Biomedical Engineering Reference
In-Depth Information
In general, the validation study design for flow cytometric PK assays is similar to
other PK methods such as ELISA and mass spectrometry-based assays. It is
recommended that the following assay performance characteristic should be eval-
uated in validation studies: specificity/selectivity, precision (intra- and interassay),
accuracy, sensitivity [limit of detection (LOD) and limit of quantification (LOQ)],
linearity of sample dilution, validation sample stability (ambient temperature: short-
term 4
C and long-term
80
C), analysis of normal donor samples and
20 to
baseline samples from treatment-na
ı
ve patients (or animal subjects), and establish-
ment of control reference ranges.
Design of validation studies to assess assay performance characteristics described
under assay development will not be discussed in this section. The emphasis of this
section is to discuss challenges that are unique to flow cytometric methods.
11.3.1 Accuracy and Precision
It is not recommended to use cells from fresh blood as target cells for PK analysis.
Instead, a cell bank of frozen PBMCs from apheresis is preferred. If cells from fresh
blood have to be used as target cells, interday assays should be completed within 48 h
after blood is drawn, due to cell stability concerns. Evaluation of different anti-
coagulants and preservatives may allow extension of fresh blood stability beyond the
typical 48 h window. Another challenge is to fully examine normal donor-to-donor
variability. The use of fresh blood from a single donor for PK analysis over time is not
practical.
Accuracy and precision assessment for flow cytometric PK assay has unique
challenges, especially if drug binding to cells is not homogeneous. Assay accuracy
and precision may not meet the criteria established for other types of assays, recovery
80-120% and CV
20%, respectively. Under such circumstances, additional inter-
pretation may be required for data generated.
Generally, a flow cytometric assay is typically performed in a single tube or at most
in triplicates if 96-well plate format is chosen. Designing a study with a large number
of sample replicates may result in acceptable accuracy and precision, but assay
performance may not reflect actual in-study/clinical testing design. The appropriate
number of repeats for intra- and interassay studies should be similar to repeats
performed for clinical sample testing. Average of data collected from multiple runs
with a large number of data points prior to calculation of recovery may not be
appropriate since over- and under-recovery in each assay may cancel out, thereby
making the assay appear accurate. The use of the absolute percentage bias is an
alternative analysis to examine assay accuracy.
11.3.2 Limit of Detection
Limit of detection provides the minimal detection of signal that can be distinguished
from background. Typically, MFI median plus 2 or 3 SD for the blank serum sample
can be interpolated from the standard curve to generate a concentration value as LOD.
At assay LOD, the method may not be able to quantify drug with acceptable precision
Search WWH ::
Custom Search