Biomedical Engineering Reference
In-Depth Information
DOE can be used to design an appropriate interference study. Recovery of 80-120%
indicates lack of interference. Identification of potential interference factors that are
present due to the disease and specifically in patient's samples is important.
11.2.7.5 Accuracy and Precision
Accuracy is defined by the exactness (closeness)
of the test value determined by the assay to the reference value (true value).
Accuracy is assessed using validation samples (drugs at three-five different levels
across the standard curve linear range). Typically, each of the two analysts collects
data from replicates for each sample on a single day (intra-assay) or over multiple days
(interassay). An appropriate number of replicates selected should be relevant to
clinical sample testing replicates. The use of a high number of replicates may result in
acceptable accuracy in validation studies, which may not represent assay's intended
use in clinical sample testing typically performed in duplicates or triplicates.
Recovery (observed concentration/expected concentration) within 80-120% is con-
sidered acceptable, demonstrating assay is accurate.
Precision measures closeness of individual measures of an analyte when repeat-
edly assessed using multiple aliquots of the analyte in a single homogeneous volume
of biological matrix.
Precision is assessed using validation samples (drugs at three-five different levels)
as described in accuracy studies for repeatability (intra-assay precision) and inter-
mediate precision (interassay precision). As described in accuracy assessment, an
appropriate number of replicates should be relevant to the clinical sample testing
replicates. For each sample examined by each analyst, CV (SD/mean
100)
15% is
considered acceptable for intra-assay and CV
20% for interassay precision.
20% for
interassay and overall assay precision. Data from assays performed outside these
generally accepted measures of assay performance should be evaluated cautiously.
Precision is examined for both MFI median and interpolated mass values (concen-
tration such as ng/mL). An example for drug Y is shown in Table 11.1.
The use of statistical approaches such as balanced design and various component
analyses to design accuracy and precision studies is encouraged. Using such an
approach, each multiple analyst performs the same number of intraday runs and
repeats the intraday runs for the same number of days. Various component analyses
allow the assessment of variability due to intraday, interday, and/or analyses that
affect accuracy and/or precision.
Typically, CV
15% is considered acceptable for intra-assay and
11.2.7.6 Evaluation of Normal Sample and Treatment Na
ve (Baseline) Disease
Sample
For human PK assays, it is recommended that 30-50 normal serum samples
are examined. In general, normal donor samples are expected to be negative for the
presence of target drug. Potential interference factors such as endogenous counterpart
of the drug (e.g., cytokines or growth factors) are sometimes present in samples from
healthy subjects and treatment na
ı
ve patients. These interference factors, especially
those sharing similar characteristics with the target drug, may contribute to positive
signals. Understanding any baseline sample positivity is important for development
strategies to reduce interference in PK analysis.
ı
Search WWH ::
Custom Search