IMMUNOGENICITY TESTING USING FLOW CYTOMETRY - Flow Cytometry in Drug Discovery and Development - page 212

instruments in the middle of the study influence the signal to the degree that
comparisons cannot be made longitudinally, and the signal differences cannot be
overcome by normalization, some investigators choose to dedicate specific instru-
ments to individual studies.

10.3.2 Determining Neutralization Readout

Signal output from the flow cytometer can be reported as median fluorescent intensity
(MFI) or GMFI, with either output being used to subsequently calculate percent
inhibition of binding or percent of maximal signal.
Reported values from immunogenicity assays are qualitative and, at best, semi-
quantitative for reasons that are well described in the literature [20]. The assay cut
point value, which represents the neutralization capacity of 95% of the normal
population, is used to score samples as either positive or negative.
As cut point values are critical to defining the neutralizing status of an antibody
positive sample, it is essential that they are determined appropriately. Ideally, cut point
values are determined using sera collected from the untreated patient population. If
these samples are not available at the time of assay development and validation,
normal donor sera may be used and the assay cut point confirmed or re-established at a
later date using patient sample data.
To establish the cut point, 30-50 samples, inclusive of both male and female, are
analyzed using the recommended assay dilution scheme. A total of three independent
runs by at least two analysts are conducted. The data are tested for normal distribution
and transformed if necessary. The cut point is then set at the 95% confidence interval
or mean

1.645 SD. Aggregate data from the different dilutions may be used as
long as lack of a difference between them has been demonstrated statistically.
Various strategies can be taken to assess the presence of neutralizing antibody. For
example, to obtain a qualitative (positive/negative) result, MFI or GMFI signal can be
compared against an assay cut point that is calculated from the output signal of drug-
na

/

þ

ve or normal individuals. Alternatively, to obtain quasi-quantitative results,
percent inhibition can be calculated as [1 - (sample signal/max signal)]

ı

100.
Background signal may be subtracted when appropriate. Another approach to
obtaining a quasi-quantitative result is to compute percent of maximal signal or
ratio of signal/maximal signal that is defined as (sample signal/max signal)

100.
Again background signal may be subtracted when appropriate. Regardless of the
computation approach that is used, an assay cut point must be used for scoring
patient samples.

10.3.3 Software

When testing samples from nonclinical and clinical studies, the ability to use a
multifunctional data analysis software package that facilitates batch analysis, allows
template-driven testing, and standardizes repetitive functions is advantageous. Two
examples of such software are FlowJo and FCS Express. Other data analysis software