Definition of biomarkers and efficacy end-points

3.1. Technical Challenges

• Biomarkers used in clinical trials must be validated (‘qualified’): is there scientific evidencethat the biomarker test is accurate, reliable, sensitive and specific enough? Companies will have to collect all this data for each biomarker they develop.
  
  
• Biomarkers must be valid measures: For example, if a biomarker is used to predict how severe a disease may get, is there enough evidence of this ‘predictive ability’? In other words, has it been shown to predict severity enough times before?
  
  
• Sample collection and processing methods must be consistent (e.g. for biopsies or blood samples). Variations in the way samples are handled can affect the results of biomarker tests. This could make the results of a trial invalid.
  
  
• The IT systems used for data management and data analysis must be robust and fast to cope with the amount of data generated. All biomarker measurements must be correctly linked with individual patients.

• When a companion diagnostic is needed to prescribe a new medicine , a new ‘platform’ or kit for testing patients in the clinic may need to be developed. The test kit will usually need to be ready for use during the large confirmatory trials of the medicine (Phase III). It is not ready until it has been validated and the validation results show that the diagnostic is accurate and clinically useful.