Prediction of Long-Term Factor IX Durability Among Hemophilia B Patients Treated with Fidanacogene Elaparvovec Gene Therapy

Speaker(s)

Morris A1, Ortega L2, Thakkar S3, Wilcox L4, Cappelleri JC5, Incerti D6, Ines M7, Deb P8, Chhabra A9, Vytlacil E10, Batt K11, Lakdawalla D12
1EntityRisk, Inc., Princeton, NJ, USA, 2EntityRisk, Inc., Washington, DC, USA, 3Pfizer Inc, Cambridge, MA, USA, 4Pfizer Canada, Kirkland, QC, Canada, 5Pfizer Inc, Groton, CT, USA, 6EntityRisk, Inc., San Francisco, CA, USA, 7Pfizer Inc, Porto Salvo, Lisboa, Portugal, 8Hunter College, New York, NY, USA, 9Pfizer Inc, New York, NY, USA, 10Yale University, New Haven, CT, USA, 11EntityRisk, Inc., Raleigh, NC, USA, 12USC Leonard D. Schaeffer Center for Health Policy and Economics, Los Angeles, CA, USA

Presentation Documents

OBJECTIVES: Phase 1/2 (NCT02484092 and NCT03307980) and 3 (NCT03861273) clinical trials assess the safety and efficacy of fidanacogene elaparvovec gene therapy for the treatment of hemophilia B. Since gene therapies for hemophilia have been only recently developed, there is limited clinical trial and real-world evidence available to evaluate their long-term effects. This study aims to predict the long-term distribution of factor IX (FIX) activity within moderately-severe and severe hemophilia B (FIX:C ≤ 2%) patients treated with fidanacogene elaparvovec.

METHODS: Year-2 data (last measurement as of 8/30/2023) from the three non-randomized clinical trials were pooled to increase sample size (14 and 43 patients from Phase 1/2 and 3, respectively) and to extend the length of follow-up time (up to six years after infusion); age, body mass index, and baseline FIX levels were included as regressors to adjust for baseline differences across trials. FIX activity levels tend to reach stability by or before six months post-infusion, designated as our baseline. The models generate FIX activity level predictions based on SynthASil assay. Model parameters were estimated using Bayesian methods with weakly informative priors. Posterior predictive checks were used to evaluate model performance. Predictions were made among a population of 1,000 simulated patients sampled with replacement from the combined trial data.

RESULTS: Predicted median FIX activity in the treated population based on predicted SynthASil assay measurements was 8.2% (95% CrI: 2.3%, 27.1%) at 30 years post-infusion. By 30 years, 86.5% (95% CrI: 54.7%, 99.6%) of the population was predicted to have FIX activity above 2%. Results were similar when using alternative modeling scenarios such as lognormal, gamma, and non-parametric regression models.

CONCLUSIONS: Our model predicts that a vast majority of hemophilia B patients receiving fidanacogene elaparvovec would experience sustained long-term effects, maintaining FIX activity levels above 2% for up to 30 years after infusion.

Code

CO3

Topic

Clinical Outcomes, Methodological & Statistical Research, Study Approaches

Topic Subcategory

Artificial Intelligence, Machine Learning, Predictive Analytics, Clinical Trials, Relating Intermediate to Long-term Outcomes

Disease

Genetic, Regenerative & Curative Therapies, Systemic Disorders/Conditions (Anesthesia, Auto-Immune Disorders (n.e.c.), Hematological Disorders (non-oncologic), Pain)