The successful management of hepatocellular carcinoma (HCC) requires a multidisciplinary approach, incorporating hepatologists, oncologists, surgical oncologists, transplant surgeons, and radiologists. With improvements in technology and better long-term outcomes data, management strategies for HCC have become more methodical and more successful. This article focuses on some of the most critical advances relating to carcinogenesis, surveillance, and management.
Key points
- •
The progress made in the diagnosis and management of hepatocellular carcinoma (HCC) represents one of the growing successes in surgical oncology.
- •
Despite advances in HCC diagnosis and management, the incidence of HCC is still increasing, and HCC represents the fifth most common cancer and the third most common cause of cancer death worldwide.
- •
Over the last 20 years alone, advances have been made to elucidate the mechanisms of carcinogenesis, to diagnose disease at an earlier stage, and to improve local and systemic treatment of HCC.
Introduction
The progress made in the diagnosis and management of hepatocellular carcinoma (HCC) represents one of the growing successes in surgical oncology. Despite these advances, the incidence of HCC is still increasing, and HCC represents the fifth most common cancer and the third most common cause of cancer death worldwide. Over the last 20 years alone, advances have been made to elucidate the mechanisms of carcinogenesis, to diagnose disease at an earlier stage, and to improve local and systemic treatment of HCC.
Introduction
The progress made in the diagnosis and management of hepatocellular carcinoma (HCC) represents one of the growing successes in surgical oncology. Despite these advances, the incidence of HCC is still increasing, and HCC represents the fifth most common cancer and the third most common cause of cancer death worldwide. Over the last 20 years alone, advances have been made to elucidate the mechanisms of carcinogenesis, to diagnose disease at an earlier stage, and to improve local and systemic treatment of HCC.
Carcinogenesis and diagnosis
Genetics
As technology has improved, the mechanisms for the generation of cirrhosis and subsequent HCC behave become better understood. The well-known environmental risk factors that may lead to underlying cirrhosis include hepatitis B virus (HBV), hepatitis C virus (HCV), exposure to toxins such as aflatoxin, and alcohol intake. For each of these causes of HCC, specific genetic mutations have been isolated. In HCV-related HCC, mutations have been identified in p53, in the disintegrin and metalloproteinase domain–containing protein 22 (ADAM22), in the Janus kinase/signal transducer and activator of transcription (JAK) pathway, in the beta-catenin gene CTNNB1, in the transport protein particle (TRAPP), in the never in mitosis A–related kinase 8 (NEK8) gene, and in the AT-rich interactive domain 2 (ARID2) gene. HBV-related HCC is associated with p53 mutations but also with exclusive mutations in ATPase family AAA domain–containing 2 (ATAD2) and interferon regulatory factor 2 (IRF2) genes. Although there is some overlap in the genetic mutations responsible for HCC in the background of HBV versus HCV, there are notable differences, with HCV being associated with increased CTNNB1 mutations and fewer p53 mutations. Alcohol consumption has shown a correlation to mutations in the chromatin remodelers, which predispose to dysregulation and the development of HCC. The mechanism for the development of aflatoxin-induced HCC has been genetically described by specific base substitutions, which can lead to HCC in the absence of any underlying liver disease.
From a population-based perspective, specific patient polymorphisms have also recently been identified as potential risk factors for the development of chronic hepatitis and cirrhosis. Using genome-wide association analyses and single-nucleotide polymorphisms, subsets of patients have been identified at a specifically higher risk for the development of HCC, independent of the well-established external exposures.
Molecular Mechanisms
The downstream pathways in which HBV and HCV promote HCC are becoming better understood. One such pathway is through promoting stem-cell activity, which has been shown by the upregulation of the well-known stemness-associated marker epithelial cell adhesion molecule (EpCAM) and beta-catenin. Other more recent markers to show stem cell–like properties in HCC include the NANOG transcription factor, octamer-binding transcription factor 4 (OCT4), sex-determining region Y box 2 (SOX2), and Kruppel-like factor 4.
Another newly uncovered mechanism of HCC carcinogenesis is secondary to the relative hypoxia and subsequent angiogenesis incurred by the cirrhotic liver. At the macroscopic level, nodular cirrhosis leads to a decrease in hepatic vasculature, which is followed by a hypoxic environment. In the setting of hypoxia, there is upregulation of hypoxia inducible factor 1 alpha (H1F1α). Stimulation of this factor leads to upregulation of vascular endothelial growth factor (VEGF), cyclo-oxygenase 2, angiopoietin 2, and several matrix metalloproteinases. These inappropriately upregulated angiogenic and inflammatory signals predispose the underlying parenchyma to damage, inhibition of regeneration, and subsequent HCC. Specific mechanisms have also shown that both HBV and HCV have unique upregulation of the transcription factor HIF1α at the genetic level. The downstream activation of these proangiogenic growth signals has shown promise in the systemic management of HCC, because some of these factors may be targeted and blocked with agents like sorafenib.
HBV and HCV have also been shown to self-inhibit their viral clearance from infected liver cells. This phenomenon of avoiding clearance is multifaceted. First, immune cell types and their reactive dysfunction from HBV and HCV can include dendritic cells that are made defective, regulatory T cells that are inappropriately induced, and CD8+ effector T cells that are downregulated. The innate immune response and cytokines are also disrupted with HCV and HBV, with interferon decreased, and natural killer cells and polymorphonuclear cells upregulated. This immune imbalance leads to unregulated inflammation and an environment for the establishment of cirrhosis and subsequent cancer.
Tumor Markers and Screening
Screening for HCC has been studied in many trials and suggested in numerous guidelines. The most accepted and updated guidelines are available from the American Association for the Study of Liver Diseases (AASLD). One of the notable recent changes from prior algorithms is the decreased use of serum levels of alfa-fetoprotein (AFP), and the increased reliance on surveillance ultrasound imaging of the liver every 6 months. When a lesion is suspected on ultrasound, contrast enhanced dynamic imaging is indicated. The hallmark radiologic signs of HCC include intense arterial uptake followed by washout of contrast in the venous phase ( Fig. 1 ). If the lesion does not have these characteristics, a biopsy should be considered. If the lesion is less than 1 cm it is more difficult to make a definitive diagnosis of HCC versus a regenerative nodule; as such, the lesion should, at a minimum, be closely followed with surveillance imaging every 3 to 6 months.
Although AFP has been used in the past as a serum tumor marker for HCC, its sensitivity and specificity are limited. Factors that influence the usefulness of AFP include the disadvantage that AFP is increased in patients with hepatitis and chronic liver disease without HCC, and the finding that AFP is correlated with tumor size (ie, small tumors are less likely to have increased AFP). To address the problem of poor specificity for HCC, focus has been dedicated on differentiating specific AFPs based on the degree of glycosylation and correlation with disease state. One of the first markers examined has been the fucosylation variant lectin ( Lens culinaris agglutinin) fraction of AFP (AFP-L3), which has been shown to be highly specific for HCC. The relative percentage of AFP-L3 to AFP has also been shown to be an indicator of poor prognosis on multivariate analysis for patients with HCC.
Another tumor marker being adopted into practice is Des-gamma-carboxy-prothrombin (DCP), a protein secreted in the setting of abnormal hepatocellular function. Sensitivities and specificities for DCP range from 28% to 89% and 87% to 96%, respectively. Like AFP-L3, DCP has shown promise to correlate with HCC stage, portal vein invasion, and prognosis; furthermore, DCP does not seem to be increased in the setting of chronic liver disease.
AFP, AFP-L3, and DCP have shown no relationship to each other. As such, attempts have been made to improve screening accuracy of patients with cirrhosis using all three markers in an additive fashion. Tateishi and colleagues, using AFP greater than 200 ng/mL, DCP greater than 40 mAU/mL, and AFP-L3 greater than 15%, found that the accuracy of these tumor markers in combination was higher than that of any test in isolation. However, one of the greatest hindrances to developing an accurate combinatorial test is standardizing assays and cutoff values. Despite better serologic markers becoming available, some groups have recommended abandoning serologic markers and relying on screening ultrasound. Our practice has been to not abandon AFP, but to use it while also recognizing that a normal AFP cannot be relied on to exclude HCC.
Biomarkers have also shown promise as potential indicators of developing HCC in the setting of cirrhosis. Their role in screening protocols and staging guidelines is still in development. Some of these biomarkers are listed in Table 1 .
| Marker | Level | Mechanism | Usefulness |
|---|---|---|---|
| Osteopontin | >156 ng/mL | Upregulated phosphoprotein in setting of tumor invasion, progression, or metastasis | When combined with AFP cutoff of 20 ng/mL, sensitivity and specificity were 95% and 96% in diagnosing HCC |
| Vascular endothelial growth factor | >245 pg/mL; >450 pg/mL | Marker of angiogenesis | Predicts overall and cancer-free survival |
| Hepatocyte growth factor | >0.6 ng/mL | Nonspecific growth factor | In patients with HCV, increased levels correlated with diagnosis of HCC |
| Insulinlike growth factor 1 | Reduction of serum levels | Endocrine, paracrine, and mitogenic role | Decreasing insulinlike growth factor 1 levels preceded development of HCC |
Surgical therapy
Once the diagnosis of HCC is suspected, therapeutic options include surgical resection, transplant, systemic chemotherapy, or locoregional therapies (eg, ablation, transarterial chemoembolization [TACE]). The factors that determine which treatment plan is most appropriate include tumor burden, underlying liver function, and patient performance status.
Resection
The perioperative mortality for HCC surgical resection is higher (4%–4.7%) than for resection for benign disease or colorectal liver metastasis, which is likely a reflection of the burden of chronic liver disease in patients with HCC. Therefore, patient selection is especially critical in the evaluation of patients with HCC for surgical resection. The best candidates for surgical resection of HCC are those with early stage, minimal, or well-compensated cirrhosis, and good performance status. Adequate cross-sectional contrast-enhanced imaging is one of the first tests in diagnosing HCC. If clinical suspicion is high (by history and increased serum markers), combined with improved imaging, there is no indication for biopsy. Imaging is also critical in determining suitability for resection because as it provides insight into tumor burden for staging, presence of vascular invasion of the primary tumor, presence of cirrhosis, and information for predicting the future liver remnant (FLR). Factors that could exclude appropriate resection are multiple tumors in multiple segments, vascular invasion, evidence of severe cirrhosis, or an FLR that is too small ( Fig. 2 ).
In order to evaluate FLR, both the function and the volume of the liver need consideration. For assessing volume, both computed tomography (CT) and magnetic resonance imaging (MRI) volumetrics have been used. Data by Kubota and colleagues were instrumental in establishing early guidelines for extent of resection and FLR prediction. This group was able to correlate preoperative volumetric CT imaging and liver function with guidelines for the extent of safe resection of HCC. From their assessments and the adoption by others, hepatectomy with a remnant of 20% to 30% is considered safe for patients with normal liver, 30% to 40% for patients with chronic hepatitis, and 40% to 50% for cirrhotics.
To assess function, Child-Pugh classification has historically been the assessment of choice. This score is based on 3 biochemical parameters (bilirubin, albumin, and prothrombin time) as well as 2 clinical parameters (ascites and encephalopathy). The Child-Pugh score correlates with morbidity and mortality after hepatectomy. More recently, native Model for End-Stage Liver Disease (MELD) score has also been shown to be helpful in identifying patients at highest risk of morbidity, mortality, and specifically postoperative liver failure. Delis and colleagues retrospectively examined a population of 69 patients with HCC and cirrhosis and showed that, along with American Society of Anesthesiologists (ASA) score, MELD scores (≤9) were independent predictors of perioperative mortality (7.2% vs 19%; P <.02) and overall morbidity (36.23% vs 48%; P <.02).
Functional tools to assess liver function include measurement of indocyanine green clearance, galactose elimination capacity, lidocaine metabolism, and ratios of arterial body ketones. Although these tools have shown promise in predicting perioperative outcomes after hepatectomy with patients with HCC, they are still limited in that they measure global liver function. To more accurately assess more focal liver function, nuclear imaging techniques like 99 mTc-galactosyl serum albumin scintigraphy and 99 mTc-mebrofenin hepatobiliary scintigraphy have been used and show promise for clinical application.
Some patients with inadequate FLR should be considered for portal vein embolization (PVE). PVE causes atrophy of the ipsilateral lobe of the liver, which is embolized with contralateral lobar hypertrophy, thereby increasing the FLR. Access to selective PVE is typically performed percutaneously and the portal vein is occluded with coils, gel foam, or glue. Prospective trials and mechanistic studies have supported the application of PVE in the management of HCC. Ribero and colleagues described the safety of PVE and its efficacy for growing the FLR. These investigators showed that patients with an adequate FLR were at lower risk of postoperative morbidity and liver insufficiency. Furthermore, they noted that the degree/rate of liver hypertrophy was equally, or more, important as the absolute final FLR volume after PVE. These investigators reported that a degree of hypertrophy less than 5% had a sensitivity of 80% and specificity of 94% in predicting hepatic dysfunction. Minimal hypertrophy after PVE was associated with a greater risk for longer hospital stays and 90-day mortality.
If patients are deemed appropriate candidates, they may undergo resection. Outcomes from surgical resection are variable, with 5-year survival rates ranging from 20% to 70%. Explanations for this variability are attributed to the heterogeneity of the population of patients who are resected. This variability includes small tumors (<5 cm), tumors in the setting of multinodular disease, large tumors (>5–10 cm), and tumors with major vascular invasion (MVI). Many groups report outcomes of resection for tumors that would not be considered for liver transplantation, and the outcomes among these patients are associated with earlier recurrence and shorter survival compared with patients with more limited disease.
Many groups have characterized the best outcomes after hepatic resection for small (<5 cm tumors). Poon and colleagues examined a series of Child-Pugh class A patients with HCC with solitary tumors less than 5 cm, fewer than 3 tumors less than 3 cm, and tumors without MVI, and showed excellent 5-year survival (68%). Shi and colleagues showed similar extended survival after resection in this subset of patients, and additionally reported 5-year survival for tumors less than 2 cm at 100%. In a recent review by Nathan and colleagues, examining a 10-year era of the Surveillance Epidemiology and End Results (SEER) database, the investigators reported that, among patients with small tumors, outcomes were excellent and that surgery may even have been underused in this group that would benefit from resection.
In the past, multinodular HCC has represented a subset of patients who have had poor perioperative and long-term outcomes following resection, with 5-year survival rates as low as 25% to 30%. However, secondary to advances in diagnosis, surgical techniques, and perioperative care, recent data have shown markedly better outcomes in this cohort of high-risk patients. Ishizawa and colleagues showed that resection of multinodular HCC had an improved 5-year overall survival of 58%. Although transplantation should be considered the mainstay of therapy for these patients, up to 20% of patients with multinodular disease progress and do not make it to transplant. As such, some investigators have argued that resection of HCC in the setting of multinodular disease should not be considered an absolute contraindication to resection.
Large tumors (>5–10 cm) pose unique prognostic and treatment challenges for resection as well. As with multinodular disease, the 5-year overall survival after resection for large tumors in the past has been poor, ranging from 16% to 33%. However, more recently, with the advent of better techniques and therapies, specifically PVE, the outcomes for resection of large tumors has improved. Vauthey and colleagues examined outcomes after resection of large HCC (or multinodular) tumors and found an overall 5-year survival of 39%. In a separate study, Yang and colleagues similarly noted a 5-year overall survival of 38.5% among patients with large HCC, with an acceptable perioperative morbidity and mortality.
Major vascular invasion represents one of the most important poor prognostic factors after resection, with 5-year survival approximately 60% without MVI, and 10% with MVI; worse survival correlates with more proximal and extensive invasion of the portal veins. For patients with proximal MVI, hepatic resection is generally contraindicated. However, if MVI is not extensive and involves only sectorial branches, then hepatic resection should be considered in the setting of adequate FLR, potentially after treating with locoregional therapies. In a multicenter study, Pawlik and colleagues showed that resection of HCC with MVI improved median survival and that 5-year survival was improved from 5% to 23% depending on the presence of minimal or no liver fibrosis ( P = .001). However, perioperative mortality in these patients needs to be carefully considered because it may be as high as 7%.
Transplantation
Liver transplantation is the only treatment modality that offers the potential to both treat the HCC and cure the underlying liver disease. Throughout most of the 1980s, transplantation for HCC had recurrence rates of greater than 50%. Survival was also disappointing, with 5-year survival rates ranging from 10% to 35%. From these early data, it became evident that the successes and prognosis of transplantation for HCC relied on the biology of the tumor (size, number of nodules, vascular invasion).
In 1996, Mazzaferro and colleagues published their seminal study, which showed that patients with a solitary tumor of less than 5 cm, or those with a maximum of 3 nodules less than or equal to 3 cm, without vascular invasion or extrahepatic spread, had improved recurrence (5%–15%) and survival (>70%). Since the adoption of the Milan criteria, other groups have validated the success of this prioritization tool for transplantation. The United Network of Organ Sharing (UNOS) has since standardized liver allocation in the United States for patients with HCC who meet the Milan criteria, automatically appointing them 22 MELD points in an effort to prioritize them. Given the reported successes of the criteria, other centers have questioned how far the Milan criteria can be expanded without affecting transplantation outcomes. The best known expanded criteria are probably those proposed by the University of California, San Francisco (UCSF). The UCSF Criteria are more inclusive than the Milan criteria: 1 tumor less than 6.5 cm, or 2 or 3 nodules less than 4.5 cm with a total tumor diameter less than 8 cm. The UCSF group published results showing that transplant using these expanded criteria can lead to results comparable with those obtained with the Milan criteria. Other groups have similarly reported reasonable long-term results with expanded criteria, particularly in the context of adjuvant locoregional therapies for patients awaiting transplant (eg Kyoto, Dallas, Pamplona, Asian, Edmonton, Shanghai, Hangzhou, and Chengdu) ( Table 2 ).
| Total Patients | Patients by Milan | Patients by Expanded | 1-y Survival by Milan (%) | 1-y Survival by Expanded (%) | 5-y Survival by Milan (%) | 5-y Survival by Expanded (%) | |
|---|---|---|---|---|---|---|---|
| Yao et al, 2002 | 70 | 46 | 24 | 91 | 71 | 72 | 57 |
| Leung et al, 2004 | 144 | 74 | 14 | 86 | NA | 51 | NA |
| Todo and Furukawa, 2004 | 316 | 138 | 171 | 81 | 75 | 78 | 60 |
| Decaens et al, 2006 | 479 | 279 | 188 | 80 | 78 | 60 | 46 |
| Onaca et al, 2007 | 1206 | 631 | 575 | 85 | 67 | 62 | 43 |
| Duffy et al, 2007 | 467 | 173 | 294 | 91 | 88 | 79 | 64 |
With more accumulated data for patients now being transplanted outside the Milan criteria, the overall 5-year survival for patients within and outside the Milan criteria do not seem comparable. This concept of the further the distance, the further the price has been named the Metroticket concept and shows that the farther patients stray from the Milan criteria after transplantation, the more their 5-year survival also decreases.
In the subset of patients with well-compensated cirrhosis and early stage HCC, transplantation still offers the best disease-free and overall survival compared with resection. Bellavance and colleagues, in a comparison of transplant and resection for early stage HCC, showed a 66% 5-year survival for transplantation versus 46% for matched resections. Despite the general overall better outcomes with transplantation, the scarcity of organs remains an issue. While patients are on the waiting list for transplants, greater than 20% of patients with HCC may progress or have other medical issues causing them to be removed from the list. The main risk factors associated with an increased likelihood of being removed from the transplant list include increased AFP, initial tumor size, multinodularity, and increased MELD. Strategies designed to improve access to transplants and decrease progression of disease include increasing the pool of donors through living donor liver transplantation (LDLT), treatment with locoregional therapy, or resection while on the waiting list.
There is a growing body of literature with inconsistent data regarding the role of LDLT for patients both inside and outside the Milan criteria. Outcomes for LDLT inside the Milan criteria are similar to deceased liver transplants. Although this may increase the donor liver availability, donor safety for LDLT remains a concern. Approximately 40% of donors experience a complication (most being minor) and mortality is 0.5% to 1%. When considering the usefulness of LDLT, Sarasin and colleagues showed that LDLT is cost-effective when the expected waiting time exceeds 7 months, similar to considering locoregional therapies.
Locoregional therapies can be used for bridging and downstaging. Bridging refers to the use of locoregional therapy for patients within transplant criteria who are awaiting transplant. In contrast, downstaging pertains to patients who are not within transplantable criteria; the locoregional therapy is designed to treat the tumor to an extent that patients are subsequently deemed appropriate for transplantation. To date, there are no prospective data to indicate that the use of locoregional therapies to bridge tumors within transplant criteria improves posttransplant survival or decreases dropout from the transplant list. There are also no data to suggest that any specific locoregional therapy (eg, ablation, TACE) has a preferential benefit compared with the others.
Despite the paucity of prospective data on bridging therapy, this approach makes some intuitive sense because data show tumor response and improved survival in nonsurgical patients treated with locoregional therapy. One study of bridge therapy in an intention-to-treat analysis suggested that patients in need of transplant derive a benefit when waiting times are longer than 6 months.
Several consensus guidelines for bridging and downstaging patients have been developed. These guidelines state that the type of treatment or therapy (eg, ablation vs TACE) should be based on the extent of disease. Ablation is more effective for tumors less than 5 cm, whereas intra-arterial therapy is more applicable for larger lesions.
The role of liver resection as a bridging therapy to transplantation is also controversial. The greatest advantage of liver resection as a bridge therapy is that transplantation may be avoided for select patients with small tumors and low MELD scores. Resection also offers accurate staging and insight into the biology of the tumor before potential transplantation. Nathan and colleagues showed the usefulness of pathology review of the tumor with regard to prognosis. Examining the primary tumors after resection, the investigators noted a variable 5-year survival of 29% to 55%, and tumor size, multifocality, and vascular invasion were independent predictors of survival. Although there may be benefits of bridging therapy with resection, there are downsides that include increasing the operative risk of eventual liver transplantation and potentially causing liver decompensation.
Locoregional therapies
There are circumstances in which the patient is not an appropriate operative candidate, or the degree of underlying liver disease precludes resection or transplantation. For these patients, locoregional therapies are available and include ablation (radiofrequency, microwave), percutaneous ethanol injection (PEI), transarterial radioembolization (TARE), drug-eluting beads (DEBs), and TACE.
Ablative Therapies
These treatment modalities work by destroying liver cells through direct chemical toxicity or by modifying neoplastic cell temperature with laser, microwaves, radiofrequency, or cryoablation. These ablative therapies can be approached through percutaneous, laparoscopic, or open laparotomies. Modalities in common use include PEI, radiofrequency ablation (RFA), microwave ablation, irreversible electroporation (IRE), and light-activated drug therapy. With PEI, the distribution of toxic ethanol may be blocked by tumor septae and capsulation, which tends to make ethanol injection less effective in tumors larger than 2 cm. In contrast, the thermal energy with RFA is not limited by tumor septae, and creates a necrotic rim around the tumor, potentially eliminating satellite lesions. Microwave ablation heats the water in tumor cells and induces cell death through coagulative necrosis. Irreversible electroporation is a newer ablative therapy. This treatment disrupts the cell membrane integrity by altering the cell membrane potential, directly causing cell death. Light-activated drug therapy works through creation of toxic oxygen singlets created with injection of talaporfin sodium. The talaporfin is injected intravenously, and light focused on the tumor focally activates the talaporfin, which destroys all cells in that field.
The data for these treatment outcomes have grown as the acceptance of this technology has increased. For tumors smaller than 3 cm, treatment with these modalities has resulted in 5-year survival approaching as high as 80%. For tumors between 3 cm and 5 cm, 5-year survival rates are reported to range from 40% to 70%. Several studies have prospectively compared PEI with RFA. Two meta-analyses have summarized these conclusions, finding that RFA had fewer treatment sessions, shorter hospitalizations, improved local recurrence, better tumor necrosis, and improved progression-free survival and overall survival.
There is growing evidence around the use of microwave ablation and IRE. Shibata and colleagues, in a randomized controlled study, compared RFA with microwave ablation and found no difference in complication rates or the incidence of developing residual disease. However, the follow-up was limited to 6 to 27 months, the assessment of residual disease was not standardized, and the microwave technology used was only first generation. To date, there are no published trials examining outcomes for IRE or light-activated drug therapies.
Current recommendations support using RFA or microwave ablation rather than PEI for HCC. RFA or microwave ablation should be specifically considered for patients with HCC less than 3 cm or for HCC less than 5 cm who are not candidates for resection or transplant. However, PEI should be considered when RFA or microwave ablation is not technically appropriate, as with lesions near the hepatic hilum when there is concern for damage secondary to peritumor necrosis. The application of newer modalities like IRE and light-activated drug therapy remains to be determined.
TACE
HCC can also be treated using a transarterial approach. TACE relies on the phenomenon that HCC is typically supplied by the hepatic arteries. Treatment strategies can exploit this finding. With a catheter-based approach, chemotherapy (mitomycin C, doxorubicin, cisplatin) mixed in a lipiodol emulsion can be injected directly into the artery feeding the tumor, and subsequently the hepatic artery terminating in the tumor can be embolized. The lipiodol creates an emulsion that helps retain chemotherapy within the tumor, maximizing exposure of the tumor to the chemotherapy and minimizing normal parenchymal toxicity. However, there are no strong clinical data to support that lipiodol improves delivery systems, and it may inhibit accurate CT assessment of tumor vascularity after treatment given its hyperdense appearance on CT.
The outcomes reported for TACE have been variable in several randomized controlled trials. In one of the earliest studies examining TACE, the Groupe d’Etude et de Traitement du Carcinome Hépatocellulaire reported that TACE for unresectable HCC resulted in reduced tumor growth but did not significantly improve survival. In contrast, Lo and colleagues reported improved tumor response and survival for patients with unresectable HCC, with an improved 3-year survival of 26% in the treated arm versus 3% in the control/untreated arm ( P = .002), with a relative risk (RR) of death of 0.49 (95% confidence interval, 0.29–0.81; P = .006).
One difficulty in evaluating the efficacy and standardization of TACE relates to assessing tumor response following therapy. Evaluating the response to locoregional therapies is critical to measuring the effects of treatment and applying that response to prognosis. Different criteria have been developed to objectively define tumor response. These assessments include the World Health Organization (WHO) and the Response Evaluation Criteria in Solid Tumors (RECIST) criteria, both of which evaluate unidimensional and bidimensional tumor measurements in response to therapy. In contrast, the European Association for the Study of the Liver (EASL) criteria examine the degree of necrosis in response to therapy. The modified RECIST (mRECIST) assesses not just size of the tumor but also the decrease in arterial enhancement of the targeted lesion ( Fig. 3 ). Even more recently, others have suggested the use of diffusion-weighted MRI. Notwithstanding the difficulty in assessing tumor response, 2 meta-analyses have investigated the efficacy of transarterial therapies in the treatment of unresectable HCC. Using RECIST criteria, complete and partial responses were noted among 35% of treated patients. With regard to survival, 2-year survival following TACE was 41% versus 27% in the non-TACE–treated group. The survival benefit of intra-arterial therapy was only noted in those studies that used chemotherapy embolization (TACE) and not those that simply used bland embolization (TAE).
