Electronic noses detecting patterns of volatile molecules have recently been introduced for different diagnostic purposes. The diagnostic accuracy of a prototype e-nose device (Bruins et al (2013) in Bangladesh showed sensitivity of 76.5-95.9% and specificity of 85.3-98.5%. Here the investigators test a production type point-of-care hand-held device with less detectors. The investigators explore factors such as food intake, smoking, and co-morbidity, as well as the impact of TB treatment, and address the question whether the device could help monitor disease and response to treatment.

Study participants provide WIC and after time for consideration, when they approve, provide demographic anonymized data.

The study population consists of six groups:

Group 1 - 20 TB patients aged > 18 yrs Group 2 - 20 non---TB patients > 18 yrs (screened for TB - other conditions) Group 3 - 20 patients with a lung disease - no TB suspects (Lung Clinics in Yogyakarta,Indonesia; Group 4 - 50 apparently healthy matched controls Group 5 - 20 newly diagnosed TB patients enrolled before start of treatment, to be followed over time, until after end of treatment (8 months FU).

Group 6 - 50 apparently healthy volunteers.

Study participants breath normally by mouthpiece into the device using a nose clamp - for 5 min during each sampling.

Electronic data are downloaded and transmitted by internet to Zutphen, Netherlands where data are processed to build the model.

Inclusion Criteria:

• age >18yrs
• voluntarily participating

Exclusion Criteria:

-