Prognostic role of vascular endothelial growth factor in hepatocellular carcinoma: systematic review and meta-analysis.

Hepatocellular carcinoma (HCC) is a highly vascular tumour that expresses vascular endothelial growth factor (VEGF). Various studies have evaluated the prognostic value of VEGF levels in HCC. Its overall test performance remains unclear, however. The aim was to perform a systematic review and meta-analysis of prognostic cohort studies evaluating the use of VEGF as a predictor of survival in patients with treated HCC. Eligible studies were identified through multiple search strategies. Studies were assessed for quality using the Newcastle-Ottawa Tool. Data were collected comparing disease-free and overall survival in patients with high VEGF levels as compared to those with low levels. Studies were pooled and summary hazard ratios were calculated. A total of 16 studies were included for meta-analysis (8 for tissue and 8 for serum). Methodological analysis indicated a trend for higher study quality with serum studies as compared to tissue-based investigations. Four distinct groups were pooled for analysis: tissue overall survival (n=251), tissue disease-free survival (n=413), serum overall survival (n=579), and serum disease-free survival (n=439). High tissue VEGF levels predicted poor overall (HR=2.15, 95% CI: 1.26-3.68) and disease-free (HR=1.69, 95% CI: 1.23-2.33) survival. Similarly, high serum VEGF levels predicted poor overall (HR=2.35, 95% CI: 1.80-3.07) and disease-free (HR=2.36, 95% CI 1.76-3.16) survival. A high degree of inter-study consistency was present in three of four groups analysed. Tissue and serum VEGF levels appear to have significant predictive ability for estimating overall survival in HCC and may be useful for defining prognosis in HCC.

Proteasome inhibition causes apoptosis of normal human plasma cells preventing alloantibody production.

Antibody production by normal plasma cells (PCs) against human leukocyte antigens (HLA) can be a major barrier to successful transplantation. We tested four reagents with possible activity against PCs (rituximab, polyclonal rabbit antithymocyte globulin (rATG), intravenous immunoglobulin (IVIG) and the proteasome inhibitor, bortezomib) to determine their ability to cause apoptosis of human bone marrow-derived PCs and subsequently block IgG secretion in vitro. IVIG, rituximab and rATG all failed to cause apoptosis of PCs and neither rituximab nor rATG blocked antibody production. In contrast, bortezomib treatment led to PC apoptosis and thereby blocked anti-HLA and antitetanus IgG secretion in vitro. Two patients treated with bortezomib for humoral rejection after allogeneic kidney transplantation demonstrated a transient decrease in bone marrow PCs in vivo and persistent alterations in alloantibody specificities. Total IgG levels were unchanged. We conclude that proteasome activity is important for PC longevity and its inhibition may lead to new techniques of controlling antibody production in vivo.

mir-29 regulates Mcl-1 protein expression and apoptosis.

Cellular expression of Mcl-1, an anti-apoptotic Bcl-2 family member, is tightly regulated. Recently, Bcl-2 expression was shown to be regulated by microRNAs, small endogenous RNA molecules that regulate protein expression through sequence-specific interaction with messenger RNA. By analogy, we reasoned that Mcl-1 expression may also be regulated by microRNAs. We chose human immortalized, but non-malignant, H69 cholangiocyte and malignant KMCH cholangiocarcinoma cell lines for these studies, because Mcl-1 is dysregulated in cells with the malignant phenotype. By in silico analysis, we identified a putative target site in the Mcl-1 mRNA for the mir-29 family, and found that mir-29b was highly expressed in cholangiocytes. Interestingly, mir-29b was downregulated in malignant cells, consistent with Mcl-1 protein upregulation. Enforced mir-29b expression reduced Mcl-1 protein expression in KMCH cells. This effect was direct, as mir-29b negatively regulated the expression of an Mcl-1 3' untranslated region (UTR)-based reporter construct. Enforced mir-29b expression reduced Mcl-1 cellular protein levels and sensitized the cancer cells to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) cytotoxicity. Transfection of non-malignant cells (that express high levels of mir-29) with a locked-nucleic acid antagonist of mir-29b increased Mcl-1 levels and reduced TRAIL-mediated apoptosis. Thus mir-29 is an endogenous regulator of Mcl-1 protein expression, and thereby, apoptosis.

TRAIL resistance results in cancer progression: a TRAIL to perdition?

Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL, APO-2L) is a mediator of cell death that preferentially targets cancer cells. The potential of TRAIL as a chemotherapeutic agent is limited, however, because of the emergence of TRAIL resistance. Furthermore, recent studies have demonstrated that alternative TRAIL signaling is unmasked in TRAIL resistant cells. In these cells, the predominant effect of TRAIL receptor activation is the activation of nuclear factor-kappaB (NF-kappaB), which promotes tumor metastases and invasion. TRAIL resistance can occur at the level of the death inducing signaling complex via upregulation of cFLIP or via an increase in antiapoptotic proteins of the Bcl-2 family. A paradigm emerges from this information, that chemotherapy, targeting NF-kappaB, cFLIP, or antiapoptotic proteins of the Bcl-2 family, in combination with TRAIL maybe more rational than TRAIL therapy alone.

Increases of intracellular magnesium promote glycodeoxycholate-induced apoptosis in rat hepatocytes.

Retention of bile salts by the hepatocyte contributes to liver injury during cholestasis. Although cell injury can occur by one of two mechanisms, necrosis versus apoptosis, information is lacking regarding apoptosis as a mechanism of cell death by bile salts. Our aim was to determine if the bile salt glycodeoxycholate (GDC) induces apoptosis in rat hepatocytes. Morphologic assessment included electron microscopy and quantitation of nuclear fragmentation by fluorescent microscopy. Biochemical studies included measurements of DNA fragmentation, in vitro endonuclease activity, cytosolic free Ca2+ (Cai2+), and cytosolic free Mg2+ (Mgi2+). Morphologic studies demonstrated typical features of apoptosis in GDC (50 microM) treated cells. The "ladder pattern" of DNA fragmentation was also present in DNA obtained from GDC-treated cells. In vitro endonuclease activity was 2.5-fold greater with Mg2+ than Ca2+. Although basal Cai2+ values did not change after addition of GDC, Mgi2+ increased twofold. Incubation of cells in an Mg(2+)-free medium prevented the rise in Mgi2+ and reduced nuclear and DNA fragmentation. In conclusion, GDC induces apoptosis in hepatocytes by a mechanism promoted by increases of Mgi2+ with stimulation of Mg(2+)-dependent endonucleases. These data suggest for the first time that changes of Mgi2+ may participate in the program of cellular events culminating in apoptosis.

Glycochenodeoxycholate-induced lethal hepatocellular injury in rat hepatocytes. Role of ATP depletion and cytosolic free calcium.

Chenodeoxycholate is toxic to hepatocytes, and accumulation of chenodeoxycholate in the liver during cholestasis may potentiate hepatocellular injury. However, the mechanism of hepatocellular injury by chenodeoxycholate remains obscure. Our aim was to determine the mechanism of cytotoxicity by chenodeoxycholate in rat hepatocytes. At a concentration of 250 microM, glycochenodeoxycholate was more toxic than either chenodeoxycholate or taurochenodeoxycholate. Cellular ATP was 86% depleted within 30 min after addition of glycochenodeoxycholate. Fructose, a glycolytic substrate, maintained ATP concentrations at 50% of the initial value and protected against glycochenodeoxycholate cytotoxicity. ATP depletion in the absence of a glycolytic substrate suggested impairment of mitochondrial function. Indeed, glycochenodeoxycholate inhibited state 3 respiration in digitonin-permeabilized cells in a dose-dependent manner. After ATP depletion, a sustained rise in cytosolic free calcium (Cai2+) was observed. Removal of extracellular Ca2+ abolished the rise in Cai2+, decreased cellular proteolysis, and protected against cell killing by glycochenodeoxycholate. The results suggest that glycochenodeoxycholate cytotoxicity results from ATP depletion followed by a subsequent rise in Cai2+. The rise in Cai2+ leads to an increase in calcium-dependent degradative proteolysis and, ultimately, cell death. We conclude that glycochenodeoxycholate causes a bioenergetic form of lethal cell injury dependent on ATP depletion analogous to the lethal cell injury of anoxia.

Selective induction of apoptosis in Hep 3B cells by topoisomerase I inhibitors: evidence for a protease-dependent pathway that does not activate cysteine protease P32.

Progress in the treatment of hepatocellular carcinoma (HCC), a common tumor worldwide, has been disappointing. Inhibitors of topoisomerases are being widely studied as potential inducers of tumor cell apoptosis. Our aims were to determine whether topoisomerase-directed drugs would induce apoptosis in a human HCC cell line (Hep 3B) and, if so, to investigate the mechanism. The topoisomerase I poison camptothecin (CPT) induced apoptosis of Hep 3B cells in a time- and concentration-dependent manner. In contrast, the topoisomerase II poison etoposide failed to induce apoptosis despite the apparent stabilization of topoisomerase II-DNA complexes. Unexpectedly, CPT-induced apoptosis in this cell type occurred without any detectable cleavage of poly(ADP-ribose) polymerase or lamin B, polypeptides that are commonly cleaved in other cell types undergoing apoptosis. Likewise, Hep 3B cell apoptosis occurred without a detectable increase in interleukin-1beta-converting enzyme (ICE)-like or cysteine protease P32 (CPP32)-like protease activity. In contrast, trypsin-like protease activity (cleavage of Boc-Val-Leu-Lys-chloromethylaminocoumarin in situ) increased threefold in cells treated with CPT but not etoposide. Tosyl-lysyl chloromethyl ketone inhibited the trypsin-like protease activity and diminished CPT-induced apoptosis. These data demonstrate that (a) apoptosis is induced in Hep 3B cells after stabilization of topoisomerase I-DNA complexes but not after stabilization of topoisomerase II-DNA complexes as measured by alkaline filter elution; (b) Hep 3B cell apoptosis occurs without activation of ICE-like and CPP32-like protease activity; and (c) a trypsin-like protease activity appears to contribute to apoptosis in this cell type.

Intracellular pH during "chemical hypoxia" in cultured rat hepatocytes. Protection by intracellular acidosis against the onset of cell death.

The relationships between extracellular pH (pHo), intracellular pH (pHi), and loss of cell viability were evaluated in cultured rat hepatocytes after ATP depletion by metabolic inhibition with KCN and iodoacetate (chemical hypoxia). pHi was measured in single cells by ratio imaging of 2',7'-biscarboxy-ethyl-5,6-carboxyfluorescein (BCECF) fluorescence using multiparameter digitized video microscopy. During chemical hypoxia at pHo of 7.4, pHi decreased from 7.36 to 6.33 within 10 min. pHi remained at 6.1-6.5 for 30-40 min (plateau phase). Thereafter, pHi began to rise and cell death ensued within minutes, as evidenced by nuclear staining with propidium iodide and coincident leakage of BCECF from the cytoplasm. An acidic pHo produced a slightly greater drop in pHi, prolonged the plateau phase of intracellular acidosis, and delayed the onset of cell death. Inhibition of Na+/H+ exchange also prolonged the plateau phase and delayed cell death. In contrast, monensin or substitution of gluconate for Cl- in buffer containing HCO3- abolished the pH gradient across the plasma membrane and shortened cell survival. The results indicate that intracellular acidosis after ATP depletion delays the onset of cell death, whereas reduction of the degree of acidosis accelerates cell killing. We conclude that intracellular acidosis protects against hepatocellular death from ATP depletion, a phenomenon that may represent a protective adaptation against hypoxic and ischemic stress.

Overexpression of the TGFbeta-regulated zinc finger encoding gene, TIEG, induces apoptosis in pancreatic epithelial cells.

Members of the TGFbeta family of peptides exert antiproliferative effects and induce apoptosis in epithelial cell populations. In the exocrine pancreas, these peptides not only regulate normal cell growth, but alterations in these pathways have been associated with neoplastic transformation. Therefore, the identification of molecules that regulate exocrine pancreatic cell proliferation and apoptotic cell death in response to TGFbeta peptides is necessary for a better understanding of normal morphogenesis as well as carcinogenesis of the pancreas. In this study, we have characterized the expression and function in exocrine pancreatic epithelial cells of the TGFbeta-inducible early gene (TIEG), a Krüppel-like zinc finger transcription factor encoding gene previously isolated from mesodermally derived osteoblastic cells. We demonstrate that this gene is expressed in both acinar and ductular epithelial cell populations from the exocrine pancreas. In addition, we show that the expression of TIEG is regulated by TGFbeta1 as an early response gene in pancreatic epithelial cell lines. Moreover, overexpression of TIEG in the TGFbeta-sensitive epithelial cell line PANC1 is sufficient to induce apoptosis. Together, these results support a role for TIEG in linking TGFbeta-mediated signaling cascades to the regulation of pancreatic epithelial cell growth.

Cathepsin B contributes to TNF-alpha-mediated hepatocyte apoptosis by promoting mitochondrial release of cytochrome c.

TNF-alpha-induced apoptosis is thought to involve mediators from acidic vesicles. Cathepsin B (cat B), a lysosomal cysteine protease, has recently been implicated in apoptosis. To determine whether cat B contributes to TNF-alpha-induced apoptosis, we exposed mouse hepatocytes to the cytokine in vitro and in vivo. Isolated hepatocytes treated with TNF-alpha in the presence of the transcription inhibitor actinomycin D (AcD) accumulated cat B in their cytosol. Further experiments using cell-free systems indicated that caspase-8 caused release of active cat B from purified lysosomes and that cat B, in turn, increased cytosol-induced release of cytochrome c from mitochondria. Consistent with these observations, the ability of TNF-alpha/AcD to induce mitochondrial release of cytochrome c, caspase activation, and apoptosis of isolated hepatocytes was markedly diminished in cells from CatB(-/-) mice. Deletion of the CatB gene resulted in diminished liver injury and enhanced survival after treatment in vivo with TNF-alpha and an adenovirus construct expressing the IkappaB superrepressor. Collectively, these observations suggest that caspase-mediated release of cat B from lysosomes enhances mitochondrial release of cytochrome c and subsequent caspase activation in TNF-alpha-treated hepatocytes.

Toxic bile salts induce rodent hepatocyte apoptosis via direct activation of Fas.

Cholestatic liver injury appears to result from the induction of hepatocyte apoptosis by toxic bile salts such as glycochenodeoxycholate (GCDC). Previous studies from this laboratory indicate that cathepsin B is a downstream effector protease during the hepatocyte apoptotic process. Because caspases can initiate apoptosis, the present studies were undertaken to determine the role of caspases in cathepsin B activation. Immunoblotting of GCDC-treated McNtcp.24 hepatoma cells demonstrated cleavage of poly(ADP-ribose) polymerase and lamin B1 to fragments that indicate activation of effector caspases. Transfection with CrmA, an inhibitor of caspase 8, prevented GCDC-induced cathepsin B activation and apoptosis. Consistent with these results, an increase in caspase 8-like activity was observed in GCDC-treated cells. Examination of the mechanism of GCDC-induced caspase 8 activation revealed that dominant-negative FADD inhibited apoptosis and that hepatocytes isolated from Fas-deficient lymphoproliferative mice were resistant to GCDC-induced apoptosis. After GCDC treatment, immunoprecipitation experiments demonstrated Fas oligomerization, and confocal microscopy demonstrated DeltaFADD-GFP (Fas-associated death domain-green fluorescent protein, aggregation in the absence of detectable Fas ligand mRNA. Collectively, these data suggest that GCDC-induced hepatocyte apoptosis involves ligand-independent oligomerization of Fas, recruitment of FADD, activation of caspase 8, and subsequent activation of effector proteases, including downstream caspases and cathepsin B.

Review article: the modern diagnosis and therapy of cholangiocarcinoma.

Cholangiocarcinomas are epithelial neoplasms that originate from cholangiocytes and can occur at any level of the biliary tree. They are broadly classified into intrahepatic tumours, (extrahepatic) hilar tumours and (extrahepatic) distal bile duct tumours. In spite of well-understood predispositions, most cholangiocarcinomas arise in the absence of risk factors. In suspected cases, the diagnosis can be established with non-invasive imaging studies. Biliary invasion should be reserved for patients with obstruction. In high-risk patients, advanced cytological tests of aneuploidy (digital image analysis and fluorescent in situ hybridization) aid early diagnosis. In the absence of primary sclerosing cholangitis, curative surgical resection has 5-year survival rates of 2-43%, higher survival observed in patients with clear surgical margins and concomitant hepatic resection for hilar tumours. Patients with unresectable cholangiocarcinoma or pre-existing primary sclerosing cholangitis should be considered for liver transplantation with neoadjuvant chemoirradiation, in specialized centres.

Lysophosphatidylcholine suppresses apoptotic cell death by inducing cyclooxygenase-2 expression via a Raf-1 dependent mechanism in human cholangiocytes.

PURPOSE: The high incidence of biliary tract carcinoma in patients with anomalous pancreaticobiliary ductal junction (APBDJ) implicates that a compositional alteration in bile may contribute to the genesis of this cancer. Lysophosphatidylcholine (LPC) is generated in the bile of these patients. Given the role of cyclooxygenase-2 (COX-2) in biliary tract carcinogenesis, we postulated that LPC induces COX-2 in cholangiocytes. METHODS: The effect of LPC on COX-2 expression in cholangiocytes was evaluated by immunoblot analysis, real-time PCR and reporter gene assay. Apoptosis was induced by TRAIL treatment, and quantified using DAPI staining. RESULTS: Lysophosphatidylcholine increased COX-2 protein expression in cholangiocytes in a concentration- and time-dependent manner. LPC-induced Raf-1 activation was responsible for this COX-2 induction. Accordingly, LPC increased COX-2 mRNA levels in a Raf-1 dependent manner by stabilizing COX-2 mRNA. Finally, LPC attenuated TRAIL-mediated apoptosis through a COX-2/PgE2 dependent mechanism. CONCLUSIONS: Collectively, these results implicate that LPC inhibits cholangiocyte apoptosis by inducing COX-2 expression via a Raf-1 dependent mechanism. This anti-apoptotic signaling may participate in biliary tract carcinogenesis in APBDJ patients, and therefore, its interruption may be a viable chemopreventative strategy.

Inflammatory cytokines induce DNA damage and inhibit DNA repair in cholangiocarcinoma cells by a nitric oxide-dependent mechanism.

Chronic infection and inflammation are risk factors for the development of cholangiocarcinoma, a highly malignant, generally fatal adenocarcinoma originating from biliary epithelia. However, the link between inflammation and carcinogenesis in these disorders is obscure. Because nitric oxide (NO) is generated in inflamed tissues by inducible nitric oxide synthase (iNOS) and because DNA repair proteins are potentially susceptible to NO-mediated nitrosylation, we formulated the hypothesis that inflammatory cytokines induce iNOS and sufficient NO to inhibit DNA repair enzymes leading to the development and progression of cholangiocarcinoma. iNOS and nitrotyrosine were demonstrated in 18/18 cholangiocarcinoma specimens. Furthermore, iNOS and NO generation could be induced in vitro by inflammatory cytokines (mixture of interleukin-1beta, IFN-gamma, and tumor necrosis factor alpha) in three human cholangiocarcinoma cell lines. NO-dependent DNA damage as assessed by the comet assay was demonstrated during exposure of the three cholangiocarcinoma cell lines to cytokines. Moreover, global DNA repair activity was inhibited by 70% by a NO-dependent process after exposure of cells to cytokines. Our data indicate that activation of iNOS and excess production of NO in response to inflammatory cytokines cause DNA damage and inhibit DNA repair proteins. NO inactivation of DNA repair enzymes may provide a link between inflammation and the initiation, promotion, and/or progression of cholangiocarcinoma.

Human Ogg1, a protein involved in the repair of 8-oxoguanine, is inhibited by nitric oxide.

NO-mediated inhibition of base excision DNA repair may potentiate oxidativeDNA damage in cells and could be relevant to carcinogenesis associated with chronic inflammation. Because 8-oxoguanine, a ubiquitous oxidative DNA lesion, is repaired predominantly by human 8-oxoguanine glycosylase (hOgg1), our aim was to determine whether NO directly inhibits its repair activity. Neither induction of NO-generating enzyme inducible NO synthase nor treatment with S-nitroso-N-acetyl-D-L-pencillamine altered expression of hOgg1 in a human cholangiocarcinoma cell line (KMBC). In contrast, both treatments completely inhibited activity of hOgg1 immunoprecipitated from KMBC cells overexpressing hOgg1 and in a cell-free system. Both NO and peroxynitrite were capable of inhibiting hOgg1 activity. Inhibition of hOgg1 protein was characterized by formation of S-nitrosothiol adducts and loss/ejection of zinc ions. Our data indicate that NO, an inflammatory mediator, directly inhibits a key base excision repair enzyme (hOgg1) responsible for base excision repair of 8-oxoguanine. These data support the concept that NO-mediated inhibition of DNA contributes to the mutagenic environment of chronic inflammation.

Viral fusogenic membrane glycoprotein expression causes syncytia formation with bioenergetic cell death: implications for gene therapy.

Viral fusogenic membrane glycoproteins (FMGs) are candidates for gene therapy of solid tumors because they cause cell fusion, leading to formation of lethal multinucleated syncytia. However, the cellular mechanisms mediating cell death after FMG-induced cell fusion remain unclear. The present study was designed to examine the mechanisms by which FMG expression in hepatocellular carcinoma cells lead to cell death. Transfection of Hep3B cells with the Gibbon Ape leukemia virus hyperfusogenic envelope protein (GALV-FMG) resulted in the formation of multinucleated syncytia that reached a maximum 5 days after transfection (100 nuclei/syncytia). The syncytia were viable for a period of 2 days and then rapidly lost viability by day 5. Mitochondrial dysfunction occurred in GALV-FMG-induced syncytia prior to loss of viability with loss of the mitochondrial membrane potential, cellular ATP depletion, and release of mitochondrial cytochrome c-GFP into the cytosol. The pan-caspase inhibitor, Z-VAD-fmk, did not prevent cell death. However, glycolytic generation of ATP with fructose effectively increased cellular ATP and preserved syncytial viability. These data suggest that expression of FMG in hepatoma cells results in the formation of multinucleated syncytia, causing mitochondrial failure with ATP depletion, a bioenergetic form of cell death with necrosis. This form of cell death should be effective in vivo and enhance the bystander effect, suggesting that FMG-based gene therapy deserves further study for the treatment of hepatocellular and other cancers.

Induction of the mitochondrial permeability transition as a mechanism of liver injury during cholestasis: a potential role for mitochondrial proteases.

As part of this thematic series on mitochondria in cell death, we would like to review our data on: (1) the role of the mitochondrial permeability transition (MPT) in hepatocyte necrosis during cholestasis; and (2) the concept that endogenous mitochondrial protease activity may lead to the MPT. Many chronic human liver diseases are characterized by cholestasis, an impairment in bile flow. During cholestasis an accumulation of toxic hydrophobic bile salts in the hepatocyte causes necrosis. We tested the hypothesis that toxic hydrophobic bile salt, glycochenodeoxycholate (GCDC), causes hepatocyte necrosis by inducing the MPT. GCDC induces a rapid, cyclosporin A-sensitive MPT. The hydrophilic bile salt, ursodeoxycholate (UDCA), prevents the GCDC-induced MPT and hepatocyte necrosis providing an explanation for its beneficial effect in human liver disease. We have also demonstrated that the calcium-dependent MPT is associated with an increase in calpain-like protease activity and inhibited by calpain inhibitors. In an experimental model of cholestasis, mitochondrial calpain-like protease activity increases 1.6-fold. We propose for the first time that activation of mitochondrial proteases may initiate the MPT and cell necrosis during cholestasis.

Phase II study of the efficacy and safety of cisplatin-epinephrine injectable gel administered to patients with unresectable hepatocellular carcinoma.

PURPOSE: To study the efficacy and safety of percutaneous cisplatin-epinephrine (CDDP-EPI) injectable gel in patients with localized unresectable hepatocellular carcinoma (HCC). PATIENTS AND METHODS: Eligible patients had histologically proven HCC, no prior treatment except for surgery, and no more than three tumors (each measured < or = 7 cm, total tumor volume < or = 200 cm(3)). They were treated percutaneously under ultrasound or computed tomography (CT) guidance, with up to 10 mL of CDDP-EPI gel (1 mL contains 4 mg of CDDP and 0.1 mg of EPI) per treatment and four treatments in 6 weeks to a maximum of eight treatments. The primary end points were tumor response, defined by change of percentage of tumor necrosis according to CT criteria, and safety. Survival parameters were secondary end points. RESULTS: From June 1997 to April 2000, 58 patients (median age, 65 years) entered the study. All patients were assessable for safety, and 51 were assessable for efficacy. The median number of treatments was four (range, one to eight treatments). Objective response rate was 53% (27 of 51 patients), including 16 complete and 11 partial responses. Of the 27 responders, 14 (52%) subsequently developed progressive disease, but in most of them (93%), a new tumor arose at untreated liver sites. Median survival was 27 months (range, 18.4 to 35.7 months). The 1-, 2-, and 3-year survival rates were 79%, 56%, and 14% respectively. The procedure was well tolerated with only minor side effects. CONCLUSION: Percutaneous local ablation with CDDP-EPI injectable gel can induce significant tumor necrosis and local control for localized unresectable HCC, and the treatment is well tolerated.