Materials and Methods
Thirty flexible fibreoptic nasendoscopy videos of laryngeal movement were captured in a laryngology clinic. These ranged from normal vocal cord movement to complete laryngeal paralysis (nine normal cases, four palsies, three nodules, two cases each of cysts, functional dysphonia, and inflammation, one case each of Reinke’s oedema, presbyphonia, polyp, hypo pharynx lesion, supraglottic lesion, crescentic defect of vocal cord, weakness and slower right vocal cord movement). The videos were pre-processed to reduce the effect of the honeycomb artefact caused by the fibreoptic endoscopes5. Six consultant head and neck surgeons (JM, RT, OH, MB, SR, KM) were asked to subjectively assess vocal cord motion by visual inspection of the laryngeal videos and individually rate the movement of the left and right vocal cords independently on a scale of 0 to 4, see table 1. There was no clinical history or sound associated with the videos. This process was repeated with the same videos, in a different order, on three separate occasions with a minimum of two weeks between each rating session. Each consultant rated the videos three times giving a total of 180 individual ratings (2 [R & L cord] x 30 x 3 = 180 ) per consultant and 1080 (180 x 6 consultants) ratings in total. The consultants were blinded to their previous and other raters’ scores. Ethical approval was not required for this study.