Credentialing biomarkers in early phase clinical trials

25 Nov 2009

Credentialing biomarkers in early phase clinical trials


The “Proof of Concept Clinical Trial Design” session of the AACR-EORTC-NCI conference was chaired by Jaap Verweij, Erasmus University Medical Center, Rotterdam, and focused on biomarkers and development of other molecular assays in early clinical trials.

Lillian L. Su (Princess Margaret Hospital, Toronto, ON, Canada) discussed about the role played by different biomarkers in early clinical trial design. Su started off her talk by pointing out the open issues that in her opinion are responsible for the current gap in translational research (Kola and Landis, Nat Rev 2004). This gap is particularly high in cancer drug development, where the average  rate of approval is only 5 % with a median average time of 14 years. The main reasons for attrition are lack of efficacy (about 30 % of cases) or toxicity issues (more than 30 % of cases). Too much money is being wasted on drugs which will fail for one of these two reasons late in Phase III trials. To reduce the attrition rate, Su pointed out possible directions to pursue, namely: a) a better understanding of oncogenic pathwas and their potential as targets for treatment; b) improved preclinical models which can better predict human response to therapeutic agents in terms of efficacy and toxicity; c) more efficient clinical trial design and method; and, d) more intelligent and coordinated biomarker research.

Su then proceeded to sketch out some possible strategies to reduce this gap, focusing also on the distinction between different kinds of biomarkers. Indeed, not all biomarkers are the same, and the word is often used too vaguely.

There are two main types of biomarkers, the former being the predictive ones, the latter being the pharmacodynamic ones, which are not necessarily predictive of prognosis or response to therapy.

Predictive biomarkers are the kind of biomarkers every investigator would like to develop. Ideally, they help match drugs with appropriate patients and predict outcome with specific therapy. They therefore guide the stratification of patients into respondents and non respondents on the basis of the presence or absence of the biomarker. A couple of examples of these desirable biomarkers were used in the IPASS trial stratifying patients with EGFR mutations in respondents or not to gefinitib (Mok et al, NEJM 2009) , and the Co.17 trial classifying patients in respondents or not to cetyximab on the basis of Kras mutation status (Karapetis et al, NEJM 2008).

The second category of biomarkers described by Su is the 'pharmacodynamic' type, which can be further subclassified in a) reflective of clinical outcome; and b) eflective of target inhibition.

Examples of biomarkers reflective of clinical outcome are very difficult, if not impossible, to retrieve in the literature, Su claimed. Two possible issues responsible for this difficulty are the following: a) surrogate tissues are often not reflective of what actually happens in the tumour, and b) it is often impractical to perform paired or multiple tumour biopsies in large clinical trials. The only example that Su could find of a marker belonging to this category was described by Cristofanilli and coauthors in circulating tumor cells in breast cancer (Cristofanilli et al, NEJM 2004).

Pharmacodynamic biomarkers belonging to the second category are reflective of target inhibition, but not necessarily of clinical outcome, for many possible reasons, among which the fact that  many pathways may affect clinical outcome, while the drug may target only one such pathway (Tabernero J et al, JCO 2008).

To sum up, strategies for more efficiently designing trials for predictive marker validation are urgently needed (Sargent et al, JCO 2005), together with more intelligent and better coordinated biomarker research. The desire to demonstrate clinical relevance must not come before solid scientific understanding of the pathway involved, and the use of the biomarker should truly be guiding  “go/no go” decisions concerning further development of the agent under scrutiny. To put it bluntly, as Su did, biomarkers should not be mere decorations for clinical trials in the hope of publishing one's paper in a journal with a higher impact factor. Which, by the way, may not necessarily be a better journal!






Cristofanilli M, Budd GT, Ellis MJ, et al. Circulating tumor cells, disease progression, and survival in metastatic breast cancer. N Engl J Med. 2004;351(8):781-91.

Karapetis CS, Khambata-Ford S, Jonker DJ, et al. K-ras mutations and benefit from cetuximab in advanced colorectal cancer. N Engl J Med. 2008 Oct 23;359(17):1757-65.

Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? Nat Rev Drug Discov. 2004;3(8):711-5.

Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med. 2009;361(10):947-57.

Sargent DJ, Conley BA, Allegra C, et al. Clinical trial designs for predictive marker validation in cancer treatment trials. J Clin Oncol. 2005;23(9):2020-7.

Tabernero J, Rojo F, Calvo E, et al. Dose- and schedule-dependent inhibition of the mammalian target of rapamycin pathway with everolimus: a phase I tumor pharmacodynamic study in patients with advanced solid tumors. J Clin Oncol. 2008;26(10):1603-10.

Tang PA, Siu LL, Chen EX, et al. Phase II study of ispinesib in recurrent or metastatic squamous cell carcinoma of the head and neck. Invest New Drugs. 2008 Jun;26(3):257-64.