Assessing Hemifacial Spasm Severity and Prognosis via Facial Tracking Tech

Overview

This study utilized facial recognition and tracking technologies to objectively quantify and grade hemifacial spasms (HFS), aiming to overcome limitations of existing grading systems. By analyzing preoperative videos with advanced software, the study measured spasm amplitude and frequency, facilitating improved classification and treatment outcome prediction.

Background

Hemifacial spasm is characterized by involuntary contractions on one side of the face, typically caused by vascular compression at the nerve exit zone. Existing grading systems for HFS are often complex and overlapping, limiting their clinical utility. Microvascular decompression (MVD) is the preferred treatment for complete cure, but variability in spasm nature and severity raises questions about uniform treatment efficacy. Facial recognition and tracking technologies offer promising tools to objectively assess spasms and guide management decisions.

Data Highlights

Preoperative videos from patients with confirmed microvascular compression syndrome were analyzed using Apple's AR-facial tracking kit and Blender software. Key parameters measured included percentage change in distance between mouth angles and eye closure to quantify spasm amplitude, and spasms per second to quantify frequency. Spasm frequency was also categorized via patient questionnaires into four intensity levels. Clustering methods using Euclidean distance were applied to classify spasms based on these metrics.

Key Findings

• Facial tracking technology enabled creation of a reproducible "face mesh" to measure dynamic facial movements during spasms.
• Amplitude was standardized by calculating percentage changes in mouth angle distance and eye closure relative to resting state.
• Spasm frequency was objectively measured as spasms per second from video frame analysis and subjectively categorized via patient questionnaires.
• Hierarchical and k-means clustering methods effectively classified spasms based on amplitude and frequency metrics.
• Use of these technologies addresses limitations of prior grading systems by providing objective, quantifiable data to guide treatment decisions.

Clinical Implications

Incorporating facial recognition and tracking technologies into clinical practice can enhance the objectivity and accuracy of hemifacial spasm assessment. This approach may improve patient stratification for treatments such as microvascular decompression or botulinum toxin injections by better characterizing spasm severity and frequency. Ultimately, it supports personalized management and more precise monitoring of treatment outcomes.

Conclusion

Facial recognition and tracking technologies provide a novel, objective method to quantify and classify hemifacial spasms, overcoming limitations of existing grading systems. This advancement holds promise for improving treatment decision-making and prognostication in patients with HFS.

References

1. Standards for Reporting of Diagnostic Accuracy (STARD) guideline statement [4]
2. Apple Inc. AR-facial tracking kit [12, 13, 17]
3. Hemifacial spasm grading systems and clinical context [9, 14, 15, 16, 21]
4. R software and R-Studio for clustering analysis [25]