Purpose:

Immune checkpoint inhibitor therapy is emerging as the promising option for patients with advanced hepatocellular carcinoma. We aimed to investigate the heterogeneity of different tumor nodules of the same patient with multifocal hepatocellular carcinomas in response to immunotherapy and its molecular mechanisms.

Experimental Design:

We attained 45 surgical tumor samples including 33 small and 12 large nodules from 12 patients with multifocal hepatocellular carcinoma and evaluated genomic and immune heterogeneity among tumors through whole-genome sequencing and RNA sequencing. IHC was performed to validate the expression of immune markers. The responses to anti–programmed cell death protein-1 (PD-1) therapy in patients with multifocal hepatocellular carcinoma were evaluated.

Results:

The small and large tumors within the same patient presented with similar genomic characteristics, indicating their same genomic origin. We further found the small tumors had higher immune cell infiltration including more CD8+ T cells, M1 macrophages, and monocytes as compared with large tumors. Besides, the expression of interferon signature predictive of response to anti–PD-1 therapy was significantly upregulated in the small tumors. Moreover, the immune pathways were more vigorous along with less active proliferation pathways in the small tumors. In keeping with this, we found that small nodules were more sensitive to anti–PD-1 therapy than large nodules in patients with multifocal hepatocellular carcinoma.

Conclusions:

The small tumors in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration and upregulation of immune pathways as compared with the large tumors, which can partially explain the different responses of small and large tumors in the same case to anti–PD-1 therapy.

Translational Relevance

Immunotherapy is emerging as a promising strategy in hepatocellular carcinoma, but its low response rate and mixed responses among different nodules remains to be a major challenge. In this study, we deciphered genomic and immune characteristics among tumors with different sizes in patients with multifocal hepatocellular carcinoma through whole-genome sequencing and RNA sequencing. We found the small and large nodules within the same patient presented with similar genomic characteristics, indicating that the small and large nodules were intrahepatic metastasis. Further analysis revealed that the small tumors had higher immune cell infiltration and upregulation of immune pathways as compared with the large tumor of the same multifocal hepatocellular carcinoma case, which may partially explain the different responses of small and large tumors to anti–programmed cell death protein-1 (anti–PD-1) treatment.

Hepatocellular carcinoma is the fourth most lethal cancer globally (1). Over 50% of patients with hepatocellular carcinoma are at advanced stage at the time of diagnosis (2). Currently, the first-line recommended treatment for patients with advanced hepatocellular carcinoma is targeted therapy, offering a median progression-free survival of only 3.6–7.3 months and an adverse rate over 80% (3, 4). Immunotherapy uses various strategies to enhance antitumor immunity and represents a promising option for hepatocellular carcinoma. Targeting immune checkpoint anti–programmed cell death protein-1 (PD-1) inhibitors (Checkmate 040 and Keynote 240) for the treatment of patients with advanced hepatocellular carcinoma demonstrated an overall survival of 28.6 months as a first-line treatment, and 12.9–15 months as a second-line treatment (5–7). However, the overall response rate for the anti–PD-1 treatment in patients with advanced hepatocellular carcinoma is less than 20% (6, 7). Among those unresponsive patients to anti–PD-1 treatment, a certain proportion have one or more nodules with controllable response, but still show progressive disease due to the coexisting progression of other nodules. Several studies in various kinds of tumors, including hepatocellular carcinoma, have reported various responses of different nodules to anti–PD-1 treatment in the same patient (8–10), which could be partially attributed to diversity in genomic characteristics, transcriptomic signatures, and tumor-immune microenvironment. Recent studies have discussed factors affecting response to immunotherapy, such as microsatellite instability, tumor mutation burdens, and intratumoral infiltration of CD8+ T cells (11–13), but these factors have seldom been compared at nodule level within the same patient, leaving the mechanisms of intertumoral heterogeneity to immunotherapy poorly understood.

In this study, we performed multi-omics analysis between small tumors (ST) and large tumors (LT) in patients with multifocal hepatocellular carcinoma through whole-genome sequencing (WGS) and RNA sequencing (RNA-seq) to gain insight into the intertumoral heterogeneous responses to immunotherapy.

Patients and samples

Twelve multifocal hepatocellular carcinoma patients with STs and LTs, who underwent resection before any adjuvant therapy, were included. The LT was defined as the one whose longest diameter was at least 1.5 times longer than all other tumors in the same multifocal hepatocellular carcinoma case, and the rest of the tumors were defined as STs. The clinicopathologic information of these 12 patients with hepatocellular carcinoma is listed in Supplementary Table S1. Clinical follow-ups were terminated at the end of September 2019, with the median follow-up period of 223 days. Tumor tissues (2–7 tumors per patient, 45 tumors in total) and adjacent nontumor tissues were snap-frozen in liquid nitrogen immediately within 15 minutes after resection. The diagnosis of multifocal hepatocellular carcinoma in all cases was confirmed by two experienced pathologists through histology examination. WGS was performed in tumor and adjacent nontumor tissues from 6 patients who had two tumor nodules in the liver, and whole-exome sequencing (WES) was applied to the other 6 patients as reported previously (14). RNA-seq was performed in all tumor and adjacent nontumor tissues from 12 patients.

An additional 8 patients with advanced multifocal hepatocellular carcinoma who received anti–PD-1 therapy were enrolled for treatment evaluation. The detailed information of patient enrollment is shown in the flow chart (Supplementary Fig. S1). In brief, 41 patients met the following inclusion criteria from June 2018 to February 2020 at the First Affiliated Hospital, Sun Yat-sen University (Guangzhou, China). The inclusion criteria were: (i) diagnosed as hepatocellular carcinoma according to the guideline of European Association for the Study of the Liver (15), (ii) received at least two cycles of anti–PD-1 therapy, and (iii) Child–pugh class A or B liver function. Thirty-three patients were excluded according to the exclusion criteria: (i) combined with focal treatment or other immunotherapy (n = 4), (ii) no CT/MRI data (n = 12), (iii) no measurable target lesion according to RECIST version 1.1 (RECIST 1.1; n = 2), (iv) only one lesion or no lesion in the liver (n = 14), and (v) the diameter of the largest tumor was less than 1.5 times the diameter of other tumors in the same hepatocellular carcinoma case (n = 1). In total, 7 patients with anti–PD-1 monotherapy (nivolumab and pembrolizumab) and one patient with combination therapy of anti–PD-1 (nivolumab) and lenvatinib were enrolled. RECIST 1.1 was used to evaluate the treatment response of all the tumors (16). After treatment, tumor was classified as four categorizations: complete response, partial response, stable disease, or progressive disease based on the radiographic evaluation of the target lesion.

This study was approved by the Institutional Ethics Review Board of the First Affiliated Hospital, Sun Yat-sen University (Guangzhou, China) and conducted according to the Declaration of Helsinki. Written informed consent was obtained from each patient.

Library preparation and WGS

Total DNA was extracted from the snap-frozen liver tissues using the QIAGEN DNeasy Blood & Tissue Kit (Qiagen). The qualified genomic DNA of tumor and adjacent nontumor tissues from patients with hepatocellular carcinoma were fragmented by an ultrasonicator Covaris E-210 (Covaris). DNA fragments were concentrated in 300-bp peaks with one library for each sample by adjusting shearing parameters. These fragments were purified, end blunted, “A” tailed, and adaptor ligated. A total of 10 to 12 cycles of PCR were performed after size selection in the gel. The concentration of the libraries was quantified by a bioanalyser (Agilent Technologies) and real-time PCR method using ABI StepOne plus real-time PCR system (Life technologies). Paired-end 150-bp read-length sequencing was performed in the HiSeq X TEN platform according to manufacturer's instructions (Illumina).

RNA-seq

Beads (Invitrogen) with oligo (dT) were used to isolate poly (A) mRNA after total RNA was collected. Fragmentation buffer was added for interrupting mRNA to short fragments. Taking these short fragments as templates, random hexamer-primer was used to synthesize the first-strand cDNA. The second-strand cDNA was synthesized using buffer, dNTPs, RNase H, and DNA polymerase I. Short fragments were purified with QIAQuick PCR Extraction Kit (Qiagen) and resolved with EB buffer for end reparation and adding poly (A). After that, the short fragments were connected with sequencing adaptors. For PCR amplification, we selected suitable fragments as templates, based on agarose gel electrophoresis. Finally, the library was sequenced using HiSeq X TEN platform and 150-bp paired-end reads were generated.

WGS data analysis

The adapter sequence in the raw data was removed, and low-quality reads which had more than 5 Ns and low-quality bases were discarded. Then high-quality reads were gapped aligned to the NCBI human reference genome (hg19) using Burrows-Wheeler Aligner (BWA; ref. 17) by default parameters. We performed local realignment of the original BAM alignment using the GATK (18) and followed by Picard to mark duplicates reads.

Somatic substitutions were detected by MuTect (19) based on BWA alignment. High confident somatic single-nucleotide variants (SNV) were called if the following criteria were met: (i) both the tumor and adjacent nontumor samples should be covered sufficiently (not less than 10×) at the genomic position, (ii) the variants should be supported by at least 5% of the total reads in the tumor while less than 1% in the adjacent nontumor tissue, (iii) the variants should be supported by at least five reads in the tumor, and (iv) distance between adjacent somatic SNV distance should be over 10 bp.

High-confidence somatic insertions and deletions (indels) were called using the following steps: (i) candidate somatic indels were predicted with GATK Somatic Indel Detector with default parameters and (ii) high confident somatic indels were defined after filtering germline events. All high-confidence somatic mutations were filtered out by the dbSNP (version 138) site which is commonly polymorphic without known medical impact. The remaining mutations were annotated with ANNOVAR (20) and subjected to subsequent analyses. Potentially functional mutations of immune-related genes from ImmPort data repository (21) and InnateDB analysis platform (22) were estimated using a threshold of SIFT < 0.05 and Polyphen-2 > 0.85 (23, 24).

Copy-number alterations (CNA) were analyzed by Control-FREEC (25). The copy ratio of tumor to adjacent normal tissue larger than 1.25 was defined as copy-number gain, and copy ratio less than 0.75 was defined as copy-number loss. Tumor cellularity was estimated using ABSOLUTE (26) algorithm based on copy-number results. Manta (27) was used to predict somatic structure variations (SV) breakpoints. High confident SVs were those with paired-end read ≥ 3 and split-read ≥ 1. The interaction of somatic SVs among different chromosomes was plotted with Circos (http://circos.ca/). Hepatitis B virus (HBV) integrations were detected by SeekSV (28), which simultaneously used read depth signal, discordant paired-end read signal, split-read signal, and the fragment with two ends unmapped. Mutational cancer cell fractions (CCF) were estimated through PyClone (29). The CCF of each SNV was computed on the basis of its variant allele fraction and copy numbers at the SNV locus. Tumor purity was used to adjust the estimated CCF.

RNA-seq data analysis

Qualified reads were obtained after removing raw reads with adapters or of low quality and then aligned to the human genome (hg19) by HISAT (30) with default parameters. RSeQC (31) was used to measured gene expression abundance as reads per kilobase per million mapped reads (RPKM). Unsupervised clustering of all samples in each case was performed by R package heatmap. Cellular composition analysis of immune infiltrates was performed by CIBERSORT in the absolute mode based on RNA-seq gene expression data (32). To calculate the activity of cancer hallmark-related pathways, the RPKM-based gene expression was subjected to single-sample gene set enrichment analysis (ssGSEA; ref. 33). In ssGSEA, gene expression value for a given sample was rank-normalized. An enrichment score representing the pathway activity was produced using the empirical cumulative distribution functions of the genes in the signature and the remaining genes. Several published hepatocellular carcinoma molecular classifications (34, 35) were analyzed using nearest template prediction method (36) based on the gene expression data of each tumor.

PCR and Sanger sequencing

To validate HBV integration found by genomic sequencing, we further performed PCR and Sanger sequencing. PCR primers were designed for sequences around the HBV integration sites (Supplementary Table S2). DNA was amplified with the AmpliTaq Gold 360 Master Mix (Applied Biosystems) and the amplicons were sequenced using ABI 3730 DNA analyzer (Life Technologies).

IHC

Hepatocellular carcinoma tissues were blocked with 10% normal goat serum for 10 minutes after deparaffinized, and were incubated with antibodies against CD8 (M7103, 1:200; DAKO), CD20 (ab9475, 1:200; Abcam), CD68 (ab955, 1:500; Abcam), Foxp3 (catalog no. 98377, 1:100; Cell Signaling Technology), CD14 (ab181470, 1:200; Abcam), PD-1 (catalog no. 86163, 1:200; Cell Signaling Technology), and inducible T-cell costimulator (ICOS; ab105227, 1:100; Abcam) overnight in a humidified chamber. The tissues were subsequently probed with secondary antibody (Vectastain ABC kit). The sections were examined with a ZEISS Axio Scan.Z1 Slide Scanner (AxioScan.Z1, ZEISS).

The expression of the above markers was assessed independently by two pathologists. PD-1 was scored as the mean percentage of positive cells in four separate tumor regions on a scale of 0–3: 0 (<1%), 1 (1%–5%), 2 (5%–10%), and 3 (>10%). ICOS was evaluated with an immunoscore according to the mean percentage of positive cells in four separate tumor regions on a scale of 0–3: 0 (<5%), 1 (5%–25%), 2 (25%–50%), and 3 (>50%). Immunoscore ≥2 was defined as high immunoscore, and immunoscore <2 was defined as low immunoscore for PD-1 and ICOS. Besides, the expression levels of CD8, CD20, CD68, Foxp3, and CD14 in different tumors were defined by the median immune cell density according to methods reported by others (37). The cut-off thresholds of CD8, CD20, CD68, Foxp3, and CD14 were 77 cells/mm2, 40 cells/mm2, 84 cells/mm2, 15 cells/mm2, and 60 cells/mm2, respectively. On the basis of the established threshold, a binary score (0: low; 1: high) was given for each immune marker in each tumor region. An immunoscore of each marker for the tumor was derived from the summation of four binary scores (range, 0–4). Immunoscore ≥3 was defined as high immunoscore, and immunoscore <3 was defined as low immunoscore.

Statistical analysis

Paired t test was used to compare the gene expression levels, enrichment scores, and immune cell density between the STs and LTs. For the 6 patients with more than one ST, the average expression level of all STs was used for paired t test analysis. A statistics model (38) was used to analyze TERT expression between two groups. Fisher exact test was used to evaluate the molecular subtypes difference. The Pearson correlation coefficients between number of chromosomal rearrangements and immune/inflammatory related gene sets (The Molecular Signatures Database, MSigDB) or immune cell infiltration were calculated in R software. All statistical analyses were carried out using R version 3.3.2 (http://www.r-project.org). Error bar indicated SEM. For all statistical analyses, two-tailed P value < 0.05 was considered statistically significant.

Data availability

The raw sequencing data reported in this article have been deposited in the Genome Sequence Archive (39) in National Genomics Data Center (40), Beijing Institute of Genomics (China National Center for Bioinformation), Chinese Academy of Sciences, under accession number HRA000176 that are publicly accessible at http://bigd.big.ac.cn/gsa-human.

The small and large nodules within the same patient with hepatocellular carcinoma presented with similar genomic characteristics

In the observation that different nodules had various responses to anti–PD-1 treatment in the same patient (8–10), we investigated the underlying molecular mechanisms by evaluating the genomic and immune heterogeneity of tumors with different sizes in 12 patients with multifocal hepatocellular carcinoma. Six of them were patients with hepatocellular carcinoma with two tumor nodules, including one LT and one ST measured by CT scans (Supplementary Fig. S2). WGS was applied to the tumor nodules and adjacent nontumor tissues in these 6 patients (Supplementary Table S3). The purity of hepatocellular carcinoma tumor cells reached over 60.9% in each nodule (Supplementary Fig. S3).

We explored the HBV-host interaction of different nodules in the patients with WGS data (Supplementary Fig. S4). Most HBV integration sites were located at the promoter region of the TERT and the exonic region of MLL4, which are reported to be the most prevalent genes integrated in hepatocellular carcinoma (Fig. 1A; Supplementary Table S4; ref. 41). mRNA expression levels of TERT were significantly upregulated in 11 of 12 tumor tissues, as compared with adjacent liver samples (Supplementary Fig. S5). Of note, the small and large nodules of each patient presented with identical integration sites while the adjacent liver samples did not, which was further confirmed by Sanger sequencing (Fig. 1A). We also investigated SVs of different nodules, and found some common interchromosomal rearrangements were shared in the ST and LT of each patient, with increased chromosomal rearrangements in the STs except for one case (HCC406; Fig. 1B; Supplementary Figs. S6 and S7; Supplementary Table S5). Besides, CNAs were similar in the ST and LT of the same patient (Fig. 1B; Supplementary Fig. S8; Supplementary Table S6). The number of SNVs (Supplementary Fig. S9) and mutation profiles (Supplementary Fig. S10; Supplementary Table S7) were also alike in the ST and LT of the same patient. CCF analysis showed both nodules shared most clonal mutations (Supplementary Fig. S11). Simplified phylogenetic trees were then constructed on the basis of the somatic mutations, with reported driver mutations, CNAs, and HBV integration sites mapping to the shared and private parts (Fig. 1C). We further compared the mutation profiles of 12 patients with hepatocellular carcinoma, including the other 6 patients with WES data (14), and found the mutation numbers in coding regions were similar between the STs and LT (Fig. 1D). Taken together, these results indicated that STs and LT in the same case might be intrahepatic metastatic tumors with the same origin.

Figure 1.

The small and large nodules within the same patient presented with similar genomic characteristics. A, Validation of common HBV integration sites in MLL4 exons or TERT promoter by Sanger sequencing. B, Circle plots show the genomic landscape of interchromosomal translocations and copy-number alterations in STs and LTs of case HCC372 and case HCC374. Chromosome ideograms are shown around the outer ring with copy-number plots on the inner ring. Common rearrangements between STs and LTs are highlighted in green while private rearrangements are shown as black arcs. Common CNAs between STs and LTs are labeled with purple lines. C, Simplified phylogenetic trees constructed using somatic mutations. The trunk and independent branches are represented in brown and yellow, respectively. Driver alterations of hepatocellular carcinoma are mapped to the phylogenetic tree. Purple: driver somatic mutations; red: copy-number gain; blue: copy-number loss; red box: HBV integration sites. D, The STs and LTs of each patient with hepatocellular carcinoma shared some common mutations in the coding regions, indicating that they were intrahepatic metastases. N, adjacent normal liver tissue; LT, large tumor; ST, small tumor.

Figure 1.

The small and large nodules within the same patient presented with similar genomic characteristics. A, Validation of common HBV integration sites in MLL4 exons or TERT promoter by Sanger sequencing. B, Circle plots show the genomic landscape of interchromosomal translocations and copy-number alterations in STs and LTs of case HCC372 and case HCC374. Chromosome ideograms are shown around the outer ring with copy-number plots on the inner ring. Common rearrangements between STs and LTs are highlighted in green while private rearrangements are shown as black arcs. Common CNAs between STs and LTs are labeled with purple lines. C, Simplified phylogenetic trees constructed using somatic mutations. The trunk and independent branches are represented in brown and yellow, respectively. Driver alterations of hepatocellular carcinoma are mapped to the phylogenetic tree. Purple: driver somatic mutations; red: copy-number gain; blue: copy-number loss; red box: HBV integration sites. D, The STs and LTs of each patient with hepatocellular carcinoma shared some common mutations in the coding regions, indicating that they were intrahepatic metastases. N, adjacent normal liver tissue; LT, large tumor; ST, small tumor.

Close modal

The small nodules in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration

Tumor microenvironment has been demonstrated to be associated with immunotherapy response (13). To explore the immune microenvironment of the STs and LTs, RNA-seq was applied on 45 intrahepatic metastatic tumors including 12 large nodules and 33 small nodules from 12 patients with multifocal hepatocellular carcinoma. Unsupervised clustering analysis showed that the STs and LT in the same case had similar gene expression profiles (Supplementary Fig. S12), which was consistent with their genomic similarity.

Immune cell infiltration determined by CIBERSORT analysis was significantly higher in the STs, including more CD8+ T cells, M1 macrophages, and monocytes as compared with the LTs (P < 0.05; Fig. 2A), while B cells, M2 macrophages, and regulatory T cells (Tregs) did not show significant difference (Fig. 2A). The protein expression levels of CD8, CD20, CD68, Foxp3, and CD14 were validated by IHC staining, which were markers of T cells, B cells, macrophages, Tregs, and monocytes, respectively (Fig. 2B). We further investigated the exhaustion and activation status of T cells in the STs and LTs. PD-1 expression was used to represent the exhaustion status of T cells, while ICOS expression was used to represent the activation status of T cells. Both RNA expression (Fig. 2C) and protein expression data (Fig. 2D) showed more activated T cells in the STs as compared with LTs, while exhausted T cells showed no significant difference (Fig. 2C and D). We also compared immune functional genes between the STs and LTs. Immune functional genes refer to genes that play a crucial role during the immune response, including immune stimulator genes (facilitating immune response) and immune inhibitor genes (suppressing immune reaction; refs. 41–43). Immune stimulators were upregulated in the STs (P < 0.01; Fig. 3A; Supplementary Figs. S13 and S14), while there was no significant difference of the immune inhibitors between two groups (P = 0.13; Fig. 3A; Supplementary Figs. S13 and S15).These finding indicated that antitumor immune response might be more active in the STs. We further found that the STs exhibited higher expression of IFN signature (P < 0.05; Fig. 3A), which was reported to predict the response to anti–PD-1 therapy in multiple solid tumor types (44, 45). These data collectively suggested that the STs in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration than the LTs.

Figure 2.

The small nodules in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration. A, Immune cell populations in the STs and LTs determined by CIBERSORT analysis from RNA-seq data. B, The representative pictures (top) and quantification (bottom) of protein expression levels of CD8, CD20, CD68, Foxp3, and CD14 in the STs and LTs by IHC analysis, which represented T cells, B cells, macrophages, Tregs, and monocytes, respectively. C, The RNA expression levels of PD-1, CD8A, and ICOS in the STs and LTs. D, The representative pictures (left) and quantification (right) of protein expression of PD-1 and ICOS in the STs and LTs by IHC. PD-1 was used to represent the exhaustion status of T cells, while ICOS was used to represent the activation status of T cells. LT, large tumor; ST, small tumor.

Figure 2.

The small nodules in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration. A, Immune cell populations in the STs and LTs determined by CIBERSORT analysis from RNA-seq data. B, The representative pictures (top) and quantification (bottom) of protein expression levels of CD8, CD20, CD68, Foxp3, and CD14 in the STs and LTs by IHC analysis, which represented T cells, B cells, macrophages, Tregs, and monocytes, respectively. C, The RNA expression levels of PD-1, CD8A, and ICOS in the STs and LTs. D, The representative pictures (left) and quantification (right) of protein expression of PD-1 and ICOS in the STs and LTs by IHC. PD-1 was used to represent the exhaustion status of T cells, while ICOS was used to represent the activation status of T cells. LT, large tumor; ST, small tumor.

Close modal
Figure 3.

The small nodules were more active in immune pathways and less active in proliferation signaling pathways. A, ssGSEA and molecular subtypes analysis revealed that immune pathways were more active in STs while proliferation pathways were more active in LTs. B, A summary of genomic and immunologic characteristics of the ST and LT of multifocal hepatocellular carcinoma. In the genomic level, the ST and LT showed identical HBC integration sites and similar structural variations, mutation profiles, and copy-number alterations, which indicated that they were intrahepatic metastases. On the other hand, immune pathways were more vigorous along with less active proliferation pathways in the ST. LT, large tumor; ST, small tumor.

Figure 3.

The small nodules were more active in immune pathways and less active in proliferation signaling pathways. A, ssGSEA and molecular subtypes analysis revealed that immune pathways were more active in STs while proliferation pathways were more active in LTs. B, A summary of genomic and immunologic characteristics of the ST and LT of multifocal hepatocellular carcinoma. In the genomic level, the ST and LT showed identical HBC integration sites and similar structural variations, mutation profiles, and copy-number alterations, which indicated that they were intrahepatic metastases. On the other hand, immune pathways were more vigorous along with less active proliferation pathways in the ST. LT, large tumor; ST, small tumor.

Close modal

The small nodules were more active in immune pathways and less active in proliferation signaling pathways

To gain further insight into the biological pathway differences between small and large nodules in patients with multifocal hepatocellular carcinoma, we performed ssGSEA in 45 tumor nodules from 12 patients with multifocal hepatocellular carcinoma. We found immune pathways including IFNα response (P < 0.001) and TNFα signaling pathway (P < 0.01) were more active in STs than LTs (Fig. 3A; Supplementary Table S8), which could activate immune cells and contribute to inflammation and immune enhancement (46, 47). On the other hand, proliferation and oncogenic pathways including VEGF signaling pathway (P < 0.001) and PI3K/Akt/mTOR signaling pathway (P < 0.001) were more active in LTs (Fig. 3A; Supplementary Table S8). VEGF signaling pathway is involved in both vascular genesis and angiogenesis (48) and PI3K/Akt/mTOR signaling pathway is crucial to cell growth (49).

We also investigated the pathway differences in the STs and LTs with reported hepatocellular carcinoma molecular classifications (34, 35). Sia immune class analysis showed that more STs were classified as “Active Immune Response Subtype” than the LTs (13/33 vs. 1/12), although no significant difference was found. Chiang classification analysis showed that IFN class was more enriched in the small foci (11/33 vs. 1/12), while proliferation class was slightly enriched in the large tumors (20/33 vs. 8/12), without significant differences (Fig. 3A). The heterogeneity of pathway enrichment in small and large nodules highlighted the importance of precise intervention for different nodules in patients with multifocal hepatocellular carcinoma. It deserves further investigation whether the combination of antiproliferation/oncogenic therapy such as tyrosine kinase inhibitors (TKI) and immunotherapy could achieve synergistic effect for patients with multifocal hepatocellular carcinoma (Fig. 3B; ref. 50).

To evaluate the impact of genomic alterations on the immune signaling, we analyzed the mutations of known immune-related genes (21, 22) in the STs and LTs. We found that 85.4% (76/89) of mutations of immune-related genes were shared between the ST and LT in the same patient, while some different mutations of immune-related genes were observed between the ST and LT (Supplementary Figure S16). We further estimated the potentially functional mutations from the different mutations of immune-related genes using a threshold of SIFT < 0.05 and Polyphen-2 > 0.85 (23, 24). The LTs contained three specific mutations including PGR, SCARF1, and RETN, while the STs had four particular mutations including TP53, DDX21, SMARCA2, and MUC4 (Supplementary Table S9). These mutations might have potential impact on their immune signaling, and are worth further investigation in the future. Moreover, we analyzed the correlation between the number of chromosomal rearrangements and surrogates of inflammation. The result showed that there was no significant relation between the number of rearrangements and inflammatory response or immune cell infiltration (P > 0.05; Supplementary Table S10).

Small nodules were more sensitive to anti–PD-1 therapy than large nodules in patients with multifocal hepatocellular carcinoma

To evaluate whether nodules with different sizes in patients with multifocal hepatocellular carcinoma showed different responses to anti–PD-1 therapy, 8 patients received anti–PD-1 therapy and were monitored closely (Supplementary Fig. S1). Among 7 patients who received anti–PD-1 monotherapy, 5 patients showed various responses to immunotherapy in different nodules (Fig. 4A; Supplementary Fig. S17). Detailed clinical data are shown in Supplementary Table S11. In particular, the small nodule showed better response than the large nodule in the same patient. According to the RECIST 1.1, the ST had a median of 46.3% shrinkage after anti–PD-1 treatment, while the corresponding LT in the same patient had a median of 14.9% increase in the tumor size (P < 0.05). As expected, the patient who received combination therapy of anti–PD-1 and lenvatinib showed a satisfactory effect with dramatic shrinkage of the large tumor and disappearance of the small nodule (Fig. 4B).

Figure 4.

Small nodules were more sensitive to anti–PD-1 therapy than large nodules in patients with multifocal hepatocellular carcinoma. A, The representative CT or MRI of two patients with hepatocellular carcinoma before and after anti–PD-1 monotherapy. The small nodules showed better response than the large nodule in each case. B, The CT images of the patient before and after 10 cycles of combination therapy (nivolumab and lenvatinib). The large nodule showed dramatic recession and the small nodule disappeared.

Figure 4.

Small nodules were more sensitive to anti–PD-1 therapy than large nodules in patients with multifocal hepatocellular carcinoma. A, The representative CT or MRI of two patients with hepatocellular carcinoma before and after anti–PD-1 monotherapy. The small nodules showed better response than the large nodule in each case. B, The CT images of the patient before and after 10 cycles of combination therapy (nivolumab and lenvatinib). The large nodule showed dramatic recession and the small nodule disappeared.

Close modal

In this study, genomic and immune differences were compared between the STs and LTs in patients with multifocal hepatocellular carcinoma by multi-omics analysis. We found that the STs had higher immune cell infiltration and were more active in immune pathways.

Mixed response of immunotherapy with shrinkage of one tumor nodule but progression of the other nodules has been widely reported in multifocal solid cancer (8–10). Better understanding of the underlying mechanisms is important for the selection of clinical treatment strategy. We observed a tendency that nodules with smaller size showed better response than the larger one in patients receiving anti–PD-1 monotherapy. However, the multi-omics comparison of nodules in the same patient according to tumor size has not been reported to the best of our knowledge.

In this study, we explored the comprehensive heterogeneity of STs and LTs so as to identify the potential factors related to immunotherapy response. It has been demonstrated that genomic characteristics including chromosomal rearrangements, CNA profile, and mutation pattern may potentially affect the antitumor immune response (51–53). We found that the STs in the patients with multifocal hepatocellular carcinoma with WGS data contained increased number of chromosomal rearrangements except one case (HCC406), but there was no significant correlation between the number of chromosomal rearrangements and immune response. In addition, although some different CNAs and mutations of immune-related genes were observed between the small and large foci, there were no certain rules that could be followed. These genomic alterations might have potential impact on the immune signaling, and are worth further investigation in the future.

We found the small and large tumors of each patient with multifocal hepatocellular carcinoma showed great genomic similarity. In particular, identical HBV integrated sites and some common SVs were found between the ST and LT in the same patient, indicating that they might be intrahepatic metastases. It has been widely demonstrated that HBV integration occurs in early timing of hepatocellular carcinoma and spatially separated tumor cells with the same HBV integrations are regarded as intrahepatic metastases (54). Besides, similar SVs pattern implied the possibility of a subclone with translocation of one lesion seeded metastases, as described for pancreatic cancer (55) and prostate cancer (56). Moreover, other 6 patients with multifocal hepatocellular carcinoma in our recent study showed that 33 of 34 tumor nodules belonged to intrahepatic metastatic foci (14). Taken together, our findings indicated that the majority of multiple nodules in hepatocellular carcinoma were intrahepatic metastases (IM), at least in our center. Some studies documented that multicentric origin was more frequent than IM in multifocal hepatocellular carcinoma (57, 58), while other studies indicated higher frequency of IM type (59, 60). One study with 160 patients showed that 59.4% of patients with multifocal hepatocellular carcinoma were with IM (59). Different clinicopathologic characteristics including etiology and liver cirrhosis may lead to various metastatic features.

Although small and the large nodules in the same patient had the same genomic origin, their immune microenvironments were quite different, which has been described to be related with immunotherapy response (13). Immune cells including M1 macrophages, monocytes, and CD8+ T cells were significantly enriched in the small nodules than those in the large nodules. M1 macrophages and monocytes have been reported to coordinate the local adaptive antitumor immune response (61, 62), and intratumoral intensity of CD8+ T has been proposed as a positive biomarker of anti–PD-1 response (63). The RNA and protein expression levels of ICOS were also upregulated in the STs. ICOS is a costimulatory molecule which is expressed on T cells after their activation. Patients with increased expression of ICOS had better outcome in response to immune checkpoint inhibitors therapy (64). Moreover, an IFN signature containing 28 genes related to antigen presentation, chemokine expression, cytotoxic activity, and adaptive immune response was upregulated in the STs, which was predictive of response to anti–PD-1 therapy across multiple solid tumors (44, 45). With higher immune cell infiltration and active immune pathway enrichment in the small nodules, we speculated that the STs may be more sensitive to immunotherapy. On the other side, the LTs were found to have upregulated proliferation pathways including VEGF signaling pathway. Our previous study has demonstrated that hepatocellular carcinoma with high expression of VEGFR2 might be sensitive to TKI therapy (65), thus we assumed that the LTs might be responsive to TKI therapy and combination therapy of anti–PD-1 therapy and TKIs might be a potential choice for patients with hepatocellular carcinoma with unresectable multifocal nodules. This was supported by recent clinical trials of this innovative combination treatment that have shown promising antitumor efficacy with objective response rate over 30% in the interim analysis (66, 67).

However, there were some limitations in our study. First, the sample size of patients with sequencing data in our study was rather small and they did not have anti–PD-1 treatment history, which may somehow weaken our conclusions. Second, the number of enrolled hepatocellular carcinoma patients with anti–PD-1 treatment was relatively small. Third, the STs and LT from the same patient presented some different chromosomal rearrangements, CNAs, and mutations, which might have potential impact on the immune response. Fourth, the underlying mechanism of different responses to immunotherapy in different nodules was not fully explored. In fact, tumor samples of patients with anti–PD-1 therapy were inherently difficult to collect because radical therapy was generally not available for this population. Therefore, further preclinical studies may help to reveal the potential cellular and molecular mechanism.

In conclusion, we found that the STs in patients with multifocal hepatocellular carcinoma had higher immune cell infiltration and upregulation of immune pathways as compared with the LT of the same case, which may partially contribute to the different responses to anti–PD-1 treatment.

No potential conflicts of interest were disclosed.

M. Huang: Data curation, formal analysis, investigation, writing-original draft. M. He: Formal analysis, investigation. Y. Guo: Data curation, investigation. H. Li: Resources. S. Shen: Resources. Y. Xie: Formal analysis. X. Li: Writing-review and editing. H. Xiao: Writing-review and editing. L. Fang: Formal analysis. D. Li: Resources. B. Peng: Resources. L. Liang: Resources. J. Yu: Conceptualization, writing-review and editing. M. Kuang: Funding acquisition, writing-review and editing. L. Xu: Conceptualization, supervision, funding acquisition, methodology, writing-review and editing. S. Peng: Conceptualization, supervision, funding acquisition, methodology, writing-review and editing.

This work was supported by the National Natural Science Foundation of China (81825013, 81771958), Guangzhou Science, Technology and Innovation Commission (Bureau of Science and Information Technology of Guangzhou Municipality; 201704020215), Bureau of Science and Information Technology of Guangzhou Municipality | Pearl River S and T Nova Program of Guangzhou (201906010086), Natural Science Foundation of Guangdong Province (Guangdong Natural Science Foundation; 2019B151502009), and Bureau of Science and Information Technology of Guangzhou Municipality | Guangzhou Research Collaborative Innovation Projects (201704020224).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
International Agency for Research on Cancer, World Health Organization
. 
Cancer today
.
Available from
: https://gco.iarc.fr/today/home.
2.
Park
JW
,
Chen
M
,
Colombo
M
,
Roberts
LR
,
Schwartz
M
,
Chen
PJ
, et al
Global patterns of hepatocellular carcinoma management from diagnosis to death: the BRIDGE study
.
Liver Int
2015
;
35
:
2155
66
.
3.
Kudo
M
,
Finn
RS
,
Qin
S
,
Han
KH
,
Ikeda
K
,
Piscaglia
F
, et al
Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial
.
Lancet
2018
;
391
:
1163
73
.
4.
Llovet
JM
,
Ricci
S
,
Mazzaferro
V
,
Hilgard
P
,
Gane
E
,
Blanc
JF
, et al
Sorafenib in advanced hepatocellular carcinoma
.
N Engl J Med
2008
;
359
:
378
90
.
5.
Crocenzi
TS
,
El-Khoueiry
AB
,
Yau
TC
,
Melero
I
,
Sangro
B
,
Kudo
M
, et al
Nivolumab (nivo) in sorafenib (sor)-naive and-experienced pts with advanced hepatocellular carcinoma (HCC): CheckMate 040 study
.
J Clin Oncol
35
:
15s
, 
2017
(
suppl; abstr 4013
).
6.
El-Khoueiry
AB
,
Sangro
B
,
Yau
T
,
Crocenzi
TS
,
Kudo
M
,
Hsu
C
, et al
Nivolumab in patients with advanced hepatocellular carcinoma (CheckMate 040): an open-label, non-comparative, phase 1/2 dose escalation and expansion trial
.
Lancet
2017
;
389
:
2492
502
.
7.
Zhu
AX
,
Finn
RS
,
Edeline
J
,
Cattan
S
,
Ogasawara
S
,
Palmer
D
, et al
Pembrolizumab in patients with advanced hepatocellular carcinoma previously treated with sorafenib (KEYNOTE-224): a non-randomised, open-label phase 2 trial
.
Lancet Oncol
2018
;
19
:
940
52
.
8.
Koelzer
VH
,
Rothschild
SI
,
Zihler
D
,
Wicki
A
,
Willi
B
,
Willi
N
, et al
Systemic inflammation in a melanoma patient treated with immune checkpoint inhibitors-an autopsy study
.
J Immunother Cancer
2016
;
4
:
13
.
9.
George
S
,
Miao
D
,
Demetri
GD
,
Adeegbe
D
,
Rodig
SJ
,
Shukla
S
, et al
Loss of PTEN is associated with resistance to anti-PD-1 checkpoint blockade therapy in metastatic uterine leiomyosarcoma
.
Immunity
2017
;
46
:
197
204
.
10.
Lu
L-C
,
Hsu
C
,
Shao
Y-Y
,
Chao
Y
,
Yen
C-J
,
Shih
IL
, et al
Differential organ-specific tumor response to immune checkpoint inhibitors in hepatocellular carcinoma
.
Liver Cancer
2019
;
8
:
480
90
.
11.
Le
DT
,
Durham
JN
,
Smith
KN
,
Wang
H
,
Bartlett
BR
,
Aulakh
LK
, et al
Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade
.
Science
2017
;
357
:
409
13
.
12.
Legrand
FA
,
Gandara
DR
,
Mariathasan
S
,
Powles
T
,
He
X
,
Zhang
W
. 
Association of high tissue TMB and atezolizumab efficacy across multiple tumor types
.
J Clin Oncol
36
:
15s
, 
2018
(
suppl; abstr 12000
).
13.
Tumeh
PC
,
Harview
CL
,
Yearley
JH
,
Shintaku
IP
,
Taylor
EJ
,
Robert
L
, et al
PD-1 blockade induces responses by inhibiting adaptive immune resistance
.
Nature
2014
;
515
:
568
71
.
14.
Xu
LX
,
He
MH
,
Dai
ZH
,
Yu
J
,
Wang
JG
,
Li
XC
, et al
Genomic and transcriptional heterogeneity of multifocal hepatocellular carcinoma
.
Ann Oncol
2019
;
30
:
990
7
.
15.
European Association for the Study of the Liver
. 
EASL clinical practice guidelines: management of hepatocellular carcinoma
.
J Hepatol
2018
;
69
:
182
236
.
16.
Schwartz
LH
,
Litiere
S
,
de Vries
E
,
Ford
R
,
Gwyther
S
,
Mandrekar
S
, et al
RECIST 1.1-update and clarification: from the RECIST committee
.
Eur J Cancer
2016
;
62
:
132
7
.
17.
Li
H
,
Durbin
R
. 
Fast and accurate short read alignment with Burrows-Wheeler transform
.
Bioinformatics
2009
;
25
:
1754
60
.
18.
McKenna
A
,
Hanna
M
,
Banks
E
,
Sivachenko
A
,
Cibulskis
K
,
Kernytsky
A
, et al
The genome analysis toolkit: a mapreduce framework for analyzing next-generation DNA sequencing data
.
Genome Res
2010
;
20
:
1297
303
.
19.
Cibulskis
K
,
Lawrence
MS
,
Carter
SL
,
Sivachenko
A
,
Jaffe
D
,
Sougnez
C
, et al
Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples
.
Nat Biotechnol
2013
;
31
:
213
9
.
20.
Wang
K
,
Li
M
,
Hakonarson
H
. 
ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data
.
Nucleic Acids Res
2010
;
38
:
e164
.
21.
Bhattacharya
S
,
Dunn
P
,
Thomas
CG
,
Smith
B
,
Schaefer
H
,
Chen
J
, et al
ImmPort, toward repurposing of open access immunological assay data for translational and clinical research
.
Sci Data
2018
;
5
:
180015
.
22.
Breuer
K
,
Foroushani
AK
,
Laird
MR
,
Chen
C
,
Sribnaia
A
,
Lo
R
, et al
InnateDB: systems biology of innate immunity and beyond–recent updates and continuing curation
.
Nucleic Acids Res
2013
;
41
:
D1228
33
.
23.
Ng
PC
,
Henikoff
S
. 
SIFT: predicting amino acid changes that affect protein function
.
Nucleic Acids Res
2003
;
31
:
3812
4
.
24.
Adzhubei
I
,
Jordan
DM
,
Sunyaev
SR
. 
Predicting functional effect of human missense mutations using PolyPhen-2
.
Current Protocols in Human Genetics
2013
;
76
:7.20.
1
41
.
25.
Boeva
V
,
Popova
T
,
Bleakley
K
,
Chiche
P
,
Cappo
J
,
Schleiermacher
G
, et al
Control-FREEC: a tool for assessing copy number and allelic content using next-generation sequencing data
.
Bioinformatics
2012
;
28
:
423
5
.
26.
Carter
SL
,
Cibulskis
K
,
Helman
E
,
McKenna
A
,
Shen
H
,
Zack
T
, et al
Absolute quantification of somatic DNA alterations in human cancer
.
Nat Biotechnol
2012
;
30
:
413
21
.
27.
Chen
X
,
Schulz-Trieglaff
O
,
Shaw
R
,
Barnes
B
,
Schlesinger
F
,
Kallberg
M
, et al
Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications
.
Bioinformatics
2016
;
32
:
1220
2
.
28.
Liang
Y
,
Qiu
K
,
Liao
B
,
Zhu
W
,
Huang
X
,
Li
L
, et al
Seeksv: an accurate tool for somatic structural variation and virus integration detection
.
Bioinformatics
2017
;
33
:
184
91
.
29.
Roth
A
,
Khattra
J
,
Yap
D
,
Wan
A
,
Laks
E
,
Biele
J
, et al
PyClone: statistical inference of clonal population structure in cancer
.
Nat Methods
2014
;
11
:
396
8
.
30.
Kim
D
,
Langmead
B
,
Salzberg
SL
. 
HISAT: a fast spliced aligner with low memory requirements
.
Nat Methods
2015
;
12
:
357
60
.
31.
Wang
L
,
Wang
S
,
Li
W
. 
RSeQC: quality control of RNA-seq experiments
.
Bioinformatics
2012
;
28
:
2184
5
.
32.
Newman
AM
,
Liu
CL
,
Green
MR
,
Gentles
AJ
,
Feng
W
,
Xu
Y
, et al
Robust enumeration of cell subsets from tissue expression profiles
.
Nat Methods
2015
;
12
:
453
7
.
33.
Barbie
DA
,
Tamayo
P
,
Boehm
JS
,
Kim
SY
,
Moody
SE
,
Dunn
IF
, et al
Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1
.
Nature
2009
;
462
:
108
12
.
34.
Chiang
DY
,
Villanueva
A
,
Hoshida
Y
,
Peix
J
,
Newell
P
,
Minguez
B
, et al
Focal gains of VEGFA and molecular classification of hepatocellular carcinoma
.
Cancer Res
2008
;
68
:
6779
88
.
35.
Sia
D
,
Jiao
Y
,
Martinez-Quetglas
I
,
Kuchuk
O
,
Villacorta-Martin
C
,
Castro de Moura
M
, et al
Identification of an immune-specific class of hepatocellular carcinoma, based on molecular features
.
Gastroenterology
2017
;
153
:
812
26
.
36.
Hoshida
Y
. 
Nearest template prediction: a single-sample-based flexible class prediction with confidence assessment
.
PLoS One
2010
;
5
:
e15543
.
37.
Gabrielson
A
,
Wu
Y
,
Wang
H
,
Jiang
J
,
Kallakury
B
,
Gatalica
Z
, et al
Intratumoral CD3 and CD8 T-cell densities associated with relapse-free survival in HCC
.
Cancer Immunol Res
2016
;
4
:
419
30
.
38.
Audic
S
,
Claverie
JM
. 
The significance of digital gene expression profiles
.
Genome Res
1997
;
7
:
986
95
.
39.
Wang
Y
,
Song
F
,
Zhu
J
,
Zhang
S
,
Yang
Y
,
Chen
T
, et al
GSA: Genome Sequence Archive
.
Genomics Proteomics Bioinformatics
2017
;
15
:
14
8
.
40.
National Genomics Data Center Members and Partners
. 
Database resources of the national genomics data center in 2020
.
Nucleic Acids Res
2020
;
48
:
D24
33
.
41.
Cancer Genome Atlas Research Network.
Comprehensive and integrative genomic characterization of hepatocellular carcinoma
.
Cell
2017
;
169
:
1327
41
.
42.
Zhang
Q
,
Lou
Y
,
Yang
J
,
Wang
J
,
Feng
J
,
Zhao
Y
, et al
Integrated multiomic analysis reveals comprehensive tumour heterogeneity and novel immunophenotypic classification in hepatocellular carcinomas
.
Gut
2019
;
68
:
2019
31
.
43.
Mascaux
C
,
Angelova
M
,
Vasaturo
A
,
Beane
J
,
Hijazi
K
,
Anthoine
G
, et al
Immune evasion before tumour invasion in early lung squamous carcinogenesis
.
Nature
2019
;
571
:
570
5
.
44.
Ayers
M
,
Lunceford
J
,
Nebozhyn
M
,
Murphy
E
,
Loboda
A
,
Kaufman
DR
, et al
IFN-gamma-related mRNA profile predicts clinical response to PD-1 blockade
.
J Clin Invest
2017
;
127
:
2930
40
.
45.
Ribas
A
,
Robert
C
,
Hodi
FS
,
Wolchok
JD
,
Joshua
AM
,
Hwu
W-J
, et al
Association of response to programmed death receptor 1 (PD-1) blockade with pembrolizumab (MK-3475) with an interferon-inflammatory immune gene signature
.
J Clin Oncol
33
:
15s
, 
2015
(
suppl; abstr 3001
).
46.
Josephs
SF
,
Ichim
TE
,
Prince
SM
,
Kesari
S
,
Marincola
FM
,
Escobedo
AR
, et al
Unleashing endogenous TNF-alpha as a cancer immunotherapeutic
.
J Transl Med
2018
;
16
:
242
.
47.
Crouse
J
,
Kalinke
U
,
Oxenius
A
. 
Regulation of antiviral T cell responses by type I interferons
.
Nat Rev Immunol
2015
;
15
:
231
42
.
48.
Stacker
SA
,
Achen
MG
. 
The VEGF signaling pathway in cancer: the road ahead
.
Chin J Cancer
2013
;
32
:
297
302
.
49.
Porta
C
,
Paglino
C
,
Mosca
A
. 
Targeting PI3K/Akt/mTOR signaling in cancer
.
Front Oncol
2014
;
4
:
64
.
50.
Jiao
Q
,
Bi
L
,
Ren
Y
,
Song
S
,
Wang
Q
,
Wang
YS
. 
Advances in studies of tyrosine kinase inhibitors and their acquired resistance
.
Mol Cancer
2018
;
17
:
36
.
51.
Weiss
GJ
,
Beck
J
,
Braun
DP
,
Bornemann-Kolatzki
K
,
Barilla
H
,
Cubello
R
, et al
Tumor cell-free DNA copy number instability predicts therapeutic response to immunotherapy
.
Clin Cancer Res
2017
;
23
:
5074
81
.
52.
Davoli
T
,
Uno
H
,
Wooten
EC
,
Elledge
SJ
. 
Tumor aneuploidy correlates with markers of immune evasion and with reduced response to immunotherapy
.
Science
2017
;
355
:
eaaf8399
.
53.
Zaretsky
JM
,
Garcia-Diaz
A
,
Shin
DS
,
Escuin-Ordinas
H
,
Hugo
W
,
Hu-Lieskovan
S
, et al
Mutations associated with acquired resistance to PD-1 blockade in melanoma
.
N Engl J Med
2016
;
375
:
819
29
.
54.
Murakami
Y
,
Saigo
K
,
Takashima
H
,
Minami
M
,
Okanoue
T
,
Brechot
C
, et al
Large scaled analysis of hepatitis B virus (HBV) DNA integration in HBV related hepatocellular carcinomas
.
Gut
2005
;
54
:
1162
8
.
55.
Campbell
PJ
,
Yachida
S
,
Mudie
LJ
,
Stephens
PJ
,
Pleasance
ED
,
Stebbings
LA
, et al
The patterns and dynamics of genomic instability in metastatic pancreatic cancer
.
Nature
2010
;
467
:
1109
13
.
56.
Quigley
DA
,
Dang
HX
,
Zhao
SG
,
Lloyd
P
,
Aggarwal
R
,
Alumkal
JJ
, et al
Genomic hallmarks and structural variation in metastatic prostate cancer
.
Cell
2018
;
174
:
758
69
.
57.
Torrecilla
S
,
Sia
D
,
Harrington
AN
,
Zhang
Z
,
Cabellos
L
,
Cornella
H
, et al
Trunk mutational events present minimal intra- and inter-tumoral heterogeneity in hepatocellular carcinoma
.
J Hepatol
2017
;
67
:
1222
31
.
58.
Furuta
M
,
Ueno
M
,
Fujimoto
A
,
Hayami
S
,
Yasukawa
S
,
Kojima
F
, et al
Whole genome sequencing discriminates hepatocellular carcinoma with intrahepatic metastasis from multi-centric tumors
.
J Hepatol
2017
;
66
:
363
73
.
59.
Li
Q
,
Wang
J
,
Juzi
JT
,
Sun
Y
,
Zheng
H
,
Cui
Y
, et al
Clonality analysis for multicentric origin and intrahepatic metastasis in recurrent and primary hepatocellular carcinoma
.
J Gastrointest Surg
2008
;
12
:
1540
7
.
60.
Xue
R
,
Li
R
,
Guo
H
,
Guo
L
,
Su
Z
,
Ni
X
, et al
Variable intra-tumor genomic heterogeneity of multiple lesions in patients with hepatocellular carcinoma
.
Gastroenterology
2016
;
150
:
998
1008
.
61.
Mills
CD
,
Lenz
LL
,
Harris
RA
. 
A breakthrough: macrophage-directed cancer immunotherapy
.
Cancer Res
2016
;
76
:
513
6
.
62.
Hagerling
C
,
Gonzalez
H
,
Salari
K
,
Wang
CY
,
Lin
C
,
Robles
I
, et al
Immune effector monocyte-neutrophil cooperation induced by the primary tumor prevents metastatic progression of breast cancer
.
Proc Natl Acad Sci U S A
2019
;
116
:
21704
14
.
63.
Kurebayashi
Y
,
Ojima
H
,
Tsujikawa
H
,
Kubota
N
,
Maehara
J
,
Abe
Y
, et al
Landscape of immune microenvironment in hepatocellular carcinoma and its additional impact on histological and molecular classification
.
Hepatology
2018
;
68
:
1025
41
.
64.
Zeidan
AM
,
Knaus
HA
,
Robinson
TM
,
Towlerton
AMH
,
Warren
EH
,
Zeidner
JF
, et al
A multi-center phase I trial of ipilimumab in patients with myelodysplastic syndromes following hypomethylating agent failure
.
Clin Cancer Res
2018
;
24
:
3519
27
.
65.
Peng
S
,
Wang
Y
,
Peng
H
,
Chen
D
,
Shen
S
,
Peng
B
, et al
Autocrine vascular endothelial growth factor signaling promotes cell proliferation and modulates sorafenib treatment efficacy in hepatocellular carcinoma
.
Hepatology
2014
;
60
:
1264
77
.
66.
Ikeda
M
,
Sung
MW
,
Kudo
M
,
Kobayashi
M
,
Baron
AD
,
Finn
RS
, et al
A phase 1b trial of lenvatinib (LEN) plus pembrolizumab (PEM) in patients (pts) with unresectable hepatocellular carcinoma (uHCC)
.
J Clin Oncol
36
:
15s
, 
2018
(
suppl; abstr 4076
).
67.
Hsu
CH
,
Lee
MS
,
Ryoo
BY
,
Stein
S
,
Lee
K-H
,
Verret
W
, et al
Safety and clinical activity results from atezolizumab + bevacizumab in hepatocellular carcinoma: updates from a phase Ib study [abstract]
. In:
Proceeding of the Asian Pacific Association for the Study of the Liver (APASL); Feb 20–24; Manila, Philippines
.
Tokyo, Japan
:
APASL
; 
2019
.

Supplementary data