Objectives

Consistency of procedures for the evaluation of a predictive biomarker (including sample collection, processing, assay and scoring system) based on adequate evidence is necessary to implement research findings in clinical practice. As a case study we evaluated how a particular predictive biomarker, ERCC1, was assessed in research on platinum-based chemotherapy in non-small-cell lung cancer and what motivated the choice of procedure.

Materials and methods

A systematic review of studies completed since 2007 and ongoing was undertaken. Questionnaires on details of ERCC1 evaluation procedures and the rationale for their choice were sent to contacts of identified studies.

Results

Thirty-three studies of platinum-based chemotherapy in non-small-cell lung cancer using ERCC1 were identified. A reply to the questionnaire was received for 16 studies. Procedures for ERCC1 evaluation varied substantially and included reverse transcriptase quantitative polymerase chain reaction (nine studies), immunohistochemistry (five studies) and other methods (multiple methods–two studies, NER polymorphism–one study). In five studies ERCC1 use was planned, but not undertaken. In nine data was insufficient to identify the procedure. For each assay there was variation across studies in the details of the laboratory techniques, scoring systems and methods for obtaining samples.

Conclusions

We found large variation across studies in ERCC1 evaluation procedures. This will limit the future comparability of results between these different studies. To enable evidence-based clinical practice, consensus is needed on a validated procedure to assess a predictive biomarker in the early phase of research. We believe that ERCC1 is not untypical of biomarkers being investigated for stratified medicine.