Files

Abstract

Acute compartment syndrome (ACS) is an orthopedic emergency caused by elevated pressure within a muscle compartment that can lead to permanent tissue damage and death. Current ACS diagnosis relies heavily on patient-reported symptoms, a method that is subjective and often followed by invasive intracompartmental pressure measurements that can be faulty in motion settings. Reliable motion diagnosis is critical for long-term monitoring that involves limb movement. This study proposes an objective and noninvasive diagnostic for ACS. Our device utilizes a random forest machine learning model that uses analog readings from force-sensitive resistors (FSRs) placed on the skin. To validate the machine learning diagnostic model, a data set containing FSR measurements and the corresponding simulated pressure was created for motion and motionless scenarios. Our diagnostic achieved up to 98% accuracy and excelled in key performance metrics, including sensitivity (97%) and specificity (98%), with a statistically insignificant (±5% error bars) performance difference in motion present cases. These results demonstrate the potential of noninvasive ACS diagnostics to meet clinical accuracy standards in real world scenarios.

Details

PDF

Statistics

from
to
Export
Download Full History