Predicting inhibitor development using a random peptide phage-display library approach in the SIPPET cohort

Area of Science:

• Immunology
• Hematology
• Bioinformatics

Background:

• Inhibitor development is a major complication in hemophilia A (HA) treatment.
• This complication increases morbidity and mortality in patients.
• Predicting inhibitor development is essential for personalized HA care.

Purpose of the Study:

• To explore the factor VIII (FVIII)-specific epitope profile in severe HA patients.
• To develop an inhibitor prediction model based on epitope mapping.
• To assess the predictive performance of developed models.

Main Methods:

• Utilized a novel immunoglobulin G epitope mapping method.
• Employed a random peptide phage-display assay to assess FVIII epitope repertoire.
• Developed and validated LASSO and random forest regression models.

Main Results:

• Identified 27,775 peptides putatively directed against FVIII.
• Achieved good predictive performance with C-statistics of 0.78 (LASSO) and 0.80 (random forest).
• Demonstrated moderately good model calibration for both statistical models.

Conclusions:

• Developed two statistical models using a novel phage display assay to predict inhibitor development.
• These models can predict inhibitor risk before initiating exogenous FVIII treatment.
• The models can inform the development of diagnostic tests for personalized HA management.