We were therefore unable to show a relationship between the antibody method and scintigraphy results. Although HCCs do not typically show neuroendocrine differentiation on histological examination, measurement of serum chromogranin A levels was undertaken on the basis of a recent report that levels can be elevated in approximately one-third of cases (Leone et Olaparib mechanism al, 2002). It was proposed as both a potential prognostic marker and a marker of neuroendocrine differentiation. In our series, six of 59 patients had levels 10Ul?1 or less (reference range 2�C8Ul?1), 35 had levels of 10�C50Ul?1, and 20 had levels higher than 50Ul?1. Although most patients progressed on radiological criteria, chromogranin A levels decreased in 15 of 59 patients (25%) and remained stable in 25 (40%), increasing in a sustained manner in only two patients (3%).

We found no evidence that receptor expression or neuroendocrine features were associated with either scan positivity or clinical outcomes: chromogranin A therefore did not appear to be a useful marker for identifying a subset of tumours more likely to respond, or as a marker for monitoring clinical progress, with the proviso that the small sample size did not provide enough statistical power to detect other than large effects. Octreotide lowers portal pressure, and is routinely used to treat patients with variceal bleeding. Since prophylactic octreotide can prevent bleeding in high-risk patients (Jenkins et al, 1997), octreotide LAR might reduce morbidity and mortality through this mechanism.

In our trial, gastrointestinal bleeding occurred in four patients (one peptic ulcer, three varices) with a total follow-up approaching 4000 patient-months. It is impossible to determine the effect of octreotide LAR on the rate of complications from portal hypertension without a randomly allocated control group; however, the rate in our study seems low compared with other series (Akanuma et al, 2002). Although more patients rated improvements than deteriorations for some symptoms, these ratings were not strongly correlated with receptor status scores, and it is unclear whether these represent an effect of treatment. However, the fact that many patients reported improvements may partly explain the clinical impression of benefit in some patients. Determining the significance and cause of these changes requires a randomly allocated control group.

Our findings support the feasibility and importance of incorporating measures of HRQL in future trials in advanced HCC. This study has shown that octreotide LAR Brefeldin_A is well tolerated and may benefit some patients with HCC. We were unable to define, from octreotide scintigraphy or chromogranin A analyses, a subgroup more likely to benefit. There was little objective evidence of antitumour activity. A phase III study is advocated and indeed warranted given the conflicting findings with previous studies.