• HEPATOCELLULAR CARCINOMA

Individuals with cirrhosis are at greatly increased risk of death from both liver failure and hepatocellular carcinoma (HCC). In this context, many specialist societies recommend using ultrasound surveillance to detect HCC early with the aim to reduce mortality from liver disease and not only from HCC.1 ,2 This is a laudable aim but the studies defining the efficacy of surveillance are generally of low quality and there is no readily applicable randomised controlled trial of surveillance in a western population. Recent evidence syntheses are conflicting and while meta-analysis of non-randomised case–control studies suggests benefits of surveillance this cannot be applied to patients with cirrhosis in general.3 ,4

The study from Cross and coworkers in this issue of Frontline Gastroenterology used an online questionnaire to examine the provision for HCC surveillance in acute hospital NHS Trusts across the UK.5 The vast majority of respondents indicated that their Trust provided a surveillance programme for HCC. In the majority of cases however it is done on an ad hoc basis without automatic recall strategies or indeed regular 6-monthly imaging in as many as 40% of cases. It is apparent that in many centres surveillance is being done in ways that are likely to substantially reduce its effectiveness. In addition, there will inevitably be responder bias in the survey and the results reported are likely to be more favourable than the overall provision on the ground.

The authors conclude that the provision of surveillance is poor and the UK is not alone in that respect.6–8 The question then is why surveillance is not done, or done poorly, in many centres. The questionnaire attempted to determine possible barriers to the provision of effective surveillance and several themes were identified: cost, difficulty accessing radiology services, and ‘doubts over effectiveness’. While several studies have indicated that surveillance is likely to be cost-effective9 ,10 there are a number of recent developments including the introduction of sorafenib treatment that are not considered and that may have significant effects on the cost-effectiveness estimates. The issues of access to radiology services and perceived effectiveness of surveillance ultimately are due to the lack of high quality evidence supporting surveillance derived from randomised controlled trials. The view of the Royal College of Radiology is that the evidence to support surveillance is lacking and any decisions regarding surveillance and the funding for this need to be made at a local level.11 They also state that patients should be counselled as to the limitations of ultrasound and the risks of false positive and false negative examinations. These data are not readily available and according to the survey from Cross and coworkers 30% of patients receive neither written nor oral information regarding the rationale for surveillance, much less detailed information about the sensitivity and specificity of ultrasound to detect small HCC.

The information required for patients considering the benefits and harms of surveillance and for those clinicians doubting its effectiveness would best come from a well-designed randomised clinical trial. There are many arguments against such a study12 and for one to be done successfully it would require engagement from all stakeholders. At present surveillance is done poorly and is likely ineffective at a population level. If such a randomised trial were to start those patients allocated to receive surveillance would do so with appropriate consent, with scans done by sonographers and radiologists with expertise in liver disease and the diagnosis of HCC, and with appropriate recall strategies in place so that any benefits of surveillance were accurately determined. That this would in all probability substantially increase the number of patients receiving effective surveillance provides the most compelling argument of all to initiate a trial of surveillance for HCC in the UK.