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

doi:10.3969/j.issn.1674-8115.2024.05.002

Abstract

Abstract:

Objective ·To explore the predictive capability of retinal vascular features in preeclampsia (PE) based on artificial intelligence (AI) models. Methods ·This retrospective study enrolled 789 pregnant women who registered from June 2020 to January 2021 at Shanghai First Maternity and Infant Hospital of Tongji University in the first 16 weeks of gestation. These women underwent regular prenatal examinations, had retinal fundus images captured, and delivered singleton live births at the hospital. According to whether they developed hypertensive disorders of pregnancy (HDP), they were divided into unaffected group (n=685) and HDP group (n=104). Within the HDP group, pregnancies were further categorized into gestational hypertension (GH) group (n=36) and PE group (n=68) based on the occurrence of PE. Based on the gestational age at onset, the PE group was further divided into early-onset PE group (gestational age<34 weeks) and late-onset PE group (gestational age≥34 weeks). Fundus images of the pregnant women were obtained, and an AI algorithm was utilized to diagnose retinal lesions and quantify retinal vascular features. Comparative analyses were conducted on fundus features and retinal vascular features. Univariate Logistic regression model was employed to analyze the influencing factors of PE occurrence, and multivariate Logistic regression model was further utilized to assess the correlation between retinal vascular features and the occurrence of PE. The predictive capability of retinal vascular features for PE (both early- and late-onset PE) was analyzed by using area under the curve (AUC) of receiver operator characteristic curve (ROC curve). Results ·The comparative analysis of fundus features and retinal vascular features demonstrated statistically significant differences between the unaffected group and PE group in terms of central retinal artery equivalent (CRAE), central retinal vein equivalent (CRVE), arteriole-to-venular ratio (AVR), retinal artery tortuosity and retinal artery fractal dimension (all P<0.05). Univariate Logistic regression analysis indicated that second-trimester mean arterial pressure (MAP), second-trimester estimated fetal weight (EFW), CRAE, CRVE, AVR, retinal artery tortuosity and retinal artery fractal dimension were the influencing factors for PE occurrence (all P<0.05). Multivariate Logistic regression analysis revealed that second-trimester EFW, CRAE, CRVE, AVR, retinal artery tortuosity and retinal artery fractal dimension were the protective factors for the occurrence of PE, while second-trimester MAP was the risk factor for PE (all P<0.05). The analysis of ROC curves revealed that maternal risk factors along with second-trimester prenatal examination data (including MAP and EFW) and retinal vascular features model had good predictive ability for PE [AUC (95% CI)=0.784 (0.725－0.843), and this model exhibited better predictive capability for early-onset PE, with an AUC (95% CI) of 0.840 (0.756－0.924). Conclusion ·The integration of quantified retinal vascular features based on AI models with maternal risk factors and second-trimester prenatal examination data (including MAP and EFW) enables a more effective prediction of PE occurrence, particularly early-onset PE.

Key words: preeclampsia (PE), retinal vessel, artificial intelligence (AI)

CLC Number:

R714.7

ZHOU Tianfan, SHAO Feixue, WAN Sheng, ZHOU Chenchen, ZHOU Sijin, HUA Xiaolin. Feasibility study on quantifying retinal vascular features for predicting preeclampsia based on artificial intelligence models[J]. Journal of Shanghai Jiao Tong University (Medical Science), 2024, 44(5): 552-559.

Figures/Tables 5

TrendMD

References 25

1	RANA S, LEMOINE E, GRANGER J P, et al. Preeclampsia: pathophysiology, challenges, and perspectives[J]. Circ Res, 2019, 124(7): 1094-1112.
2	CHAPPELL L C, CLUVER C A, KINGDOM J, et al. Pre-eclampsia[J]. Lancet, 2021, 398(10297): 341-354.
3	Gestational Hypertension and Preeclampsia: ACOG Practice Bulletin, Number 222[J]. Obstet Gynecol, 2020, 135(6): e237-e260.
4	WRIGHT D, WRIGHT A, NICOLAIDES K H. The competing risk approach for prediction of preeclampsia[J]. Am J Obstet Gynecol, 2020, 223(1): 12-23. e7.
5	WRIGHT D, SYNGELAKI A, AKOLEKAR R, et al. Competing risks model in screening for preeclampsia by maternal characteristics and medical history[J]. Am J Obstet Gynecol, 2015, 213(1): 62.e1-62.e10.
6	TAN M Y, WRIGHT D, SYNGELAKI A, et al. Comparison of diagnostic accuracy of early screening for pre-eclampsia by NICE guidelines and a method combining maternal factors and biomarkers: results of SPREE[J]. Ultrasound Obstet Gynecol, 2020, 51(6): 743-750.
7	ZEISLER H, LLURBA E, CHANTRAINE F, et al. Predictive value of the sFlt-1: PLGF ratio in women with suspected preeclampsia[J]. N Engl J Med, 2016, 374(1): 13-22.
8	ACOG committee opinion No. 743: low-dose aspirin use during pregnancy[J]. Obstet Gynecol, 2018, 132(1): e44-e52.
9	ZHU Z T, CHEN Y F, WANG W, et al. Association of retinal age gap with arterial stiffness and incident cardiovascular disease[J]. Stroke, 2022, 53(11): 3320-3328.
10	WHELTON P K, CAREY R M, ARONOW W S, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines[J]. Hypertension, 2018, 71(6): 1269-1324.
11	ABU SAMRA K. The eye and visual system in the preeclampsia/eclampsia syndrome: what to expect?[J]. Saudi J Ophthalmol, 2013, 27(1): 51-53.
12	STERN E M, BLACE N. Ophthalmic pathology of preeclampsia[M]. Treasure Island (FL): StatPearls Publishing, 2022.
13	ZHOU L Q, WU X L, HUANG S Y, et al. Lymph node metastasis prediction from primary breast cancer US images using deep learning[J]. Radiology, 2020, 294(1): 19-28.
14	HADLOCK F P, HARRIST R B, SHARMAN R S, et al. Estimation of fetal weight with the use of head, body, and femur measurements: a prospective study[J]. Am J Obstet Gynecol, 1985, 151(3): 333-337.
15	黄烨霖, 赵昕. 用于对视网膜眼底图像进行分析的方法及其相关产品: CN114359284B[P]. 2022-06-21.
	HUANG Y L, ZHAO X. Method and related products for analyzing retinal fundus images: CN114359284B[P]. 2022-06-21.
16	黄烨霖, 史晓宇. 确定视网膜眼底血管的弯曲度的方法、设备和存储介质: CN115511883B[P]. 2023-04-18.
	HUANG Y L, SHI X Y. Method, equipment, and storage medium for determining the curvature of retinal fundus blood vessels: CN115511883B[P]. 2023-04-18.
17	黄烨霖, 赵昕, 和超, 等. 一种训练眼底图像分割模型的方法以及动静脉分割方法: CN113920077A[P]. 2022-01-11.
	HUANG Y L, ZHAO X, HE C, et al. A method for training fundus image segmentation models and arteriovenous segmentation method: CN113920077A [P]. 2022-01-11.
18	中华医学会妇产科学分会妊娠期高血压疾病学组. 妊娠期高血压疾病诊治指南(2020)[J]. 中华妇产科杂志, 2020, 55(4): 227-238.
	Hypertensive Disorders in Pregnancy Subgroup, Chinese Society of Obstetrics and Gynecology, Chinese Medical Association. Diagnosis and treatment of hypertension and pre-eclampsia in pregnancy: a clinical practice guideline in China(2020)[J]. Chinese Journal of Obstetrics and Gynecology, 2020, 55(4): 227-238.
19	STIRNEMANN J, VILLAR J, SALOMON L J, et al. International estimated fetal weight standards of the INTERGROWTH-21^st Project[J]. Ultrasound Obstet Gynecol, 2017, 49(4): 478-486.
20	KALAFAT E, LAORETI A, KHALIL A, et al. Ophthalmic artery Doppler for prediction of pre-eclampsia: systematic review and meta-analysis[J]. Ultrasound Obstet Gynecol, 2018, 51(6): 731-737.
21	CILOGLU E, OKCU N T, DOGAN N Ç. Optical coherence tomography angiography findings in preeclampsia[J]. Eye, 2019, 33(12): 1946-1951.
22	CHAEMSAITHONG P, GIL M M, CHAIYASIT N, et al. Accuracy of placental growth factor alone or in combination with soluble fms-like tyrosine kinase-1 or maternal risk factors in detecting preeclampsia in asymptomatic women in the second and third trimesters: a systematic review and meta-analysis[J]. Am J Obstet Gynecol, 2023, 229(3): 222-247.
23	NICOLAIDES K H, SARNO M, WRIGHT A. Ophthalmic artery Doppler in the prediction of preeclampsia[J]. Am J Obstet Gynecol, 2022, 226(2S): S1098-S1101.
24	MCCARTHY F P, RYAN R M, CHAPPELL L C. Prospective biomarkers in preterm preeclampsia: a review[J]. Pregnancy Hypertens, 2018, 14: 72-78.
25	CHEUNG C Y, XU D J, CHENG C Y, et al. A deep-learning system for the assessment of cardiovascular disease risk via the measurement of retinal-vessel calibre[J]. Nat Biomed Eng, 2021, 5(6): 498-508.

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
Age/year	31.00 (29.00, 34.00)	32.00 (30.00, 34.00)	0.307	32.00 (29.00, 35.00)	0.153	0.234
Pre-pregnancy BMI/(kg·m^-2)	22.43 (20.32, 25.15)	22.72 (21.04, 27.29)	0.194	23.87 (20.77, 26.66)	0.013	0.025
Previous PE history/n(%)	5 (0.73)	1 (2.78)	0.265	0 (0)	1.000	0.317
Primiparity/n(%)	493 (71.97)	29 (80.56)	0.339	54 (79.41)	0.203	0.270
ART/n(%)	51 (7.45)	2 (5.56)	1.000	12 (17.65)	0.009	0.021
CH/n(%)	14 (2.04)	1 (2.78)	0.540	8 (11.76)	0.000	0.000
Renal disease/n(%)	3 (0.44)	0 (0)	1.000	2 (2.94)	0.067	0.080
AD/n(%)	29 (4.23)	2 (5.56)	0.664	5 (7.35)	0.222	0.318
PGDM/n(%)	4 (0.58)	1 (2.78)	0.227	3 (4.41)	0.019	0.016
GDM/n(%)	76 (11.09)	7 (19.44)	0.173	9 (13.24)	0.550	0.252
PCOS/n(%)	10 (1.46)	1 (2.78)	0.443	0 (0)	0.612	0.436
Second-trimester MAP/mmHg	82.67 (76.33, 88.33)	88.17 (85.67, 96.33)	0.000	91.33 (84.67, 95.33)	0.000	0.000
Second-trimester EFW/g	1 445.89 (1 320.37, 1 566.41)	1 443.56 (1 341.89, 1 583.57)	0.604	1 347.73 (1 218.09, 1 511.01)	0.003	0.009
Delivery method/n(%)			0.167		0.000	0.000
Spontaneous delivery	393 (57.37)	16 (44.44)		20 (29.41)
Caesarean section	292 (42.63)	20 (55.56)		48 (70.59)
Gestational weeks/week	39.43 (38.57, 40.14)	39.00 (38.29, 39.71)	0.028	37.64 (36.04, 39.00)	0.000	0.000
PPH/n(%)	19 (2.77)	1 (2.78)	1.000	4 (5.88)	0.146	0.283
Postpartum hospitalization days/n(%)	3.00 (3.00, 4.00)	4.00 (3.00, 5.00)	0.001	6.00 (5.00, 7.00)	0.000	0.000
Fetal birth weight/g	3 380.00 (3 115.00, 3 670.00)	3 407.50 (3 135.00, 3 655.00)	0.725	3 005.00 (2 446.25, 3 292.50)	0.000	0.000
1 min Apgar score/score	9.00 (9.00, 10.00)	9.00 (9.00, 10.00)	0.964	9.00 (9.00, 9.00)	0.000	0.000
5 min Apgar score/score	10.00 (10.00, 10.00)	10.0 (10.00, 10.00)	0.029	10.00 (9.00, 10.00)	0.000	0.000

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
Age/year	31.00 (29.00, 34.00)	32.00 (30.00, 34.00)	0.307	32.00 (29.00, 35.00)	0.153	0.234
Pre-pregnancy BMI/(kg·m^-2)	22.43 (20.32, 25.15)	22.72 (21.04, 27.29)	0.194	23.87 (20.77, 26.66)	0.013	0.025
Previous PE history/n(%)	5 (0.73)	1 (2.78)	0.265	0 (0)	1.000	0.317
Primiparity/n(%)	493 (71.97)	29 (80.56)	0.339	54 (79.41)	0.203	0.270
ART/n(%)	51 (7.45)	2 (5.56)	1.000	12 (17.65)	0.009	0.021
CH/n(%)	14 (2.04)	1 (2.78)	0.540	8 (11.76)	0.000	0.000
Renal disease/n(%)	3 (0.44)	0 (0)	1.000	2 (2.94)	0.067	0.080
AD/n(%)	29 (4.23)	2 (5.56)	0.664	5 (7.35)	0.222	0.318
PGDM/n(%)	4 (0.58)	1 (2.78)	0.227	3 (4.41)	0.019	0.016
GDM/n(%)	76 (11.09)	7 (19.44)	0.173	9 (13.24)	0.550	0.252
PCOS/n(%)	10 (1.46)	1 (2.78)	0.443	0 (0)	0.612	0.436
Second-trimester MAP/mmHg	82.67 (76.33, 88.33)	88.17 (85.67, 96.33)	0.000	91.33 (84.67, 95.33)	0.000	0.000
Second-trimester EFW/g	1 445.89 (1 320.37, 1 566.41)	1 443.56 (1 341.89, 1 583.57)	0.604	1 347.73 (1 218.09, 1 511.01)	0.003	0.009
Delivery method/n(%)			0.167		0.000	0.000
Spontaneous delivery	393 (57.37)	16 (44.44)		20 (29.41)
Caesarean section	292 (42.63)	20 (55.56)		48 (70.59)
Gestational weeks/week	39.43 (38.57, 40.14)	39.00 (38.29, 39.71)	0.028	37.64 (36.04, 39.00)	0.000	0.000
PPH/n(%)	19 (2.77)	1 (2.78)	1.000	4 (5.88)	0.146	0.283
Postpartum hospitalization days/n(%)	3.00 (3.00, 4.00)	4.00 (3.00, 5.00)	0.001	6.00 (5.00, 7.00)	0.000	0.000
Fetal birth weight/g	3 380.00 (3 115.00, 3 670.00)	3 407.50 (3 135.00, 3 655.00)	0.725	3 005.00 (2 446.25, 3 292.50)	0.000	0.000
1 min Apgar score/score	9.00 (9.00, 10.00)	9.00 (9.00, 10.00)	0.964	9.00 (9.00, 9.00)	0.000	0.000
5 min Apgar score/score	10.00 (10.00, 10.00)	10.0 (10.00, 10.00)	0.029	10.00 (9.00, 10.00)	0.000	0.000

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

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

Characteristic	Univariate Logistic analysis			Multivariate Logistic analysis
Characteristic	OR	95% CI	P value	aOR	95% CI	P value
Second-trimester MAP	1.110	1.076‒1.148	0.000	1.106	1.068‒1.147	0.000
Second-trimester EFW	0.702	0.571‒0.869	0.001	0.700	0.560‒0.870	0.000
CRAE	0.936	0.904‒0.965	0.000	0.940	0.910‒0.970	0.000
CRVE	0.972	0.952‒0.993	0.009	0.970	0.950‒0.990	0.010
AVR	0.010	0.000‒0.288	0.008	0.020	0.000‒0.500	0.020
Retinal artery tortuosity	0.000	0.000‒0.012	0.012	0.000	0.000‒0.050	0.020
Retinal artery fractal dimension	0.046	0.006‒0.393	0.004	0.070	0.010‒0.720	0.020