DOI 10.1016/j.clae.2023.102090
A Hue-Value method for semi-automated assessment of Lid Wiper Epitheliopathy.
Аннотация
BACKGROUND
Lid wiper epitheliopathy (LWE) is a marker of an abnormal lid/cornea interaction. This study proposes an automated Hue-Value grading algorithm of LWE staining following manual selection of the region of interest.
METHODS
Images of LWE staining were processed using Hue and Value from HSV (Hue-Saturation-Value) color space with a custom MATLAB program. Thirty-one images were successfully analyzed. Examiners analyzed images in random order twice, separated by more than a week. Bland Altman and Intraclass Correlation Coefficients (ICC) were performed.
RESULTS
There was no difference (p > 0.05) between upper (UL) and lower (LL) eyelids for LWE height (UL: 0.12 ± 0.12 mm, LL: 0.12 ± 0.07 mm), width (UL: 10.70 ± 3.84 mm, LL: 10.26 ± 3.49 mm), or area (UL: 2.85 ± 2.67 mm2, LL: 2.63 ± 1.71 mm2). There was no between examiner difference for all eyelid LWE height or area (p > 0.05), but a difference in LWE width (0.16 mm; p = 0.031). ICC for LWE height, width and area were 0.996 (95% CI: 0.993 to 0.998), 0.997 (95% CI: 0.992 to 0.998) and 0.999 (95% CI: 0.998 to 0.999). There was no between examiner difference for height or area (p > 0.05) for UL, but a difference in LWE width (0.28 mm; p = 0.026). ICC for height, width and area were 0.999 (95% CI: 0.996 to 1.00), 0.995 (95% CI: 0.982 to 0.999) and 1.00 (95% CI: 0.999 to 1.00). There was no difference in LWE height, width or area for LL (all p > 0.05). ICC were 0.991 (95% CI: 0.973 to 0.997) for height, 0.998 (95% CI: 0.995 to 0.999) for width and 0.997 (95% CI: 0.990 to 0.999) for area.
CONCLUSIONS
This novel method results in highly repeatable interexaminer measures of LWE staining after general lid region delineation. Small differences in LWE width were observed between examiners.
Предметные рубрики MeSH
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.