PET(CO2) measurement and feature extraction of capnogram signals for extubation outcomes from mechanical ventilation

Capnography is a continuous and noninvasive method for carbon dioxide (CO2) measurement, and it has become the standard of care for basic respiratory monitoring for intubated patients in the intensive care unit. In addition, it has been used to adjust ventilatory parameters during mechanical ventilation (MV). However, a substantial debate remains as to whether capnography is useful during the process of weaning and extubation from MV during the postoperative period. Thus, the main objective of this study was to present a new use for time-based capnography data by measuring the end-tidal CO2 pressure ([Formula: see text]), partial pressure of arterial CO2 ([Formula: see text]) and feature extraction of capnogram signals before extubation from MV to evaluate the capnography as a predictor of outcome extubation in infants after cardiac surgery. Altogether, 82 measurements were analysed, 71.9% patients were successfully extubated, and 28.1% met the criteria for extubation failure within 48 h. The ROC-AUC analysis for quantitative measure of the capnogram showed significant differences (p < 0.001) for: expiratory time (0.873), slope of phase III (0.866), slope ratio (0.923) and ascending angle (0.897). In addition, the analysis of [Formula: see text] (0.895) and [Formula: see text] (0.924) obtained 30 min before extubation showed significant differences between groups. The [Formula: see text] mean value for success and failure extubation group was 39.04 mmHg and 46.27 mmHg, respectively. It was also observed that high CO2 values in patients who had returned MV was 82.8  ±  21 mmHg at the time of extubation failure. Thus, [Formula: see text] measurements and analysis of features extracted from a capnogram can differentiate extubation outcomes in infant patients under MV, thereby reducing the physiologic instability caused by failure in this process.