Translating Knowledge into Guidelines and Practice

Chairs: Josep M. Llovet, MD (Spain/USA) and Morris Sherman MD, PhD (Canada)

  • 'Overview of the Role of Biomarkers in Guidelines in Oncology'
    Markus Peck-Radosavljevic, MD (Austria)

    A biomarker, a biological marker, in oncology can be any marker that can help to predict, detect, prognosticate, guide therapy, or predict recurrence of a cancer. A biomarker can be a protein, a gene, a microarray, a metabolic intermediate, anything that can help in selecting patients.

    There are different types of biomarkers: a prognostic biomarker is a test that is able to subdivide patients with a given disease into different prognostic groups. A predictive biomarker is able to predict which patients will respond to a certain treatment and which will not. A pharmacodynamic biomarker is able to give information about differences in drug metabolism of a certain drug. Even though there are a large number of studies ongoing to find biomarkers in particular in oncology, there are significant hurdles to the routine clinical use of biomarkers. First of all there are no clear regulatory proceedings on health to evaluate a complex genetic test for use in clinical routine, which should include analytical validity, clinical validity, and clinical utility. To cost of a thorough prospective validation of such biomarkers can be very high and also suitable tumour tissue that might be needed to establish a tissue biomarkers of a certain tumour is often hard to obtain.

    There are four major roadblocks to biomarker development:

    • The inability to mechanistically link a candidate biomarker to tumour biology
    • The lack of validation in well controlled human clinical study sets
    • Failure to plan the trial for the intended use of a biomarker with an appropriate control group
    • The liability and extremely lower abundance of cancer biomarkers in peripheral blood

    There are already several tumours, where biomarkers have been integrated into the management guidelines. The most notable example is breast cancer, where treatment is primarily guided according to the expression of estrogen receptors. Estrogen receptor expressing tumours are treated with endocrine therapy, while negative tumours are not. And in addition to that HER-2 /neu receptor expression is evaluated and HER-2 positive tumours are treated with trastuzumab.

    But biomarkers are also used to predict prognoses of patients with breast cancer. Notably, histological grade and estrogen receptor expression can be integrated with other markers into prognostic models and molecular predictors of prognosis like the Amsterdam 70-gene signature, Mammaprint, or the 16-gene recurrence score may out performed traditional prognostic markers in some patients’ populations

    But biomarkers are also used to predict prognoses of patients with breast cancer. Notably, histological grade and estrogen receptor expression can be integrated with other markers into prognostic models and molecular predictors of prognosis like the Amsterdam 70-gene signature, Mammaprint, or the 16-gene recurrence score may out performed traditional prognostic markers in some patients’ populations.

    But biomarkers are also used to predict prognoses of patients with breast cancer. Notably, histological grade and estrogen receptor expression can be integrated with other markers into prognostic models and molecular predictors of prognosis like the Amsterdam 70-gene signature, Mammaprint, or the 16-gene recurrence score may out performed traditional prognostic markers in some patients’ populations.

    But biomarkers are also used to predict prognoses of patients with breast cancer. Notably, histological grade and estrogen receptor expression can be integrated with other markers into prognostic models and molecular predictors of prognosis like the Amsterdam 70-gene signature, Mammaprint, or the 16-gene recurrence score may out performed traditional prognostic markers in some patients’ populations.

    In summary, biomarkers are an indispensable tool for the advancement of personalised medicine in oncology. They can be prognostic, predictive, or pharmacodynamic. The development of suitable biomarkers is not easy and faces several hurdles. For that reason, complex genomic biomarkers have only been established for breast cancer so far.

  • Criteria for Accepting Subclasses/Biomarkers in HCC Guidelines
    Josep M. Llovet, MD (Spain/USA)

    Several molecular subclasses of hepatocellular carcinoma (HCC) have been proposed during the last years. At least 3 groups agreed on the fact that Wnt-CTNNB1 subclass and proliferation subclass (related or non with progenitor- cells) represent common biological themes in HCC classification. Similarly, several biomarkers (mir 26, AFP, Ang2) or gene signatures from tumour (5-gene, EpCAM) or adjacent tissue (poor prognosis signature) have been reported as predictors of survival. The current EASL-EORTC guidelines for management of HCC (J Hepatol 2012) established a modification of previosuly reported criteria (SImon, JNCI 2009) to adopt novel prognostic markers in HCC:

    1. The gene-signature or marker should retain prognostic value in uni/multivariate analysis, when tested along with known clinical/pathological variables.
    2. The signature needs to be supported by a properly powered training/validation set testing approach.
    3. The signature needs to be confirmed by an independent cohort of samples from an independent group.

    Up to now, two signatures, the 5-gene signature from the tumour (Nault, Gastroenterology 2013) and the poor- porgnosis signature from the adjacent tissue (Hoshida, NEJM 2008) fulfill these criteria and can be considered as bonafide prognostic biomarkers. Similarly AFP levels of 400ng/mL and mi26 as isolated markers retain prognostic capacity according to the criteria depicted above. Thus the next ILCA-EASL Guidelines on biomarkers and HCC molecular subclasses should take these criteria into consideration for being adopted in clincial practice.

  • Tissue Biomarkers Predictors of Response in Oncology
    Richard Finn, MD (USA)

    The approval of sorafenib has validated the concept that molecular targeted therapies can be active in the treatment of advanced HCC. While sorafenib has provided a large step forward, there is still a great need for new treatments for the disease. Over the past 6 years we have seen several negative Phase III studies in advanced HCC and there are several ongoing. Still, there is even a larger number of agents being pursued in single-arm and randomised Phase II studies. Many of these studies are based on some scientific rational to pursue any number of molecular targets. The challenge for us in the future is to understand the molecular drivers in HCC and use this information to better select patients for clinical trials. To date, there have been a number of informative molecular profiling studies done in HCC tissue, yet the true translation of this data to clinical trial design is lacking. Unless this is done, we are sure to see additional disappointing results in the future. To date, the largest impact of new agents in solid tumour oncology has come from the introduction of predictive markers of response into clinical trials. If we look to other malignancies as a model of drug development, we can better understand how the introduction of molecular therapeutics in HCC can have a similar impact. In this presentation we will review the successful clinical development of several targeted agents in other tumour types and explore how these approaches can be applied to HCC trials in an attempt to avoid future failures.