RESUMO
The sensitivity of in vivo low-dose high-resolution micro-computed tomography imaging enables monitoring the lung damage caused by tuberculosis. Here, we propose a radiological score integrated in the experimental workflow that enables longitudinal monitoring for prospective efficacy studies in drug development programs. The score is based on an automatic measurement of total unaffected lung volume in vivo normalized for inter-subject comparison. It was validated on well-characterized progression of chronic tuberculosis in Erdman and H37Rv strains in C3HeB/FeJ-based models. We demonstrated that a decrease in the score value indicates increasing adverse effects and vice versa. The colony-forming units count confirmed the variability in the host response suggested by the score values. The correlation between changes in the mice's weight and the score is consistent with disease progression. The classification of disease extent by k-means clustering of the score values provided the definition of the lung damage severity according to the bacillus strain. The proposed score will reduce sources of bias and improve the statistical robustness of studies by the attrition of non-infected subjects or subjects with a weak immune response. Readily available quantifications allow for a fast assessment of the therapeutic potential in drug-resistant tuberculosis strains.