REGULATORY COMPLIANCE AND METHOD VALIDATION - Flow Cytometry in Drug Discovery and Development - page 249

subsets are at least 10% of the parental population, precision of 20%CV may be

attainable. However, when measuring rare cell populations (less than 10% of
parental population), 30%CV may be more likely. When assay imprecision is
greater than expected, and the lack of reliability does not exclude the analysis from
the clinical plan, analysis of samples in replicates is advised. Evenwhen samples are
assayed in replicates, clinical teams must be made aware of the liability imposed by
intra-assay variability when interpreting assay results in their trials. Perhaps, the
only acceptable place where assay imprecision can be discounted is in the case of
cell-depleting agents because intra-assay variability approaches zero in samples
devoid of the population of interest.

<

INTER - ASSAY PRECISION Inter-assay precision is an assessment of between-run
variability. Multiple experimental runs are performed using the same source
samples; therefore, the measured parameter must be stable enough to allow
analysis at different times while generating the same results. Consideration of cell
population frequency and sample stability is important in determining acceptance
criteria for inter-assay precision. Whole blood samples with limited postcollection
stability present challenges for conducting inter-assay precision assessment where all
analytical runs must be assayed within the limited sample stability window, usually
1-2 days. Multiple runs must be performed on the same day and can lead to long hours
in the laboratory. Frozen PBMC, however, usually have longer storage stability and,
therefore, PBMC assays offer more flexibility in sample management performing
multiple analytical runs for inter-assay precision assessment. When two or more
analysts participate in this stage of assay validation, care should be taken to ensure all
analysts perform analysis comparably. To reduce the analyst-to-analyst variability, all
analysts should be formally trained in assay performance, and training results be
documented with acceptance criteria. Whenever more than one analyst is performing
an assay for precision assessment or training, although aliquots of the same sample
will be used, special consideration should be made to ensure they are truly
“independent” analytical runs.

INTER - ANALYST COMPARISON Operator training is a key component in obtaining high-
quality data. Visual observation followed by practice is recommended to obtain
reproducible assay results. A training plan with documented performance and
acceptance criteria upon assay transfer can be valuable to successful analysis no
matter where it is applied in the drug development process. Precision validation
assessment performed by one analyst is the best-case scenario for reproducibility in
anassaywhentransitionismadetotheclinic. Although one analyst may primarily
be responsible for sample analysis, it is important to have more than one analyst
trainedonproperlyconductingtheassayasabackup.FrozenPBMC-basedassays,
in many cases, can be scheduled and batched, facilitating the one-analyst scenario.
For whole blood assays that have limited sample stability, it is more likely that a
second analyst will be required to perform sample analysis. In addition, a primary
analyst who is responsible for assay validation may not be the primary analyst for
sample analysis; therefore, training and documentation of meeting acceptance