Prediction of Microvascular Invasion and Recurrence After Curative Resection of LI-RADS Category 5 Hepatocellular Carcinoma on Gd-BOPTA Enhanced MRI

Figures & data

Figure 1 Flowchart of the study population.

Table 1 Detailed Sequences and Parameters of Gd-BOPTA Enhanced MRI

Sequences	TR/TE (msec)	FOV (mm)	Flip Angle	Thickness (mm)	Matrix
T1WI	190/4.3(2)	420 × 420	80	6	256 × 160
T2WI	6667/85	420 × 420	160	6	320 × 224
DCE	3.7/1.7	420 × 420	15	2.5	256 × 192
HBP	3.7/1.7	420 × 420	15	2.5	256 × 192
2D MRCP	4000/847	320 × 320	160	50	320 × 256

Abbreviations: DCE, dynamic contrast-enhanced; HBP, hepatobiliary phase.

Table 2 Clinicopathologic Characteristics of the Study Population

Characteristic	Total (n=132)	MVI Positive (n=43)	MVI Negative (n=89)	p value	Recurrence (n=62)	Non-recurrence (n=70)	p value
Age (years)	55.5 ± 10.9	56.2 ± 10.7	55.1 ± 11.0	0.602	57.2 ± 10.5	53.9 ± 11.1	0.078
Gender				0.561			0.876
Male	110 (83.3)	37 (86)	73 (82)		52 (83.9)	58 (82.9)
Female	22 (16.7)	6 (14)	16 (18)		10 (16.1)	12 (17.1)
HBsAg				0.790			0.244
Absent	20 (15.2)	6 (14)	14 (15.7)		7 (11.3)	13 (18.6)
Present	112 (84.8)	37 (86)	75 (84.3)		55 (88.7)	57 (81.4)
Cirrhosis				0.293			0.036
Absent	48 (36.4)	17 (39.5)	27 (30.3)		15 (24.2)	29 (41.4)
Present	84 (63.6)	26 (60.5)	62 (69.7)		47 (75.8)	41 (58.6)
Serum AFP level				0.357			0.124
≤20 ng/mL	69 (52.3)	20 (46.5)	49 (55.1)		28 (45.2)	41 (58.6)
>20 ng/mL	63 (47.7)	23 (53.5)	40 (44.9)		34 (54.8)	29 (41.4)
Serum PT level	12.1 ± 1.0	12.0 ± 0.7	12.1 ± 1.1	0.540	12.3 ± 1.0	11.9 ± 1.0	0.069
Serum ALT level	27 (19, 39)	30 (20.7, 55.3)	27 (19, 38)	0.208	29 (19.5, 40)	27 (19, 39)	0.421
Serum AST level	27 (20, 35)	30.5 (23, 45)	26 (19, 32.5)	0.014	29 (21, 36)	25 (19, 33.3)	0.096
Serum GGT level	44 (26, 99)	53 (30.8, 110.5)	42 (24, 72)	0.076	49 (29.5, 109)	36 (24, 77)	0.049
Serum ALB level	42.3 (39.5, 45.1)	41.7 (39.6, 44.5)	42.8 (39.4, 45.5)	0.490	42.6 (39.2, 45.2)	42.6 (39.2, 45.2)	0.749
Serum pro-ALB level	213 (168.8, 271)	200 (169, 250)	217 (166, 277)	0.457	185 (155.5, 247)	232.5 (188.8, 285)	0.003
Serum TBiL level	13.7 (11.3, 18)	12.7 (9.0, 18.8)	14.3 (12.0, 17.6)	0.148	14.5 (12.0, 18.3)	13.4 (11.2, 17.6)	0.390
Serum dBiL level	5.1 (4.0, 6.9)	4.8 (3.5, 6.9)	5.3 (4.2, 6.7)	0.315	5.7 (3.9, 6.8)	5.1 (4.1, 6.9)	0.588
Edmondson-Steiner grade				0.529			0.146
I–II	19 (14.4)	5 (11.6)	14 (15.7)		6 (9.7)	13 (18.6)
III–IV	113 (85.6)	38 (88.4)	75 (84.3)		56 (90.3)	57 (81.4)

Note: p < 0.05 was defined as statistical significance (bold).

Abbreviations: AFP, alpha-fetoprotein; PT, prothrombin time; ALT, alanine transaminase; AST, aspartate transaminase; GGT, γ-glutamyl transpeptidase; ALB, albumin; pro-ALB, pro-albumin; Tbil, total bilirubin; dTbil, direct bilirubin.

Table 3 Univariate and Multivariate Logistic Analyses for MVI with Patients of LR-5 HCC

Characteristic	Univariate Analysis			Multivariate Analysis
	OR	95% CI	p value	OR	95% CI	p value
Tumor size >5cm	1.919	0.872, 4.222	0.105
Multiple tumors	1.027	0.029, 4.603	0.972
LI-RADS major features
Non-rim arterial phase hyperenhancement
Non-peripheral washout	2.500	0.283, 22.088	0.410
Enhancing capsule	1.048	0.489, 2.245	0.904
LI-RADS ancillary features (favoring HCC in particular)
Non-enhancing capsule	1.536	0.410, 4.360	0.420
Nodule-in-nodule architecture	0.600	0.204, 1.764	0.354
Mosaic architecture	2.837	1.328, 6.063	0.007
Fat in mass, more than adjacent liver	0.716	0.276, 1.862	0.494
Blood products in mass	2.149	0.833, 5.546	0.114
LI-RADS ancillary features (favoring malignancy, not HCC in particular)
Hepatobiliary phase hypointensity	0.646	0.311, 1.345	0.243
Mild-moderate T2 hyperintensity	2.500	0.283, 22.088	0.410
Corona enhancement	4.658	2.139, 10.143	<0.001	3.217	1.326, 7.861	0.010
Restricted diffusion	0.233	0.021, 2.643	0.240
Fat sparing in solid mass	1.969	0.663, 5.842	0.222
Iron sparing in solid mass	1.036	0.091, 11.747	0.977
Non-LIRADS imaging features
Non-smooth tumor margin	3.421	1.596, 7.333	0.002
Internal arteries	6.125	2.647, 14.170	<0.001	4.147	1.596, 10.880	0.004
Peritumoral hypointensity on HBP	8.193	3.550, 18.908	<0.001	5.165	2.064, 12.923	<0.001

Note: p < 0.05 was defined as statistical significance (bold).

Abbreviations: HCC, hepatocellular carcinoma; LIRADS, The Liver Imaging Reporting and Data System; HBP, hepatobiliary phase; CI, confidence interval; OR, odds ratio.

Table 4 Univariate and Multivariate Cox Analyses for Recurrence with Patients of LR-5 HCC

Characteristic	Univariate Analysis			Multivariate Analysis
	HR	95% CI	p value	HR	95% CI	p value
Tumor size >5cm	1.831	1.087, 3.083	0.023
Multiple tumors	2.587	1.144, 5.849	0.022
LI-RADS major features
Non-rim arterial phase hyperenhancement
Non-peripheral washout	0.656	0.238, 1.808	0.415
Enhancing capsule	1.213	0.712, 2.067	0.477
LI-RADS ancillary features (favoring HCC in particular)
Non-enhancing capsule	1.109	0.528, 2.331	0.785
Nodule-in-nodule architecture	2.021	1.113, 3.670	0.021
Mosaic architecture	3.042	1.841, 5.026	<0.001	1.982	1.147, 3.427	0.014
Fat in mass, more than adjacent liver	1.036	0.562, 1.911	0.909
Blood products in mass	0.784	0.373, 1.649	0.521
LI-RADS ancillary features (favoring malignancy, not HCC in particular)
Hepatobiliary phase hypointensity	0.709	0.430, 1.168	0.177
Mild-moderate T2 hyperintensity	1.714	0.419, 7.013	0.454
Corona enhancement	2.816	1.705, 4.652	<0.001	1.783	1.031, 3.082	0.039
Restricted diffusion	1.949	0.270, 14.069	0.508
Fat sparing in solid mass	0.443	0.161, 1.220	0.115
Iron sparing in solid mass	1.789	0.437, 7.324	0.419
Non-LIRADS imaging features
Non-smooth tumor margin	1.487	0.900, 2.458	0.121
Internal arteries	1.727	1.020, 2.923	0.042
Peritumoral hypointensity on HBP	3.492	2.108, 5.786	<0.001	2.130	1.208, 3.755	0.009

Note: p < 0.05 was defined as statistical significance (bold).

Abbreviations: HCC, hepatocellular carcinoma; LIRADS, The Liver Imaging Reporting and Data System; HBP, hepatobiliary phase; CI, confidence interval; HR, hazard ratios.

Figure 2 Kaplan–Meier curves of imaging features for RFS in LR-5 HCCs. (A) Mosaic architecture (absent) and Mosaic architecture (present) (Log rank test, p =0.034). (B) Corona enhancement (absent) and Corona enhancement (present) (Log rank test, p <0.0001). (C) Peritumoral hypointensity on HBP (absent) and Peritumoral hypointensity on HBP (present) (Log rank test, p <0.0001).

Table 5 Diagnostic Performance for Predicting MVI and Recurrence

MRI Findings	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)
Prediction of MVI
Corona enhancement	60.5 (26/43)	75.3 (67/89)	70.5 (93/132)	54.2 (26/48)	79.8 (67/84)
Internal arteries	51.2 (22/43)	85.4 (76/89)	74.2 (98/132)	62.9 (22/35)	78.4 (76/97)
Peritumoral hypointensity on HBP	60.5 (26/43)	84.3 (75/89)	76.5 (101/132)	65 (26/40)	81.5 (75/92)
Prediction of recurrence
Mosaic architecture	53.2 (33/62)	70 (49/70)	62.1 (82/132)	61.1 (33/54)	62.8 (49/78)
Corona enhancement	53.2 (33/62)	78.6 (55/70)	66.7 (88/132)	68.8 (33/48)	65.5 (55/84)
Peritumoral hypointensity on HBP	45.2 (28/62)	82.9 (58/70)	65.2 (86/132)	70 (28/40)	63.0 (58/92)

Abbreviations: MVI, microvascular invasion; HBP, hepatobiliary phase.

Figure 3 Images of a 54-year-old man with chronic hepatitis B infection, categorized as LR-5 HCC with MVI and recurrence. (A) Heterogeneous mass with mosaic architecture on T2WI; (B) Non-rim arterial phase hyperenhancement (APHE) on the arterial phase, with corona enhancement and internal arteries (white arrow). (C) Peritumoral hypointensity (white arrow) in the hepatobiliary phase.

Figure 4 MR images of a 48-year-old man with chronic hepatitis B infection, categorized as LR-5 HCC without MVI and non-recurrence. (A) Non-rim arterial phase hyperenhancement (APHE) on the arterial phase. (B) Non-peripheral washout on the portal venous phase. (C) The lesion of hypointense on the HBP image.