PI-020 - PREDICTING OPTIMAL PATIENT-CENTRIC TREATMENT DURATION FOR LEVOFLOXACIN-CONTAINING TUBERCULOSIS REGIMEN: AN INTEGRATED APPROACH OF A MODEL-BASED STRATIFICATION TOOL WITH SITE-OF-DISEASE PHARMACOLOGY KNOWLEDGE

E. Yang¹, E. Dreesen^1,2, M. Morimoto¹, V. Chang¹, C. Cousins³, S. Wasserman³, R. Savic^1,4; ¹University of California, San Francisco, San Francisco, CA, United States, ²KU Leuven, Leuven, Belgium, ³Institute for Infection and Immunity, City St George's, University of London, London, United Kingdom, ⁴UCSF Center for Tuberculosis, University of California, San Francisco, San Francisco, CA, United States.

Background: Risk-stratified medicine is emerging as a promising strategy for shorter, safer, and more effective tuberculosis (TB) treatments, aligning with global health priorities. Yet, uncertainties remain about relative contributions of alternative drugs when introduced into TB regimens. Here, we leveraged our model-based stratification tool, integrating site-of-disease pharmacology knowledge, to predict clinical outcomes of a levofloxacin-containing regimen as a substitute for moxifloxacin and to inform its optimal patient-centric treatment duration based on risk phenotypes.

Methods: We utilized an in-house parametric survival model developed from pooled data on rifamycin regimens with or without moxifloxacin for drug susceptible TB. We incorporated the effect of a moxifloxacin-to-levofloxacin substitution into the model using rabbit lesion pharmacokinetic/pharmacodynamic indices from the literature, specifically drug exposure over efficacy targets for cellular lesions and caseum. This substitution was assumed to reduce hazard risk for unfavorable outcomes by one third. Using the integrated model, we simulated TB-related unfavorable outcomes under high-dose rifamycin-levofloxacin regimens across various treatment durations in 2,343 participants from Phase 3 Study 31/A5349 and calculated cure rates by risk phenotypes according to baseline disease severity, namely Xpert MTB/RIF cycle threshold and extent of disease on chest radiography. Simulations were performed with NONMEM v.7.5.1.

Results: A 4-month high-dose rifamycin regimen with levofloxacin was predicted to result in marginally more TB-related unfavorable outcomes at 12 months since treatment start (4.7%) compared to the approved 4-month regimen with moxifloxacin (3.8%). Easier-to-treat patients (75%) were expected to need a 4-month treatment, while harder-to-treat patients (25%) would require more than 4 months to achieve a 93% cure rate at 18 months with the levofloxacin-containing regimen.

Conclusion: We integrated our clinical stratification tool with translational aspects of TB pathology to optimize patient-centric treatment duration of a levofloxacin-containing regimen based on risk phenotypes. Our approach will offer a valuable rationale for designing future confirmatory clinical trials and advancing stratified medicine strategies for TB.