iHCC. Radiographically, the majority of iHCC were larger than 10 cm, and over 80% had portal venous invasion at the time of diagnosis. Only 60% had arterial enhancement with portal venous washout typically seen in nodular HCC. Among the 30 patients with liver MRI, 73% of the tumors were relatively homogeneous and mildly or moderately hyperintense on T2-weighted images and 27% were isointense. Kanematsu et al.¹³ examined typical MRI findings in 22 patients with iHCC and found that at least 50% of the liver was involved with tumor in all cases, 100% had proximal portal venous tumor invasion and on T2-weighted images all tumors appeared either mildly or moderately hyperintense. iHCC is clearly an advanced form of HCC as over 90% of patients in our series were classified as either BCLC stage C or D, and close to 50% had an AFP level greater than 1000 ng/mL.

Only 30% of our patients received any form of tumor-directed therapy, and only one patient each underwent surgical resection or liver transplantation. In one surgical series by Ochiai et al.¹⁸, the 5-year survival after surgical resection for iHCC was only 16%, and there was a three-fold increase in the risk of death in patients with iHCC compared to those with other HCC subtypes. None of the patients with iHCC were enrolled into our down-staging protocol for liver transplantation¹⁹. Barakat et al.²⁰ evaluated the effects of down-staging in 32 patients with advanced HCC outside of either Milan or University of California San Francisco (UCSF) criteria (1 lesion ≤ 6.5cm in diameter or 2-3 lesions each ≤ 4.5cm, with a total diameter ≤ 8cm) but without vascular invasion or extra-hepatic disease. None of the 12 patients with iHCC could be successfully down-staged, whereas 18 of 20 patients with nodular or multi-focal HCC were able to be successfully down-staged to within United Network for Organ Sharing (UNOS) T2 criteria, and 13 of these 18 (72%) ultimately received a liver transplant. Lopez et al.²¹ compared the outcomes after TACE in 69 patients with nodular HCC versus 19 patients with iHCC. The mean overall survival was significantly shorter in the infiltrative group (3.4 vs. 13.9 months, p<0.001). There were no patients with iHCC who did not receive TACE, and thus the impact of TACE on survival could not be determined in that study.

While iHCC had a poor overall prognosis with a median survival of 4 months in the present series, several factors were associated with improved outcomes. Not surprisingly, patients with Child’s class A cirrhosis and those with low MELD scores had a longer survival, as did patients with smaller tumors and lower AFP levels. We also found that loco-regional therapy or systemic therapy with sorafenib conferred a significant survival benefit in multivariate analysis, after adjusting for other factors including MELD score and Child’s class that might have affected treatment eligibility. The median survival was only 3.0 months for patients who received no therapy, versus 6.0 months for those treated with TACE, 7.5 months for patients who received sorafenib, and 9.2 months for those treated with RFA. There were four patients in our cohort who received a combination of sorafenib and locoregional therapy. These patients had the longest median survival of 21.6 months. The small number of patients in this subgroup precludes drawing any conclusion about the impact of combination therapy.

The retrospective study design and the relatively small numbers in the different treatment groups make it difficult to ascertain the benefits of treatment. Nevertheless, the survival benefit observed in iHCC patients treated with sorafenib is similar to the results from the SHARP trial²², where the median survival was 3 months longer among patients with HCC treated with sorafenib when compared to those who received placebo. In a recent publication by Kneuertz et al²³, patients with iHCC (n=75) that received intra-arterial therapy (IAT) compared to no IAT had a marked increase in survival (median survival of 12 vs 3 months respectively, p=0.001). This survival benefit was more pronounced than that observed in the current study, which may be due to an overall healthier study population undergoing IAT in the Kneuertz study, as well as lower AFP levels and concomitant sorafenib use. IAT patients (n=48) had a median MELD of 8, and 94% were CPT class A or B (versus MELD 13 with 57% of patients being CPT A/B in the non-treatment group (n=21)). Furthermore, the IAT group had a significantly lower AFP (244 vs 1563 ng/ml) and 25 of the 48 (52.1%) received peri-procedural sorafenib therapy in addition to IAT. As lower MELD scores, lower AFP levels, CPT A/B vs C, and receipt of sorafenib were predictive of improved survival in our study cohort, these factors likely contributed to the 9 month survival benefit observed among iHCC patients receiving IAT in the study by Kneuertz et al.

The current study has several limitations. Only 30% of patients received tumor-directed therapy, and thus the precise role and effects of therapy on iHCC cannot be ascertained. In particular, only a small number of patients received sorafenib in this study, mainly because sorafenib only became available in 2008. It is possible that patients with better prognosis were selected for loco-regional or sorafenib therapy while the sickest patients received only supportive care.

Nevertheless, the survival benefits of these treatments were still evident after controlling for other prognostic factors in the multivariate analysis. In spite of all the limitations, our data represent the largest patient cohort in the literature that allows for better characterization of the clinical and radiologic features, outcomes and prognostic factors associated with iHCC.

In summary, iHCC is a radiographically distinct and advanced form of HCC. The majority of patients with iHCC have extensive tumor burden as well as macrovascular invasion. Over 90% of these patients in this series had advanced BCLC stage C or D. The prognosis of patients who are diagnosed with iHCC is grim, with a median survival of only 4 months. Nonetheless, our results suggest that some patients may still derive survival benefit from liver-directed therapy and/or systemic therapy with sorafenib.

Abbreviations

AFP

Alpha-fetoprotein

BCLC

Barcelona Clinic Liver Cancer

CPT

Child-Pugh-Turcotte

CT

Computed Tomography

CI

Confidence intervals

HR

Hazard ratio

HBV

Hepatitis B virus

HCV

Hepatitis C virus

HCC

Hepatocellular carcinoma

iHCC

Infiltrative hepatocellular carcinoma

IAT

Intra-arterial therapy

MRI

Magnetic Resonance Imaging

MELD

Model for End-Stage Liver Disease

NAFLD

Non-alcoholic fatty liver disease

PVTT

Portal vein tumor thrombosis

RFA

Radiofrequency ablation

TACE

Transarterial chemoembolization

UNOS

United Network for Organ Sharing

UCSF

University of California San Francisco