Virtual Booth

Diabetes mellitus remains a major global concern, with many cases undiagnosed due to costly and inaccessible diagnostic methods, especially in rural areas. This study developed and evaluated BreathRight-AI, a portable, non-invasive breath analyzer for detecting acetone in exhaled breath as a marker of diabetic risk. The device integrates an MQ-138 sensor with an ESP32 microcontroller to measure acetone, classify risk, and log data. Using a quantitative correlational design, thirty (30) diabetic and non-diabetic participants were purposively selected. Breath acetone levels (ppm) were compared with fasting blood sugar (mg/dL) using a glucometer. Results revealed a very strong correlation (r = 0.981, p < 0.00001), with each 1 ppm acetone increase corresponding to a 56 mg/dL rise in blood glucose. The prototype showed 95.6% classification accuracy, with better performance in low- and high-risk cases than borderline ones. Stability tests showed consistent results (SD = 0.073 ppm). Findings suggest that BreathRight-AI is a promising, user-friendly screening tool for early diabetes detection, though further calibration and large-scale clinical validation are needed.

Keywords: Diabetes, Breath Acetone, Non-invasive screening, MQ-138 sensor, Diabetic risk monitoring