An objective, markerless videosystem for staging facial palsy

Background: The clinical classification of unilateral peripheral facial palsy (UPFP) is still based on subjective conventional methodology, leaving several missing points for an exact depiction of face deformity. Aim of the study: To propose a new objective, video recording method that relies on mathematical algorithms allowing the software to recognize numerical points on the face surface that would be indicative of facial nerve impairment, without positioning of markers on the face. Material and methods: Patients with UPFP of different House-Brackman (HB) degrees, from II to V, have been evaluated after video recording during two selected facial movements (forehead frowning and smiling) by a software trained to recognize the face points as numbers. Numerical parameters in millimeters have been obtained as indicative values of the shifting of the face points, and the shifting ratio between the healthy (denominator) and the affected side (numerator), i.e. the asymmetry index for the two movements taken into consideration. Results: For each HB grade, specific ranges of asymmetry index have been identified with a positive correlation for shift differences and negative correlation for asymmetry indexes. Conclusions: The use of the present objective system enabled the identification of numerical ranges of asymmetry index between the healthy and the affected side, that were found to be consistent with the outcome from the subjective methods currently in use.