Further analysis of assessments of the coefficient of variation of corneal endothelial cell areas from specular microscopic images

被引:30
作者
Doughty, Michael J. [1 ]
Aakre, Bente Monica
机构
[1] Glasgow Caledonian Univ, Dept Vis Sci, Glasgow G4 0BA, Lanark, Scotland
关键词
cornea; digitiser pad; endothelium; human; image analysis; morphometry;
D O I
10.1111/j.1444-0938.2008.00281.x
中图分类号
R77 [眼科学];
学科分类号
100212 ;
摘要
Purpose: The aim of this study was to compare two methods of assessments of the coefficient of variation (COV) of endothelial cell area. Methods: A single image (Topcon SP-2000P specular microscope) was obtained from the central region of the corneal endothelium of 45 healthy white (Norwegian) individuals, aged from 24 to 43 years and without a history of major eye disease or surgery. The image file was printed to A3-size, the cell-cell boundaries marked manually and the areas of the cells measured with a digitiser pad. The same image file was independently processed by the semi-automated Topcon IMAGEnet system. From either method, the cell area data from 100 contiguous cells approximately in the middle portion of the images were used to calculate the average cell area (AVG), the coefficient of variation (COV) on the cell areas and the endothelial cell density (ECD). Results: Both methods produced similar AVG and ECD values that were not statistically different (p >= 0.180). The SD values on the cell areas increased in relation to the AVG values (Pearson's r >= 0.557). The resultant COV values were only marginally higher with the manual method (27.8 versus. 26.3 per cent) but the limits of agreement (LoA) for the COV values were rather large at -4.9 to +7.9 per cent. Conclusion: A semi-automated image analysis system can be used to generate COV data for the corneal endothelium similar to those of a manual method. The limits of agreement between the methods are substantial and this probably reflects the extreme sensitivity of the COV calculation to even a few different cell area values. This poor agreement needs to be considered in any comparative studies.
引用
收藏
页码:438 / 446
页数:9
相关论文
共 41 条
[1]  
AMON M, 1988, SPEKTRUM AUGENHEILKD, V2, P245
[2]  
Benetz BAM, 1999, CORNEA, V18, P67
[3]   STATISTICAL METHODS FOR ASSESSING AGREEMENT BETWEEN TWO METHODS OF CLINICAL MEASUREMENT [J].
BLAND, JM ;
ALTMAN, DG .
LANCET, 1986, 1 (8476) :307-310
[4]   Clinical responses of the corneal endothelium [J].
Bourne, WM ;
McLaren, JW .
EXPERIMENTAL EYE RESEARCH, 2004, 78 (03) :561-572
[5]  
Cheung S., 1998, CLIN EXP OPTOM, V81, P1
[6]   Endothelial cells analysis with the TOPCON specular microscope SP-2000P and IMAGEnet system [J].
Cheung, SW ;
Cho, P .
CURRENT EYE RESEARCH, 2000, 21 (04) :788-798
[7]   Effects of laser in situ keratomileusis (LASIK) on the corneal endothelium 3 years postoperatively [J].
Collins, MJ ;
Carr, JD ;
Stulting, RD ;
Azar, RG ;
Waring, GO ;
Smith, RE ;
Thompson, KP ;
Edelhauser, HF .
AMERICAN JOURNAL OF OPHTHALMOLOGY, 2001, 131 (01) :1-6
[8]   Corneal endothelium evaluation with 2 noncontact specular microscopes and their semiautomated methods of analysis [J].
de Sanctis, U ;
Machetta, F ;
Razzano, L ;
Dalmasso, P ;
Grignolo, FM .
CORNEA, 2006, 25 (05) :501-506
[9]  
Doughty Michael J, 2007, Eye Contact Lens, V33, P180, DOI 10.1097/01.icl.0000250885.35316.85
[10]   Could the coefficient of variation (COV) of the corneal endothelium be overestimated when a centre-dot method is used? [J].
Doughty, Michael J. .
CLINICAL AND EXPERIMENTAL OPTOMETRY, 2008, 91 (01) :103-110