EASL Clinical Practice Guidelines

Trial design

The increasing number of ongoing clinical trials in HCC has created the need for a common frame to test novel drugs accepted by all disciplines. As a consequence, new guidelines on the design of clinical trials and end-points in HCC have been reported by a multidisciplinary panel of experts [149]. These statements will evolve as new evidence becomes available, including more precise information on the natural history of HCC, new drugs or predictive biomarkers. The panel endorses the trial design and selection of end-points for clinical trials in HCC proposed in previous Journal National Cancer Institute guidelines [149]. In addition, the panel wants to emphasize that the integrity of research is absolutely vital for the advancement of evidence-based medicine. If a study proposing a major change in clinical practice is accepted yet its findings are fraudulent, as recently occurred in HCC with a study that required retraction [352], the threat to patient safety and management could be enormous.

The main recommendations for trial design are summarized below:

  1. End-points. Survival and time to recurrence were proposed as primary end-points for phase III studies assessing primary and adjuvant therapies, respectively. Composite end-points such as disease free survival (DFS) or progression free survival (PFS) are suboptimal in HCC research, and should be included as secondary end-points. Randomized phase II studies were considered pivotal prior to conducting phase III trials in HCC. These studies classically consider response rate as the gold-standard for efficacy, but time to progression was recommended as the primary end-point when testing molecular targeted therapies [149]. The panel considers that further data is needed to establish response rate as per mRECIST as surrogate of survival. Quality of life assessment in HCC research suffers from the lack of a reliable, standardized and adequately validated questionnaire, and thus, it is currently recommended as ancillary information.
  2. Trial design is summarized in Fig. 5. Selection of patients should be based on BCLC staging and Child–Pugh A class, in order to minimize the competitive risk of death associated with liver failure. The control arm for clinical trials should be the standard of care, meaning chemoembolization for intermediate HCCs and sorafenib for advanced cases. Therefore, for the assessment of first-line systemic treatments for advanced HCC a design adding a new agent to sorafenib versus sorafenib alone is recommended. Comparison of single agents head-to-head with the standard of care therapy might jeopardize the recruitment of patients for ethical reasons, unless the novel agent showed very promising efficacy in early phase II studies. For second-line treatments, the new agent should be randomized against placebo/best supportive care, and the selection criteria should include patients with contraindications or failures to sorafenib. Randomized studies testing molecular targeted therapies should optimally include biomarker analysis (tissue and/or serum samples) to enable the identification of molecular markers of response and for pharmacokinetic purposes, as reported in other cancers.

    Fig. 5 Summary of trial design strategies and control groups. Adapted from Llovet et al. [164].

    View Large Image | View Hi-Res Image | Download PowerPoint Slide
  3. Assessment of tumor response. The main end-point in cancer research is overall survival. Nonetheless, tumor response and time to progression have been considered pivotal for surrogate assessment of efficacy. In oncology, tumor response was initially measured according to the World Health Organization (WHO) criteria [353], and afterwards according to the Response Evaluation Criteria In Solid Tumors (RECIST) guidelines [[354], [355]]. These criteria were designed primarily for evaluation of cytotoxic agents. They do not address measures of antitumor activity other than tumor shrinkage. As acknowledged in the original RECIST publication, assessments based solely on changes in tumor size can be misleading when applied to other anticancer drugs, such as molecular targeted therapies, or other therapeutic interventions [354]. EASL and AASLD guidelines adopted a modified version of a WHO criterion in which the evaluation of the treatment response accounted for the induction of intra-tumoral necrotic areas in estimating the decrease in tumor load, and not just a reduction in overall tumor size [[1], [56]].

Results from a number of previous clinical studies in HCC have demonstrated that RECIST criteria do not mirror the extent of tumor necrosis induced by interventional therapies or new molecular targeting drugs [[168], [356]]. Viable tumor formation needs to be assessed using CT or MRI studies and viable tumor should be defined as uptake of contrast agent in the arterial phase of dynamic imaging studies. Consequently, a modification of the RECIST criteria was first proposed by a panel of experts [149], and further expanded [100]. This proposal is based on the fact that diameter of the target lesions with viable tumor should guide all measurements. In addition, specific modifications of the original criteria regarding assessment of vascular invasion, lymph nodes, ascites, pleural effusion and news lesions have been summarized in Table 5. Objective response rates using mRECIST have been reported of 57% in patients treated with chemoembolization [357], 90Y [358], ∼20% with sorafenib [[359], [360]] and 15–25% using brivanib [335]. The panel of experts recommends to assess tumor response according to mRECIST criteria, and to test whether these criteria have better performance than conventional RECIST, and correlate with pathological studies and outcome prediction (Table 6).


Table 5
Assessment of response comparing RECIST and mRECIST.*



RECIST, Response Evaluation Criteria In Solid Tumors; mRECIST, modified Response Evaluation Criteria In Solid Tumors; CR, complete response; PR, partial response; IR, incomplete response; SD, stable disease; PD, progressive disease.
∗Adapted from Llovet et al. [149] and Lencioni and Llovet [100].

Table 6
Unmet needs in HCC research.