Construction of a community-based primary screening and hospital-based confirmatory screening pathway in pediatric nonalcoholic fatty liver disease

Table 1 Baseline characteristics of the study population, n (%)

Variables	Total (n = 222)	F0-F1 (n = 109)	≥ F2 (n = 113)	P value1
Male	195 (87.8)	94 (86.2)	101 (89.4)	0.610
Age (years)	11.6 (3.23)	12.0 (3.75)	11.2 (2.59)	0.371
Weight (kg)	61.6 (21.2)	61.6 (21.4)	61.5 (21.0)	0.941
BMI	25.0 (4.72)	24.6 (4.18)	25.4 (5.17)	0.416
BMI-z	2.25 (1.13)	2.11 (0.89)	2.39 (1.30)	0.004
G	1.46 (0.56)	1.18 (0.41)	1.73 (0.56)	< 0.001
S	1.68 (0.92)	0.85 (0.36)	2.47 (0.50)	< 0.001
SBP (mmHg)	120 (13.2)	120 (12.4)	120 (13.9)	0.865
DBP (mmHg)	72.7 (7.82)	73.3 (8.51)	72.2 (7.08)	0.266
ALT (U/L)	190 (139)	170 (138)	209 (138)	0.013
FPG (mmol/L)	4.95 (0.80)	4.88 (0.58)	5.01 (0.97)	0.888
Insulin (mU/L)	30.0 (17.7)	21.3 (12.9)	37.7 (18.0)	< 0.001
TC (mmol/L)	4.45 (1.02)	4.44 (1.11)	4.46 (0.92)	0.465
TG (mmol/L)	1.62 (0.82)	1.63 (0.90)	1.61 (0.75)	0.702
PLT (× 109/L)	282 (63.3)	279 (70.8)	286 (55.3)	0.305
Creatinine (μmol/L)	56.7 (13.5)	60.4 (16.4)	53.0 (8.51)	0.005
HDL (mmol/L)	1.14 (0.25)	1.18 (0.24)	1.10 (0.25)	0.034
LDL (mmol/L)	3.01 (0.80)	2.98 (0.88)	3.03 (0.72)	0.377
Urea (mmol/L)	3.86 (1.06)	4.01 (1.14)	3.73 (0.95)	0.068
UA (μmol/L)	380 (114)	397 (113)	365 (113)	0.053
HGB (g/L)	137 (13.2)	140 (13.4)	135 (12.5)	0.041
AST (U/L)	112 (89.8)	90.8 (84.4)	132 (90.6)	< 0.001
ALP (U/L)	290 (132)	255 (122)	323 (132)	< 0.001
GGT (U/L)	85.7 (80.2)	78.6 (84.6)	92.7 (75.4)	0.003
Albumin (g/L)	45.2 (4.62)	45.6 (4.20)	44.7 (4.98)	0.048
Prothrombin activity	100 (15.0)	102 (15.7)	98.1 (14.0)	0.052
Fibrinogen (g/L)	3.04 (0.816)	2.90 (0.809)	3.19 (0.802)	0.027
Lipoprotein a (mg/L)	90.3 (105)	89.6 (101)	91.1 (109)	0.932
ApoB (g/L)	0.839 (0.237)	0.832 (0.252)	0.846 (0.224)	0.614
ApoA1 (g/L)	1.27 (0.231)	1.29 (0.232)	1.25 (0.231)	0.203
Globulin (g/L)	27.7 (5.99)	27.2 (6.50)	28.1 (5.45)	0.082
Prealbumin (mg/L)	232 (52.5)	248 (50.5)	217 (50.0)	< 0.001
Direct bilirubin (μmol/L)	4.44 (5.19)	4.01 (2.11)	4.84 (6.97)	0.581
Total bilirubin (μmol/L)	11.4 (9.78)	11.3 (8.41)	11.5 (11.0)	0.407
TBA (μmol/L)	8.32 (11.8)	5.71 (5.53)	10.8 (15.2)	< 0.001
Cholinesterase (U/L)	9610 (2010)	9750 (1820)	9470 (2170)	< 0.001
INR	0.96 (0.06)	0.95 (0.06)	0.97 (0.06)	< 0.001
B-AST	270 (291)	206 (242)	332 (321)	< 0.001
APRI	1.06 (1.00)	0.899 (1.04)	1.22 (0.940)	< 0.001
M-APRI	2.48 (3.19)	2.03 (3.06)	2.92 (3.27)	0.001
FIB-4	0.34 (0.21)	0.30 (0.18)	0.37 (0.23)	0.027
M-FIB-4	0.74 (0.70)	0.64 (0.46)	0.84 (0.86)	0.007
Forns	-0.55 (1.57)	-0.48 (1.76)	-0.62 (1.37)	0.483
HSI	41.9 (25.1)	44.8 (34.9)	39.2 (7.06)	0.126
NFS	-4.82 (1.21)	-4.93 (1.18)	-4.71 (1.23)	0.179
PNFS	-1.79 (1.84)	-1.93 (1.85)	-1.66 (1.82)	0.112

¹Student’s t test (normal data) and the Kruskal-Wallis test (nonnormal data) were used for the continuous variables, and the χ² test was used for the categorical variables.

ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; ApoB: Apolipoprotein B; ApoA1: Apolipoprotein A1; AST: Aspartate aminotransferase; DBP: Diastolic blood pressure; FPG: Fasting plasma glucose; GGT: Glutamyl transpeptidase; HDL: High-density lipoprotein cholesterol; HGB: Hemoglobin; INR: International normalized ratio; LDL: Low-density lipoprotein cholesterol; PLT: Platelets; SBP: Systolic blood pressure; TBA: Total bile acid; TC: Total cholesterol; TG: Triglycerides; UA: Uric acid.

Table 2 Diagnostic performances of the machine learning models based on the ATS index for the diagnosis of ≥ F2 in the training and validation sets

	AUC (95%CI)	P value1	Sensitivity	Specificity	Accuracy	PPV	NPV	Kappa
Training set
LR	0.81 (0.75, 0.87)	Reference	0.64	0.88	0.75	0.86	0.68	0.51
DT	0.91 (0.87, 0.95)	0.001	0.81	0.84	0.83	0.85	0.80	0.65
SVM	0.88 (0.83, 0.93)	< 0.001	0.82	0.78	0.80	0.81	0.79	0.60
RF	1.00 (1.00, 1.00)	< 0.001	0.98	1.00	0.99	1.00	0.98	0.98
ANN	0.97 (0.95, 0.99)	< 0.001	0.92	0.93	0.92	0.94	0.91	0.84
XGBoost	1.00 (0.99, 1.00)	< 0.001	0.95	0.99	0.97	0.99	0.94	0.93
Internal validation set
LR	0.70 (0.54, 0.85)	Reference	0.53	0.77	0.67	0.63	0.69	0.30
DT	0.68 (0.51, 0.83)	0.861	0.63	0.62	0.62	0.55	0.70	0.24
SVM	0.79 (0.65, 0.92)	0.037	0.90	0.65	0.76	0.65	0.90	0.52
RF	0.91 (0.82, 0.98)	0.003	0.79	0.89	0.84	0.83	0.85	0.68
ANN	0.67 (0.51, 0.82)	0.767	0.84	0.58	0.69	0.59	0.83	0.40
XGBoost	0.82 (0.69, 0.93)	0.117	0.74	0.77	0.76	0.70	0.80	0.50

¹Delong test (two-sided 5% significance level).

The ATS index is a logistic regression model developed in this study that includes six parameters: Alkaline phosphatase, total bile acid, aspartate aminotransferase, cholinesterase, high-density lipoprotein cholesterol, and fibrinogen. LR: Logistic regression; DT: Decision tree; SVM: Support vector machine; RF: Random forest; ANN: Artificial neural network; XGBoost: EXtreme gradient boosting; AUC: Area under the curve; PPV: Positive predictive; NPV: Negative predictive value.

Table 3 Diagnostic performances of the machine learning models based on the BIU index for the diagnosis of ≥ F2 in the training and validation sets

	AUC (95%CI)	P value1	Sensitivity	Specificity	Accuracy	PPV	NPV	Kappa
Training set
LR	0.81 (0.71, 0.90)	Reference	0.84	0.74	0.80	0.81	0.79	0.59
DT	0.89 (0.82, 0.95)	0.547	0.80	0.83	0.81	0.86	0.76	0.62
SVM	0.98 (0.95, 1.00)	< 0.001	0.91	0.91	0.91	0.93	0.89	0.82
RF	1.00 (1.00, 1.00)	< 0.001	1.00	1.00	1.00	1.00	1.00	1.00
ANN	0.82 (0.73, 0.91)	0.534	0.84	0.80	0.83	0.84	0.80	0.64
XGBoost	1.00 (0.99, 1.00)	< 0.001	0.98	1.00	0.99	1.00	0.97	0.98
Internal validation set
LR	0.88 (0.70, 1.00)	Reference	0.82	0.70	0.76	0.75	0.78	0.52
DT	0.80 (0.61, 0.96)	0.187	0.73	0.80	0.76	0.80	0.73	0.53
SVM	0.95 (0.84, 1.00)	0.377	0.91	0.60	0.76	0.71	0.86	0.52
RF	0.86 (0.67, 0.99)	0.841	1.00	0.60	0.81	0.73	1.00	0.61
ANN	0.90 (0.76, 1.00)	0.762	0.91	0.80	0.86	0.83	0.89	0.71
XGBoost	0.75 (0.50, 0.95)	0.325	0.82	0.60	0.71	0.69	0.75	0.42

¹Delong test (two-sided 5% significance level).

The BIU index is a logistic regression model developed in this study that includes three parameters: Body mass index, insulin, and uric acid. LR: Logistic regression; DT: Decision tree; SVM: Support vector machine; RF: Random forest; ANN: Artificial neural network; XGBoost: EXtreme gradient boosting; AUC: Area under the curve; PPV: Positive predictive; NPV: Negative predictive value.

Table 4 Diagnostic performance of serial tests of the BIU and ATS indices in predicting significant fibrosis

Classification					Sensitivity	Specificity	Accuracy	PPV	NPV
Both BIU and ATS at 90%, sensitivity			LR (ATS)		1.00	0.04	0.58	0.57	1.00
			+	-
	LR (BIU)	+	70	11
		-	19	2
Both BIU and ATS at 90%, specificity			LR (ATS)		0.28	0.98	0.59	0.94	0.52
			+	-
	LR (BIU)	+	17	19
		-	16	50
Both BIU and ATS at 90%, sensitivity			RF (ATS)		0.95	0.31	0.67	0.64	0.82
			+	-
	LR (BIU)	+	42	39
		-	4	17
Both BIU and ATS at 90%, specificity			RF (ATS)		0.58	0.98	0.76	0.97	0.65
			+	-
	LR (BIU)	+	34	2
		-	32	34
Both BIU and ATS at 90%, sensitivity			RF (ATS)
			+	-
	RF (BIU)	+	41	11
		-	5	45	0.95	0.93	0.94	0.95	0.93
Both BIU and ATS at 90%, specificity			RF (ATS)
			+	-
	RF (BIU)	+	58	5
		-	8	31	0.96	0.93	0.95	0.95	0.94

The BIU index is a logistic regression model developed in this study that includes three parameters: Body mass index, insulin, and uric acid. The ATS index is a logistic regression model developed in this study that includes six parameters: Alkaline phosphatase, total bile acid, aspartate aminotransferase, cholinesterase, high-density lipoprotein cholesterol, and fibrinogen. PPV: Positive predictive value; NPV: Negative predictive value.