Intracellular pharmacodynamic modelling ( PDi ) is predictive of the clinical 1 activity of fluoroquinolones against Tuberculosis 2

Clinical studies of new anti-tubercular drugs are costly and time consuming. Owing 31 to the extensive TB treatment periods, the ability to identify drug candidates based on 32 their predicted clinical efficacy is vital to accelerate the pipeline of new therapies. 33 Recent failures of pre-clinical models in predicting the activity of fluoroquinolones 34 underlines the importance of developing new and more robust predictive tools that 35 will optimise the design of future trials. Here, we have used high-content imaging 36 screening and pharmacodynamic intracellular modelling ( PDi ) to identify and 37 prioritise fluoroquinolones for TB treatment. In a set of studies designed to validate 38 this approach, we show moxifloxacin to be the most effective fluoroquinolone, and 39 PDi modelling-based Monte Carlo simulations accurately predict negative culture 40 conversion (sputum sterilisation) rates when compared against 8-independent clinical 41 trials. Additionally, PDi -based simulations were used to predict the risk of relapse. 42 Our analyses show that the duration of treatment following culture conversion can be 43 used to predict relapse rate. These data further support that PDi -based modelling 44 offers a much-needed decision making tool for the TB drug development pipeline. 45