New systemic approaches and trial design


Session type:

Philip Johnson

University of Birmingham, UK


New systemic approaches and trial design

Conventional systemic cytotoxic chemotherapy has had only marginal impact on the natural history of advanced hepatocellular carcinoma (HCC). The most widely used conventional cytotoxic agent has been Doxorubicin (Dox) as a single agent or in combination, where response rates of 10-25% have been consistently reported. An early randomised trial against best supportive therapy showed significantly increased survival but this was only in the order of weeks. A more recent study where Dox was the control arm in a randomised Phase III trial against Nolatrexed, showed a significant survival benefit in favour of the Dox arm. More recently the multikinase inhibitor Sorafenib (Sor) has shown a significant improvement in survival in two large prospective randomised trials and the combination of Dox and Sor has suggested improvement in survival compared to Dox alone, in a smaller randomised phase II study. The integration of effective systemic therapies with effective locoregional treatments is now a major challenge as is the strategy for combining Sorafenib with other systemic agents.

Phase I studies may follow standard lines except that an HCC specific trial is often necessary in view of the possibility of toxicity and aberrant pharmacokinetics associated with the underlying chronic liver disease. Phase III trials of new agents do not present specific challenges since the primary endpoint of overall survival is now widely accepted as is the control arm of Sorafenib, although ‘crossover’ may cause problems. However, the major difficulty arises in phase II trials that aim to determine if an agent should proceed to phase III. These have usually used a single arm approach according to the design of Simon or Gehan. However, the conventional endpoints of some change in the physical size of the tumour are less appropriate for HCC as it has been shown that significant tumour necrosis can occur without corresponding decrease in tumour size and that significant prolongation can be realised without patients achieving response according to conventional criteria. There is thus ample scope for discarding active agents. For this reason some measure of survival, either ‘Time To Progression’ or ‘Progression Free Survival’, are being increasingly used. However such endpoints depend upon a precise understanding of the natural history of the (untreated) disease for rational interpretation and because of the underlying cirrhosis and several other co-morbidities this is difficult to establish in HCC. The randomised phase II trial design overcomes some of these problems in that it includes an unbiased control arm against which to compare the experimental arm and thus excludes the need to compare against a potentially biased historical benchmark.