PII-020 - A MATHEMATICAL MODEL FOR SELECTING REGIMEN COMPOSITION AND DURATION BASED ON PHENOTYPE FOR PATIENTS WITH DRUG-SUSCEPTIBLE TUBERCULOSIS.

E. Dreesen^1,2, M. Imperial², V. Chang², R. Savic²; ¹Katholieke Universiteit Leuven, Leuven, Belgium, ²University of California, San Francisco, San Francisco, CA, USA.

Background: Despite many large-scale clinical trials, little progress has been made in shortening duration of tuberculosis (TB) treatment. We previously developed a risk stratification tool to inform selection of optimal treatment duration for novel clinical trials.(1) However, the treatment shortening potential of individual drugs has not been quantified.

Methods: A parametric survival model was developed based on pooled TB-ReFLECT and S31/A5349 (NCT02410772) data (N=6346 patients; NONMEM 7.5).(1,2) The primary endpoint was time to unfavorable outcome ≤18(mo)nths after start of treatment (SoT). Regimens included standard dose isoniazid [H], [R]ifampin, pyra[Z]inamide, [E]thambutol, [M]oxifloxacin, and/or standard/high dose rifa[P]entine. The effects of baseline phenotype and the number of treatment days (surrogate for treatment duration) were quantified. Effects of E-to-M and R-to-P substitutions were estimated from S31/A5349 and the pooled data, resp. The final model was used to assess in silico clinical trial designs in which treatment duration was based on phenotype (low/moderate/high risk) and regimen composition, aiming at a 95% cure rate 12mo after SoT.

Results: Effects of E-to-M (-0.26 [28%]) and R-to-P (high dose; -0.28 [29%]) substitutions on baseline hazard were identified. Clinical trial simulations (N=500) suggest that inclusion of M and high dose P can compensate for treatment extension for patients with higher risk of unfavorable outcomes. Treatments can be shortened from 18 (HRZE) to 14 (HPZE) to 8 weeks (HPZM) [low risk], from 23 to 21 to 17 weeks [moderate risk], and from 26 to 25 to 23 weeks [high risk].

Conclusion: Our tool can inform selection of TB regimen compositions and durations for each patient phenotype, thereby maximizing cure rates for all patients and informing the design of innovative trials.

(1) Imperial, M.Z., Phillips, P.P.J., Nahid P., Savic, R.M. Precision-Enhancing Risk Stratification Tools for Selecting Optimal Treatment Durations in Tuberculosis Clinical Trials. Am. J. Respir. Crit. Care Med. 204, 1086-1096 (2021).
(2) Dorman, S.E. et al. Four-Month Rifapentine Regimens with or without Moxifloxacin for Tuberculosis. N. Engl. J. Med. 384: 1705-1718 (2021).