Surface imaging for real-time patient respiratory function assessment in intensive care

Area of Science:

• Biomedical Engineering
• Critical Care Medicine
• Medical Imaging

Background:

• Continuous physiological monitoring is crucial in Intensive Care Units (ICUs) for assessing organ function.
• Current respiratory monitoring methods often require physical contact, posing challenges in critical care settings.
• There is a need for efficient, noncontact respiratory monitoring solutions in ICUs.

Purpose of the Study:

• To present a novel noncontact system for real-time respiratory monitoring and function assessment in ICU patients.
• To evaluate the efficacy of a noncontact system in estimating key respiratory parameters.

Main Methods:

• Utilized a Time-of-Flight depth sensor to analyze chest wall movements.
• Employed a deep neural network for automatic patient torso detection.
• Validated the system on a mannequin and 16 mechanically ventilated ICU patients.

Main Results:

• High correlation between estimated and reference respiratory rate (r=0.99 for mannequin, r=0.95 for patients).
• High correlation between estimated and reference tidal volume (r=0.99 for mannequin, r=0.90 for patients).
• All correlations were statistically significant (P < 0.001).

Conclusions:

• The developed noncontact system is suitable for continuous respiratory monitoring in ICUs.
• This technology offers a valuable tool for disease assessment in critically ill patients.
• The system demonstrates high accuracy in estimating respiratory parameters noninvasively.