VENTURISPIRO: development of an AI-enhanced, low-cost portable spirometer

Spirometry is fundamental for diagnosing chronic respiratory diseases, yet the high cost of existing devices prevents widespread access. Measurements such as forced vital capacity (FVC) and forced expiratory volume in the first second (FEV1) are crucial for diagnosis. The lack of affordable, portable options limits patient monitoring outside of clinical visits. This article presents the design and implementation of an extremely low-cost (~USD 37) portable spirometer integrated with explainable artificial intelligence (XAI) to improve accessibility. The hardware uses a 3D-printed Fleisch tube, an MPXV7002DP differential pressure sensor, and an ATMEGA328P microcontroller with wireless transmission of sensor data. A hybrid AI module processes this data with categorization into normal, restrictive, or obstructive lung function, with explanations of the GradCAM/LIME model provided. Preliminary validation demonstrates compliance with ATS/ERS accuracy standards and confirms its potential for use, particularly in resource-limited settings.