Sensitive and reproducible methods are needed to measure the impact on the host following experimental challenge with Mycobacterium tuberculosis, in order to determine the degree of protection conferred by new vaccines. Here we compare how well different clinical and post-mortem measures of disease burden predict the response by the host to increasing doses of M. tuberculosis in rhesus and cynomolgus macaques. The total lung and lesion volume was quantified from magnetic resonance imaging (MRI) digital stacks obtained from lungs of M. tuberculosis infected animals that were formalin fixed and scanned ex-vivo. The total lung lesion volume relative to the fixed whole lung volume was superior at indicating disease burden when compared to thoracic radiography, pathology scores, changes in body weight and temperature, as well as erythrocyte haemoglobin concentrations and sedimentation rate. The total lesion volume accurately reflected differences in challenge doses of M. tuberculosis that ranged from 30 to 500 CFU delivered by aerosol. The determination of total lesion volume from MR images demonstrated a species-dependent difference between rhesus and cynomolgus macaques in susceptibility to M. tuberculosis infection. MR stereology provides an accurate, quantifiable and relatively simple assessment, which can be easily standardized between laboratories and should form an essential component of the clinical assessment of disease progression, or vaccine efficacy.