Pre-operative three-dimensional face scans for predicting difficult facemask ventilation: a prospective development...

Area of Science:

• Anesthesiology and airway management.
• Medical imaging and computational anatomy.
• Digital health and predictive diagnostics.

Background:

• Predicting difficult facemask ventilation is crucial but challenging.
• Pre-operative three-dimensional (3D) facial scanning shows potential diagnostic value.
• Interpretable facial shape features can aid in predicting ventilation difficulty.

Purpose of the Study:

• To identify interpretable facial shape features from 3D face scans.
• To quantify the diagnostic value of these features for predicting difficult facemask ventilation.
• To assess if integrating facial shape features improves existing prediction models.

Main Methods:

• Prospective observational study involving 398 patients undergoing surgery.
• Pre-operative 3D facial scans and structured airway assessments were performed.
• 3D scans were analyzed to extract shape coefficients; DIFFMASK score was evaluated with and without facial features.

Main Results:

• The DIFFMASK score had an optimism-corrected AUROC of 0.73.
• Selected facial shape features achieved an AUROC of 0.74.
• Enriching the DIFFMASK score with three facial features improved the AUROC to 0.76, with significant model fit improvement (p=0.002).
• Nasal morphology, lower mandible, neck region, and facial convexity were key predictive features.

Conclusions:

• Pre-operative 3D facial scans are effective in predicting difficult facemask ventilation.
• Integrating specific facial shape features enhances the diagnostic value of the DIFFMASK score.
• Digital phenotyping using 3D facial analysis offers a valuable complement to traditional clinical assessments.