Validation of biomarkers2018-11-20T14:00:22+00:00

Searching for human tissues in bio-banks


The validation of a biomarker requires various degrees of validation depending on its use or its phase of development: proof of concept / relevance of the biomarker, experimental validation of the biomarker, validation of the analytical performances, and standardization of the protocol.

Proof of concept / relevance of the biomarker :

This phase consists in validating In-Situ the benefit of the previously selected biomarker. A small cohort of characterized human tissues is used. The results sometimes lead to combining the use of several biomarkers for improving the relevance of the results.

Experimental validation of the biomarker(s):

This part makes it possible to confirm the relevance of the labelling protocol selected with a larger collection of human tissues (covering all patient strata). Here we can evaluate more powerfully the benefit of the method in order to develop a routine test.

Validation of analytical performances :

After the identification of one or several biomarkers, a labelling protocol must be designed, validated and eventually standardized. These important steps require the consideration of several parameters which are:

  • specificity
  • robustness
  • sensitivity
  • limits of detection / limits of quantification
  • accuracy
  • repeatability, and
  • reproducibility (multiple site testing on different automated systems, different teams, etc.)

Standardization of the protocol:

The aim of this phase is to optimize, and then to validate a method for multicentric use. Here, it is imperative to carry out a detailed analysis of the critical points by scanning the entire analytical procedure so as to identify and control the potential risks. Several parameters must be tested and specified:

This step requires:

  • precisely specifying the method of specimen collection
  • defining the criteria of acceptability of the specimens
  • describing the conditions for storage of the samples
  • identifying all the factors liable to influence the analysis results
  • developing reference guidelines for the routine use of the method

It is also possible to design custom made tools:

  • web interface for training and qualification of skills
  • Tool for image analysis for facilitating a diagnosis
  • etc.