基于人工智能模型量化视网膜血管特征参数预测子痫前期的可行性研究

周天凡, 邵飞雪, 万盛, 周晨晨, 周思锦, 花晓琳

Feasibility study on quantifying retinal vascular features for predicting preeclampsia based on artificial intelligence models

ZHOU Tianfan, SHAO Feixue, WAN Sheng, ZHOU Chenchen, ZHOU Sijin, HUA Xiaolin

表2 3组孕妇的眼底特征及视网膜血管特征参数比较

Tab 2 Comparison of fundus characteristics and retinal vascular characteristic parameters among the three groups of pregnant women

Characteristic	Unaffected group (n=685)	HDP group (n=104)				P^c value
Characteristic	Unaffected group (n=685)	GH group (n=36)	P^a value	PE group (n=68)	P^b value	P^c value
Decreased elasticity of retinal arteries/n(%)	196 (28.61)	12 (33.33)	0.572	23 (33.82)	0.401	0.532
Leopard pattern change/n(%)	456 (66.57)	23 (63.89)	0.721	48 (70.59)	0.589	0.746
Arteriosclerosis/n(%)	4 (0.58)	1 (2.78)	0.227	1 (1.47)	0.378	0.181
Vitreous warts/n(%)	42 (6.13)	2 (5.56)	1.000	4 (5.88)	1.000	1.000
Retinal sporadic bleeding/n(%)	7 (1.02)	1 (2.78)	0.338	1 (1.47)	0.533	0.337
CRAE	94.00 (87.00, 99.00)	89.00 (86.00, 95.00)	0.150	87.00 (80.00, 94.00)	0.000	0.000
CRVE	122.00 (116.00, 129.00)	120.00 (116.00, 122.50)	0.101	120.00 (111.00, 126.50)	0.017	0.019
AVR	0.75 (0.71, 0.81)	0.74 (0.70, 0.79)	0.432	0.72 (0.67, 0.77)	0.002	0.006
Retinal artery tortuosity	0.05 (0.04, 0.07)	0.04 (0.04, 0.07)	0.567	0.04 (0.03, 0.06)	0.004	0.015
Retinal vein tortuosity	0.09 (0.07, 0.11)	0.07 (0.06, 0.11)	0.186	0.08 (0.06, 0.10)	0.120	0.141
Retinal artery fractal dimension	1.48 (1.41, 1.54)	1.47 (1.41, 1.51)	0.245	1.45 (1.38, 1.51)	0.003	0.007
Retinal vein fractal dimension	1.49 (1.42, 1.56)	1.51 (1.45, 1.55)	0.569	1.48 (1.44, 1.55)	0.990	0.848
VCDR	0.28 (0.22, 0.34)	0.29 (0.20, 0.33)	0.688	0.27 (0.22, 0.34)	0.522	0.764
HCDR	0.39 (0.31, 0.46)	0.40 (0.31, 0.45)	0.967	0.39 (0.31, 0.45)	0.740	0.941