Hepatocellular carcinoma: beyond screening
Article Outline
Surveillance for hepatocellular carcinoma (HCC) has become a routine part of the management of patients with cirrhosis. This is despite the fact that there is little data to support routine surveillance for HCC. There is no evidence yet that surveillance will reduce mortality from hepatocellular carcinoma. In addition, there is little agreement on many aspects of surveillance. For example, although studies have shown that alphafetoprotein is a poor screening test [1], [2], it remains in common use. Other serological tests have been described [3], [4], [5], [6], but studies are incomplete and these tests have not gained widespread acceptance. There is little agreement on what constitutes an abnormal screening result, and there is no data to indicate the efficacy or cost-efficacy of alternate methods of investigating to confirm or refute the diagnosis. Even the ideal screening interval is not known.
As generally understood, the term ‘surveillance’ refers to the process of subjecting patients to regular ultrasonography and/or regular testing with serological screening tests. However, surveillance involves much more. After detection of an abnormality there is the recall procedure. This includes a definition of what severity or type of abnormality constitutes a result warranting further intervention, and what the intervention should be. If an HCC is identified, the next step would be therapy, but more often, the results of investigations are indeterminate. Since there is always a possibility of a false negative result, HCC can never be completely ruled out by a single test, and often, not even by the combined results of several different tests. For example, a nodule between 1 and 2 cm on ultrasound, even if not shown to be vascular on computed tomography (CT) scan may still be HCC. Patients with these indeterminate results are entered in a phase of more frequent and more intense evaluation.
For the purposes of discussion, we can label the phases of surveillance as ‘screening’, ‘recall’, and ‘enhanced follow-up’. Screening is applied to patients, who although at-risk for HCC, do not have any indication that HCC is present. Recall procedures include all the tests performed to determine whether the abnormal screening result is due to HCC. Enhanced follow-up is the series of steps taken after failure to document HCC following an abnormal screening result. The major characteristics of enhanced follow-up as different from screening is an increased frequency of radiological examinations, with CT scanning and magnetic resonance imaging (MRI) being the radiological investigations of choice, whereas ultrasonography is the usual examination of choice for screening.
Each of these phases need to be studied separately and need to be optimized before we can make any assessment of whether overall surveillance will reduce HCC mortality or not. At present, most of the literature available either confuses the separate processes, or investigates only the screening phase. However, we are starting to see studies looking at the recall phase, one of which, by Borzio et al. [7], is published in this issue.
Borzio et al. attempt to address the question of the value of liver biopsy of small nodules found on ultrasound screening of cirrhotic patients. This is a retrospective study. CT scan results are not given, we are told, because too few had high quality triphasic CT scans. Thus, we cannot determine whether a CT scan would have any influence on the decision to perform a biopsy. Nonetheless, they performed a series of biopsies in the patients and reported only on those nodules which turned out not be HCC. We are not told what proportion of all biopsied nodules were benign. Nor do we know whether they biopsied all nodules found on ultrasound, or only those in which a diagnosis of HCC could not be made with certainty on other grounds.
In the study they were careful to exclude all malignant lesions, and lesions that became malignant early in the course of follow-up and were possibly misclassified initially. Nodules that were negative for HCC were classified histologically according to what is now standard nomenclature to describe such lesions [8]. These are large regenerative nodules (LRN), high-grade dysplastic nodules (HGDN) and low-grade dysplastic nodules (LGDN). Large regenerative nodules are just cirrhotic macronodules, previously called typical adenomatous hyperplasia (a nomenclature that is now defunct). They also biopsied the surrounding liver, and identified whether large or small cell change was present or not. Although prospective validation is still lacking it is generally agreed that HGDN are malignant precursors. Extranodular large cell change is also believed to be a predictor of malignancy, although whether the lesion is itself a malignant precursor is not clear.
As expected, the presence of HGDN and large cell change predicted the development of HCC. However, only 21% of nodules were HGDN. Furthermore, almost as many nodules (17%) regressed on follow-up as were HGDN.
This study raises the question of whether all patients with nodules in the liver should be biopsied. There are no studies which compare outcome in patients who undergo biopsy vs. those who do not. However, when the nodule is larger than 2–3 cm, and occurs in the setting of chronic hepatitis B or C and cirrhosis, the accuracy of clinical and radiological diagnosis is sufficiently high that biopsy is not necessary [9], [10]. At the opposite end of the scale, approximately half of all nodules smaller than 1 cm turn out not to be HCC [11]. Based on these and other considerations, the EASL-sponsored consensus conference on HCC [12] proposed that lesions smaller than 1 cm should not be biopsied but followed with serial radiological testing. Lesions larger than 2 cm in the correct clinical setting should also not be biopsied, particularly if surgery and transplantation are being considered. Not only is biopsy unnecessary, but there is a risk of up to 5% of needle track seeding, which would convert a curable situation into an incurable one. Biopsy was recommended only for lesions between 1 and 2 cm in diameter. In their series, Borzio et al. found that 73% of the nodules they biopsied fell into this size range.
One factor that might influence the decision to biopsy is the appearance of the lesion on CT scanning or MRI. This was not assessed in the current study. However, it may be that as long as the CT scan or MRI confirms that the lesion is vascular (and not hemangioma) biopsy should still be performed for lesions between 1 and 2 cm. Certainly, on MRI up to 50% of vascular lesions below 2 cm in size become smaller or disappear, and about 20% remain stable over at least 12 months [13]. These are probably also not HCC. Non-vascular lesions are also unlikely to be HCC.
A second issue raised by this study is whether knowing the nature of a small nodule will affect how patients are managed subsequently. Clearly, those with HCC will be treated appropriately. The question is how those with non-malignant lesions should be managed. Over a 5-year follow-up 31% of all nodules showed malignant transformation. HCC arose in LRN (21% transformed), LGDN (25% transformed) and HGDN (63% transformed). HCC also developed in extra-nodular sites in a further 9% of subjects. However, the rate at which neoplastic transformation occurred in nodules, at least over the 3 years (15.5%), was not much different than the rate expected for development of HCC in a cirrhotic hepatitis C population in general. Should these patients undergoing enhanced follow-up, or would the outcome be similar if they simply continued with screening? Only future studies will tell. This also begs the question of how long should enhanced follow be continued. Some patients only developed HCC 5 years after initial biopsy. Enhanced follow-up over such a long period may be very expensive.
The authors suggest that patients at highest risk, namely those with high-grade dysplastic nodules and large cell change in the adjacent liver might be considered for liver transplantation to prevent the development of cancer. This is a somewhat drastic step in patients with preserved liver function, particularly since the outcome of transplantation remains good as long as the HCC remains under about 3 cm in diameter.
Several issues flow from this study. First, the study confirms the EASL consensus conference recommendations that biopsy is necessary in lesions between 1 and 2 cm in diameter. One may argue that since follow-up decisions might not be altered by the biopsy result, it should not be done. The counter argument is that until these lesions are biopsied in large numbers we will not be able to find out whether treatment decisions should depend on the biopsy results or not.
The second issue is whether studies of ethanol injection or other forms of local ablation that did not include biopsy results need to be redone. For example, Livraghi et al. [13] have described a large series of patients who underwent alcohol ablation for small HCC's. In their series most tumors were smaller than 3 cm. No biopsies were taken. Thus it is possible that at least some of the lesions treated were in fact not malignant. If so, the survival rates are probably overstated.
In summary, the study by Borzio et al. deals with that component of the surveillance process that I have labeled as ‘recall’. It is an attempt to minimize the uncertainty about diagnosis by performing a biopsy of small nodules seen on ultrasound. This study also represent the first step in characterizing these small lesions, raising the possibility of tailoring follow-up to the findings, an investigation which is overdue.
References
- Serum alpha-fetoprotein for diagnosis of hepatocellular carcinoma in patients with chronic liver disease: influence of HBsAg and anti-HCV status. J Hepatol. 2001;34:570–575
- . Screening for hepatocellular carcinoma in chronic carriers of Hepatitis B virus: incidence and prevalence of hepatocellular carcinoma in a North American urban population. Hepatology. 1995;22:432–438
- Clinical utility of Lens culinaris agglutinin-reactive alpha-fetoprotein in small hepatocellular carcinoma: special reference to imaging diagnosis. J Hepatol. 1999;30:125–130
- Serum alpha-l-fucosidase activity and early detection of hepatocellular carcinoma: a prospective study of patients with cirrhosis. Cancer. 1998;83:2468–2474
- . Early detection of hepatocellular carcinoma associated with cirrhosis by combined assay of des-gamma-carboxy prothrombin and alpha-fetoprotein: a prospective study. Hepatogastroenterology. 1995;42:387–393
- Impact of large regenerative, low grade and high grade dysplastic nodules in hepatocellular carcinoma development. J Hepatol. 2003;39:208–214
- . Terminology of nodular lesions of the liver. Hepatology. 1995;22:983–993
- Accurate pre-operative evaluation of liver mass lesions without fine needle biopsy. Hepatology. 1999;30:8889–8893
- . Resection of hepatocellular carcinoma without preoperative tumor biopsy. Ann Surg. 2001;234:206–209
- . Hepatocellular carcinoma. Tokyo: Springer-Verlag; 1987;
- Clinical management of hepatocellular carcinoma: conclusions of he Barcelona-2000 EASL conference. J Hepatol. 2001;35:421–430
- . Cirrhosis or chronic hepatitis: evaluation of small (<2 cm) early-enhancing hepatic lesions with serial contrast-enhanced dynamic MR imaging. Radiology. 2003;226:550–555
- Hepatocellular carcinoma and cirrhosis in 746 patients: long term results of percutaneous ethanol injection. Radiology. 1995;197:101–108
PII: S0168-8278(03)00249-6
doi:10.1016/S0168-8278(03)00249-6
© 2003 European Association for the Study of the Liver. Published by Elsevier Inc. All rights reserved.
