Diagnosis of Laryngomalacia by Fiberoptic Endoscopy: Data and Statistical Analysis

Mar-21-2015

Diagnosis of Laryngomalacia by Fiberoptic Endoscopy: Data and Statistical AnalysisAn additional 25 video recordings from infants with normal upper airways who underwent laryngoscopies (WT, 12 recordings; DAT, 13 recordings) for reasons unrelated to upper airways problems were also copied to the new video cassettes and served as healthy control subjects. The causes for performing FFL in these cases were nocturnal snoring, following adenoidectomy, chronic rhinorrhea, BAL, and suspected foreign body aspiration. Each video clip was edited to visualize the motion of the supraglottic structures only, and the observer was blinded to the patient, the clinical presentation, and the technique. Sound was not included so as not to influence the visual diagnosis. A total of 140 video clips were copied in a random sequence, and each clip was assigned a new number. Based on a previously applied scoring system for laryngomala-cia, each investigator assigned a score for each video clip that was composed of the following two parts: the “epiglottic score” (range, 0 to 4 points); and the “arytenoids score” (range, 0 to 4 points). Hence, the maximal possible laryngomalacia video score was 8 points (Table 2). generic yaz birth control
The correlation of the clinical and video scores was evaluated by the Spearman correlation coefficient for each method and for each investigator. The clinical agreement between the two investigators was assessed by comparing their clinical scores using the Kendall coefficient of concordance, the Spearman correlation coefficient, and the Wilcoxon nonparametric test. The interinvestigator agreement for the video scoring was assessed by comparing their correlations, differences, and variability regarding the epiglottic, arytenoids, and total video scores using the Kendall coefficient of concordance, the Friedman nonparametric test, and the Wilcoxon nonparametric test, respectively. The correlation between each two of the three investigators was evaluated by the Spearman correlation coefficient. The intrainvestigator variability was studied by comparing the two assessments of the same video clips (ie, the original recorded clip and the repeat clip) for each of the investigators by the Spearman correlation coefficient and the Wilcoxon nonparametric test. The accuracy of the DAT and WT in diagnosing laryngomalacia was assessed by comparing the area under the curve (AUC) of the receiver operating characteristics of each test using the methodology described by Hanley and McNeil. Models with AUC values between 0.70 and 0.79 are considered as having moderate discriminative properties, and those with AUC values of > 0.80 as having good discriminative properties. Statistical analysis was performed using a statistical software package (SPSS for Windows, version 13.0; SPSS; Chicago, IL).

Table 2—Laryngomalacia and Laryngoscopic Score

Findings Score
Arytenoid score
No discernible collapse into glottis with inspiration 0
Subtle collapse of arytenoids into glottis 1
Collapse of arytenoids into glottis, 25-50% of vocal cords obscured 2
Collapse of arytenoids into glottis, about 75% of vocal cords obscured 3
Collapse of arytenoids into glottis, 100% of vocal cords obscured 4
Epiglottic score
Normal epiglottis, no folding during inspiration 0
Slight length-wise folding of epiglottis 1
Moderate fold of epiglottis without contact between lateral edges 2
Intermittent contact of lateral edges of epiglottis 3
Continuous contact and even overlap of lateral edges of epiglottis 4