From The Stage
← Back to From The Stage
Dr. Afshin Safavi
Posted by Dr. Afshin Safavi BioAgilytix Insight, Pharmacokinetics

Preparing and Editing Calibration Curves in Quantitative Ligand Binding Assays: What are the Best Practices?

Preparing and Editing Calibration Curves in Quantitative Ligand Binding Assays: What are the Best Practices?

We all know that calibration curve quality is of critical importance in ligand binding assays (LBAs) and other quantitative methods, with the accuracy of reported sample results contingent upon the assay calibration curve’s robustness and reproducibility. While many of the requirements for calibration curves have already been defined in regulatory guidance and lead publications, there are still aspects that have yet to be established or adequately addressed: namely around editing specifications and preparation guidelines.

To help fill in these gaps, myself and my colleague Jeff Sailstad, BioAgilytix’s USA Chief Scientific Officer, recently participated in the development of a white paper aimed at presenting a collective view from members of the LBA community on best practices for the preparation and editing of calibration curves, with a focus on those used for quantitative pharmacokinetic (PK) LBAs.

The paper, “Calibration Curves in Quantitative Ligand Binding Assays: Recommendations and Best Practices for Preparation, Design, and Editing of Calibration Curves”, was recently published in The AAPS Journal and presents a consensus view of the contributing LBA experts on recommendations for the preparation of calibration curves, as well as for the treatment of calibrator data points. Below, I have summarized a few of the key takeaways from our discussions and the resulting best practices we define in the paper:

Calibration Curve Design
Regulatory agencies have established guidelines for calibration curve design, but in the white paper we lay out additional good practices and suggestions not specifically addressed in that guidance. For example, standard practice for the minimum number of calibrators is 1 + the number of unknown model parameters, but this approach does not account for assay variability. This is why the paper recommends a minimum of six calibrator points for a PL 4 curve.

Editing a Calibration Curve
Calibration curve editing involves masking or excluding a calibrator point from the standard curve regression while still keeping the calibrator available in the system. A calibration may only be edited due to an assignable cause; for example, a documented spiking or pipetting error. Editing must also be conducted independent of quality control (QC) assessment, and therefore calibrators should not be excluded to facilitate QC passing.

Although each individual laboratory must define their own approach to calibration curve editing as part of their standard operator procedures, the white paper outlines general guidelines for masking and excluding calibrators, including the point that after masking, high and low QCs should remain bracketed by valid calibrators or the assay fails.

Monitoring Calibration Drift
LBA calibration curves can be susceptible to calibration performance drift over time, due to changes in critical reagent characteristics, non-critical reagent performance, and/or reference material formulation, or introduction of new matrix pool batches. Because these changes could affect the slope or other properties of the standard curve, and potentially cause over- or under-reporting of sample concentrations, it is important to create a plan early in method development for tracking calibration drift from pre-study validation through multiple clinical studies.

While there is currently no established methodology for monitoring assay calibration performance, in the paper we provide a set of industry-accepted recommendations to evaluate assay drift through a study.

Growing Consensus, Standardizing Practices
We hope that the information compiled in this white paper will give the industry a new robust source to find scientifically sound approaches for developing calibration curves in quantitative LBAs that will eventually contribute to end-to-end method standardization. We want to see all laboratories able to perform optimal and fit-for-purpose evaluation of drug concentrations in non-clinical and clinical studies, and I see this paper being a key part of the next step forward toward that goal.

You can openly access the complete white paper at this link, and feel free to reach out to me if you have any questions or would like to continue this discussion.

Share This