Development and Clinical Application of a Novel Non-contact Early Airflow Limitation Screening System Using an Infrared Time-of-Flight Depth Image Sensor

Abstract

Obstructive pulmonary diseases, such as diffuse panbronchiolitis (DPB), asthma, chronic obstructive pulmonary disease (COPD), and asthma COPD overlap syndrome (ACOS) trigger a severe reaction at some situations. Detecting early airflow limitation caused by diseases above is critical to stop the progression. Thus, there is a need for tools to enable self-screening of early airflow limitation at home. Here, we developed a novel non-contact early airflow limitation screening system (EAFL-SS) that does not require calibration to the individual by a spirometer. The system is based on an infrared time-of-flight (ToF) depth image sensor, which is integrated into several smartphones for photography focusing or augmented reality. The EAFL-SS comprised an 850 nm infrared ToF depth image sensor (224 × 171 pixels) and custom-built data processing algorithms to visualize anterior-thorax three-dimensional motions in real-time. Multiple linear regression analysis was used to determine the amount of air compulsorily exhaled after maximal inspiration (referred to as the forced vital capacity, FVC_EAFL–SS) from the ToF-derived anterior-thorax forced vital capacity (FVC), height, and body mass index as explanatory variables and spirometer-derived FVC as the objective variable. The non-contact measurement is automatically started when an examinee is sitting 35 cm away from the EAFL-SS. A clinical test was conducted with 32 COPD patients (27/5 M/F, 67–93 years) as typical airflow limitation cases recruited at St. Marianna University Hospital and 21 healthy volunteers (10/11 M/F, 23–79 years). The EAFL-SS was used to monitor the respiration of examinees during forced exhalation while sitting still, and a spirometer was used simultaneously as a reference. The forced expiratory volume in 1 s (FEV1%_EAFL–SS) was evaluated as a percentage of the FVC_EAFL–SS, where values less than 70% indicated suspected airflow limitation. Leave-one-out cross-validation analysis revealed that this system provided 81% sensitivity and 90% specificity. Further, the FEV1_EAFL–SS values were closely correlated with that measured using a spirometer (r = 0.85, p < 0.0001). Hence, EAFL-SS appears promising for early airflow limitation screening at home.