Most people infected with Mycobacterium tuberculosis, the causative agent of tuberculosis (TB) actually maintain a strong immune response and are able to control bacterial growth (deemed latently infected (LTBI)), while approximately 10% progress to disease resulting in almost 2 million deaths per year. Determining the immune 'footprint' at specific stages of infection and disease will allow for better diagnostics, treatments and ultimately development of new vaccine candidates. In this study we performed multi-factorial flow cytometry on fresh blood from 56 TB cases, 46 Tuberculin Skin Test (TST) positive (LTBI) and 39 TST negative household contacts. We found a highly significant increase in granulocytes and decrease in B cells and invariant (Valpha24+Vbeta11+) NKT cells in TB cases compared to TST+ contacts (p<0.0001, p=0.007 and p=0.01 respectively) which were restored to LTBI levels following 6 months of TB treatment. Using support vector analysis, we found a combination of granulocyte and lymphocyte and/or NKT cell proportions allowed almost 90% correct classification into M. tuberculosis infection or disease. This work has important public health benefits in regards to diagnosis and treatment of TB in sub-Saharan Africa and in furthering our understanding of the requirements for protective immunity to TB.