| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Department of Surgery, New York University, 1 Gustave L. Levy Place, Box 1104, New York, New York, 10029, United States
Correspondence: Address correspondence and reprint requests to: Myron Schwartz, MD; E-mail: myron.schwartz{at}mssm.edu
Hepatocellular carcinoma (HCC) is, in the large majority of cases, a secondary disease; cirrhosis as the result of long-standing chronic liver disease is the premalignant precursor condition.1 Cirrhosis is a diffuse process, and thus the entire liver is at risk for the development of HCC. Furthermore, the liver is the first site of metastasis of HCC as it progresses to the point where microscopic vascular invasion commences; this often occurs at a stage where the HCC is generally considered early from a clinical standpoint, typically in the 2–3 cm range.2 Not all deaths in patients with HCC are tumor-related; apart from leading to the development of cancer, progression of cirrhosis leading to end-stage liver failure is an important cause of mortality. On first consideration, liver transplantation (LT) seems the most rational treatment for HCC, because it both radically removes the primary tumor along with all micrometastases that may be present and treats the cirrhosis, thus eliminating the risk of new HCC’s and of liver failure.
Liver transplantation for patients with HCC within the Milan criteria (one nodule 
5 cm, two to three nodules all 3 cm) has been repeatedly shown to yield 5-year survival greater than 70%, with tumor recurrence developing in under 10% of cases; the most recent data from Milan demonstrate 5-year survival of 81% in a series of 173 patients.3 While the best reported 5-year survival results for resection of HCC may rival the results of LT, HCC recurrence within 5 years is observed in more than 50% of cases in even the best of series;4 whereas survival curves after LT show little fall-off between 5 and 10 years, postresection survival continues to decline.5 Recognition of the benefit of LT for early HCC has led to the adoption in the United States of an organ allocation policy granting waiting list priority to patients with early HCC.6
In proposing resection for HCC, the presumption is that the lesions identified preoperatively represent the entire extent of tumor present. This is often an incorrect presumption. The best data concerning the likelihood of unrecognized HCC in patients with a known tumor come from pathologic studies of livers removed at transplant. In patients thought to harbor a single HCC <5 cm based on presurgical imaging, 39% are found at pathology to have other foci of disease.7 Two studies examining the sensitivity of dual-phase helical computerized tomography and magnetic resonance imaging in identifying small HCC concurred,8,9 demonstrating that fewer than one-third of lesions <1 cm are identified preoperatively. Based on these data, it is clear that a significant proportion of postresection "recurrence" is in fact progression of established HCC’s that were present at the time of resection but were undetected on preoperative imaging.
The aim of hepatic resection for HCC is to achieve local control of the index tumor, accepting that new or unrecognized lesions will likely appear in the future. It is increasingly clear that local control of early HCC can be effectively achieved without surgery. For years there have been data demonstrating survival after percutaneous ethanol injection (PEI) of small HCC equaling that of resection.10,11
More recently, with the introduction of radiofrequency ablation (RFA), the evidence is indisputable that effective local control is possible to accomplish nonsurgically. In a randomized trial comparing PEI versus RFA for HCC 4 cm, RFA achieved initial complete response in 96% of nodules treated with subsequent local progression over 3 years noted in 14%, a figure considerably lower than the likelihood of the appearance of new lesions.12 More recently, in a similar trial comparing PEI versus RFA for HCC <3 cm, results of RFA were even better, with only 1.7% local failure over 4 years. In this study, for which the primary endpoint was survival, 4-year survival after RFA was 74%. Similar to resection, this survival was achieved despite a 63% incidence of recurrence elsewhere in the liver.13 As with resection, the results achieved with RFA are heavily dependent on the experience of the operator, in terms of both case selection and technical performance. While some Western series report rates of complete tumor necrosis following RFA as low as 20%,14 a series from UCLA based on explant pathology demonstrated complete necrosis in 83% of tumors <3 cm,15 suggesting that the results from Japan and Taiwan are indeed believable.
Metastasis of HCC is primarily via the hematogenous route. Vascular invasion occurs relatively early in the course of tumorigenesis: microscopic vascular invasion can be observed in 16% of HCC’s <2 cm in diameter, and among tumors 3–4 cm in diameter the observed rate rises to 46%. Nevertheless, the rate of HCC recurrence after transplant for tumors within the Milan criteria is very low, indicating that in its initial stages vascular invasion does not equate to extrahepatic spread. After resection in the setting of HCV cirrhosis, on the other hand, recurrence even in the best series reaches 70%,4 with the majority of recurrences due to intrahepatic metastases.16 The theoretical advantage of hepatic resection over RFA is limited to the removal of unrecognized metastases at a distance from the primary tumor, yet contained in the excised liver segment. Demonstration of an advantage of anatomic over nonanatomic resection supports this view; despite this advantage, however, recurrence is still noted after 64% of anatomic resections. With the low incidence of vascular invasion seen in HCC 2 cm, it is increasingly difficult in view of the excellent results of RFA to justify hepatectomy in patients with such tumors.
Upon reflection it is clear that the primary determinants of survival after initial treatment of HCC, whether by resection or ablation, are the quality of follow-up care in identifying and treating recurrence, and the progression of the underlying liver disease; the modality chosen to treat the index lesion is of secondary significance. Even accepting that there is a higher local failure rate with RFA versus resection, this is offset by the significantly greater morbidity and mortality associated with hepatectomy in cirrhotic patients compared to percutaneous treatment. When candidacy for resection is limited to patients with strict Child’s A cirrhosis and no significant portal hypertension, perioperative mortality in the 3–4% range is the norm;17 as criteria are broadened, results rapidly deteriorate.18 The patient experience, often glossed over in the literature, is not to be minimized; hepatic resection is a formidable ordeal for a cirrhotic patient, especially when compared to an ambulatory ablation procedure.
There are some HCC’s that are either not amenable to percutaneous treatment due to their location, typically high in the dome of the liver, or that, due to an exophytic growth pattern, would require direct puncture of the tumor without first traversing uninvolved parenchyma, in which case the risk of peritoneal or subcutaneous tumor dissemination is as high as 15%.19 The same issues that make these tumors unsuitable for percutaneous treatment in most instances render them particularly easy to resect, and a role for resection remains in such cases.
A strategy has been proposed employing resection as the primary treatment for early HCC in patients with preserved liver function, with LT held in reserve as a salvage procedure. Some authors have advocated observation after resection and listing for LT upon the appearance of recurrence,20,21 while others have suggested that if, upon review of pathology after resection, the presence of microscopic VI or satellites is observed the patient should be listed for LT in anticipation of likely recurrence.22 The validity of this strategy, since the large majority of patients undergoing resection will develop HCC recurrence, depends on the likelihood that patients who have undergone liver surgery and develop recurrence will be candidates for LT and, if they are, whether outcomes will be as good as in patients being transplanted primarily.
The most favorable report advocating salvage transplant concerned a patient cohort in whom the underlying liver disease was almost universally HBV.20 As these patients have a lower incidence of recurrence, and their livers are commonly precirrhotic, the strategy holds considerable appeal in this setting, and this author agrees that resection is the correct primary treatment for early HBV-related HCC with preserved liver function.23
In the context of HCV or alcoholic liver disease, however, where tumor recurrence is more frequent and liver disease more severe, salvage transplantation is more problematic. A study conducted in Paris showed significantly higher perioperative mortality in patients who underwent LT after prior hepatic resection,5 due to the added complexity inherent in operating on patients with portal hypertension and prior right upper quadrant surgery; adhesions become highly vascular, and the transplant can be quite bloody and hazardous, particularly after major hepatectomy. Furthermore, in that study tumor recurrence was also more frequent in patients with prior resection, likely due to the fact that the large majority of recurrences after resection are in fact intrahepatic metastases, and thus correlate with the risk of extrahepatic spread. Results from Mount Sinai confirm these findings: of 18 transplants performed after resection in our series, 2 resulted in intraoperative death (an extremely rare occurrence, <0.5% in our general experience), and 7/18 have developed recurrence of HCC.
The issue of resection as a bridge to LT has been studied from a cost-effectiveness standpoint. In an analysis utilizing a Markov model, resection was shown not to be cost-effective unless waiting times exceed 1 year, whereas percutaneous treatment of HCC was cost-effective in all scenarios analyzed.24
The question arises, whether resection or ablation has been employed as the initial tumor treatment, what to do with the patient in whom a complete response has been achieved. In the above-mentioned study advocating salvage transplant after resection,20 21% of patients with recurrent HCC had tumors that were beyond the Milan criteria when recurrence was first discovered. Similarly, we have often observed patients with apparently completely-ablated HCC who, despite close follow-up, present after a variable period with multifocal HCC or gross vascular invasion that eliminates them from consideration for LT. It can be a difficult decision to proceed with LT in a patient who has had an HCC successfully ablated and has normal imaging studies a year later when his priority score has risen to where transplant is immanent. There is room at this juncture for clinician judgment and patient preference. In younger patients with HCV-related HCC, the likelihood of living out a normal lifespan with their diseased liver is negligible and we favor proceeding with LT. On the other hand, in patients approaching age 70 we lean towards continued close observation; the risk–benefit equation is shifted away from LT, since the potential gain in long-term survival afforded by transplant is diminished relative to the up-front risk of the procedure.
There are two factors that mitigate the appeal that LT holds in the treatment of HCC: the shortage of donor livers, and the impact of recurrent HCV infection on posttransplant survival. The need to wait for a donor organ is associated with the risk of dropout from the waiting list: in the United States, if HCC progresses to beyond the Milan criteria, HCC-related priority is lost. For patients with solitary HCC 3 cm the risk of dropout over 12 months is low; however, patients with two to three nodules or with tumors >3 cm have a greater than 50% risk of dropout at 12 months.25 The priority accorded patients with HCC meeting the Milan criteria under the current US allocation system results in varying waiting times that are dependent on regional organ availability; in the state of New York, for example, despite priority the median waiting time approaches 1 year.
For patients with Child’s B or C cirrhosis for whom resection is not an option, the risk of dropout must be accepted. In weighing the transplant alternative for a patient with resectable HCC, on the other hand, transplant results must be viewed on an intention-to-treat basis, incorporating the risk of dropout into the risk–benefit equation.4 In the Mount Sinai experience, though posttransplant survival exceeds survival after resection for HCC, intention-to-treat survival at 5 years is actually worse with transplant than with resection.26
Recurrence of hepatitis C is universal in patients who are HCV-RNA-positive before transplant. By 5 years post-LT, 25% of HCV patients are once again cirrhotic, and post-LT survival is diminished in HCV-positive recipients compared to patients without viral hepatitis.27 For patients transplanted for HCV-related HCV meeting the Milan criteria, recurrent HCV is a far more common cause of death than recurrent HCC.
Living donor transplantation (LDT) has particular significance in the treatment of HCC.28 For patients with early HCC for whom a suitable donor is available, LDT enables timely transplantation, eliminating the risk of dropout from the waiting list due to tumor progression. Patients being transplanted for HCC typically, though cirrhotic, have better liver function than patients who require LT for end-stage liver failure, and are thus relatively better-able to tolerate relatively small-for-size living donor grafts. The best hope for patients with HCV-related HCC is to clear HCV on antiviral therapy before proceeding to LT: in the large majority of patients who go into LT HCV-RNA-negative, HCV recurrence post-LT is not observed. This strategy is facilitated by LDT in that transplant can be performed promptly once HCV-RNA has been cleared.29 In the Mount Sinai experience, over 40% of LDT’s have been for HCC.30
What, then, is the best treatment strategy for patients with early HCC in the setting of Child’s A cirrhosis and no portal hypertension? Although it was the task assigned to this author to advocate for LT as the treatment of choice, it is clear from the above discussion that there are no easy answers. Based on the considerations presented herein, I would propose the following:
2 cm, and for larger tumors depends on wide and preferably anatomic resection. Received for publication March 4, 2006. Accepted for publication April 18, 2006.
 |
REFERENCES |
|---|
 TOP REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
