Decision-Tree Models Indicative of Microvascular Invasion on MRI Predict Survival in Patients with Hepatocellular Carcinoma Following Tumor Ablation

Figures & data

Figure 1 TTPVI and RVI are defined by a two- and three-trait tree-like decision model, respectively. TTPVI (upper row):¹³ Index lesion must include the presence of intratumoral vessels in arterial contrast phase followed by the absence of a peritumoral hypointense halo in the portal venous contrast phase. RVI (lower row):¹⁴ Index lesion must include the presence of intratumoral vessels in portal venous contrast phase, followed by the absence of a peritumoral hypointense halo in the portal venous contrast phase, and moreover, followed by the absence of a sharp tumor-liver-difference in the portal venous contrast phase. Arrows indicate the respective imaging trait. T indicated the tumoral index lesion.

Table 1 Patients, Tumor and Disease Characteristics

Demographics	iBT Cohort	cTACE/iBT Cohort	p-value
Patient characteristics
Number of patients	48	47	-
Number of completed therapies	60	51	0.68
Age, mean ± SD	70.60 ± 9.28 years	70.29 ± 9.53 years	0.84
Male:Female, % (n)	79.2%:20.8% (38:10)	76.6%:23.4% (36:11)	0.81
Target tumor characteristics
Lesions, n	89	70	-
Unifocal:Multifocal, % (n)	68.3%:31.7% (41:19)	74.5%:25.5% (38:13)	0.53
Index lesion diameter, mean ± SD	26.51 ± 10.58 mm	44.32 ± 20.36 mm	<0.01
Disease characteristics
Cirrhosis, % (n)	85.4% (41)	91.5% (43)	0.52
Etiology of cirrhosis, n (%)
Hepatitis B	2.5% (1)	2.3% (2)	0.99
Hepatitis C	26.8% (11)	20.9% (9)	0.61
Alcoholic steatohepatitis	26.8% (11)	41.9% (18)	0.17
Non-alcoholic steatohepatitis	31.7% (13)	32.6% (14)	0.99
Others	12.2% (5)	-	-
Child Pugh class, % (n)
A	82.9% (34)	97.7% (41)	0.09
B	17.1% (7)	2.3% (2)	0.09
Barcelona Clinic Liver Cancer stage, % (n)
A	85.4% (41)	46.8% (22)	<0.01
B	12.5% (6)	36.2% (17)	0.01
C	2.1% (1)	17.0% (8)	0.02
Laboratory values of liver function, mean ± SD
Albumin [g/l]	38.7 ± 6.6	38.7 ± 3.7	0.96
Bilirubin [mg/dl]	0.76 ± 0.53	0.82 ± 0.47	0.23
ALT [U/l]	38.4 ± 21.6	40.7 ± 23.6	0.63
AST [U/l]	45.9 ± 19.1	53.4 ± 36.3	0.27
γ-GT [U/l]	162.7 ± 154.0	195.0 ± 179.0	0.24
AP [U/l]	115.0 ± 60.8	106.1 ± 46.6	0.73

Note: Bold p-values indicate statistical significance in the respective statistical test (Fisher’s exact test, unpaired t-test, and Mann–Whitney U-test).

Abbreviations: iBT, interstitial brachytherapy; cTACE, conventional transarterial chemoembolization; ALT, alanine aminotransferase; AST, aspartate aminotransferase; γ-GT, gamma-glutamyl transferase; AP, alkaline phosphatase.

Table 2 Response Assessment Evaluation by Modified Response Criteria in Solid Tumors

mRECIST	8 Weeks		5 Months
	iBT (n = 51)	cTACE/iBT (n = 47)	iBT (n = 48)	cTACE/iBT (n = 42)
CR	60.8% (31)	44.7% (21)	54.2% (26)	57.2% (24)
PR	23.6% (12)	17.0% (8)	10.4% (5)	9.5% (4)
SD	7.8% (4)	10.6% (5)	6.2% (3)	11.9% (5)
PD	7.8% (4)	27.7% (13)	29.2% (14)	21.4% (9)

Abbreviations: CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; iBT, interstitial brachytherapy; cTACE, conventional transarterial chemoembolization.

Figure 2 Patient survival and tumor progression with patients treated with ablation with and without prior embolization. Overall survival (OS), progression-free survival (PFS) and time-to-progression (TTP) are depicted for patients undergoing interstitial brachytherapy (iBT, green line) and conventional transarterial chemoembolization (cTACE, blue line). Kaplan–Meier analysis reveals no significant differences in survival outcomes between both patient groups. Median survival is indicated by dashed lines.

Table 3 The Impact of Radiogenomic Venous Invasion (RVI) and Two-Trait Predictor of Venous Invasion (TTPVI) on Survival Outcomes

Outcome	Imaging Biomarker	iBT Cohort			cTACE/iBT Cohort
		Positive	Negative	p-value	Positive	Negative	p-value
OS	TTPVI	15.97	37.65	0.08	25.58	19.72	0.55
	RVI	12.40	40.37	<0.01	26.03	21.77	0.75
PFS	TTPVI	5.87	13.77	0.04	9.58	6.45	0.43
	RVI	5.87	13.17	0.04	11.50	6.23	0.70
PFSLTP	TTPVI	10.87	18.30	0.03	22.85	18.18	0.38
	RVI	8.73	17.73	<0.01	25.12	18.20	0.64
PFSIDR	TTPVI	11.30	18.77	<0.01	11.50	7.87	0.96
	RVI	7.43	14.23	0.01	14.23	7.73	0.84
TTP	TTPVI	5.97	15.03	0.03	12.10	7.73	0.27
	RVI	6.40	11.83	0.14	14.23	7.73	0.33
TTPLTP	TTPVI	N/A	N/A	0.63*	N/A	N/A	0.49*
	RVI	N/A	N/A	0.51*	N/A	N/A	0.94*
TTPIDR	TTPVI	9.60	15.67	0.05	12.20	8.00	0.88
	RVI	8.73	15.03	0.21	14.23	8.00	0.52

Notes: N/A non-assessable (*median survival was not reached; however, no significant separation of the curves was achieved). Values are depicted as medians. Bold p-values indicate statistical differences in the Log rank test (p<0.05).

Abbreviations: OS, overall survival; PFS, progression-free survival; TTP, time-to-progression; LTP, local tumor progression; IDR, intrahepatic distant recurrence; TTPVI, two-trait predictor of venous invasion; RVI, radiogenomic venous invasion.

Figure 3 The impact of RVI on patient survival and tumor progression in patients treated with ablation with and without prior embolization. Overall survival (OS), progression-free survival (PFS), and time-to-progression (TTP) are depicted for patients with a positive and negative RVI trait receiving interstitial brachytherapy (iBT, upper row) and iBT with a prior conventional transarterial chemoembolization (cTACE/iBT, lower row). For patients within the iBT group, Kaplan–Meier analysis reveals significantly poorer OS, PFS, and a trend for poorer TTP in patients that had a positive RVI trait. In patients within the cTACE/iBT group, no significant differences in survival outcomes were observed in regards of the RVI trait. Median survival is indicated by dashed lines.

Figure 4 The impact of RVI and TTPVI imaging biomarkers and MVI trait in patients within the internal and external study cohorts. Hazard ratios (HR) and 95% confidence intervals (95%-CI) and p-values were calculated for the impact of RVI, TTPVI in the internal and external study cohorts as well and microvascular invasion (MVI) in the external study cohort and depicted as colored line graphs in the respective colors of the treatment groups: Orange – TCGA cohort, green – interstitial brachytherapy (iBT) cohort, and blue – conventional transarterial chemotherapy with consecutive iBT cohort (cTACE/iBT). Triangles show indicate a significant p-value for the respective Kaplan–Meier analysis (p<0.1).

Table 4 The Confounding Effects of Clinical Parameters on the Survival Outcomes

iBT Cohort
	Overall Survival			Progression-Free Survival			Time-to-Progression
	95% CI	HR	p-value	95% CI	HR	p-value	95% CI	HR	p-value
Age	0.95–1.02	0.96	0.40	0.97–1.03	1.00	0.92	0.96–1.03	0.99	0.65
Gender	0.40–1.67	0.78	0.49	0.56–1.27	1.06	0.86	0.48–2.15	0.96	0.92
Index tumor diameter	0.98–1.04	1.01	0.44	0.97–1.03	1.00	0.93	0.95–1.02	0.99	0.44
Number of lesions	0.68–1.21	0.94	0.66	0.90–1.65	1.25	0.15	0.91–1.77	1.31	0.11
BCLC stage	0.51–2.77	1.31	0.53	0.63–2.89	1.48	0.31	0.47–2.86	1.32	0.54
RVI	1.63–6.50	3.33	<0.01	1.08–3.86	2.10	0.02	0.78–3.72	1.78	0.15
TTPVI	0.90–3.23	1.73	0.09	1.42–4.92	2.66	<0.01	1.03–4.12	2.08	0.04
0.01
cTACE/iBT Cohort
	Overall Survival			Progression-Free Survival			Time-to-Progression
	95% CI	HR	p-value	95% CI	HR	p-value	95% CI	HR	p-value
Age	093–1.00	0.97	0.07	0.97–1.04	1.00	0.93	0.96–1.03	1.00	0.89
Gender	0.40–1.98	0.84	0.66	0.55–2.23	1.07	0.86	0.44–1.84	0.86	0.68
Index tumor diameter	1.01–1.03	1.02	<0.01	0.98–1.01	1.00	0.60	0.98–1.01	1.00	0.60
Number of lesions	1.27–3.62	2.19	<0.01	0.68–1.64	1.09	0.70	0.35–1.16	0.68	0.20
BCLC stage	1.15–2.58	1.73	<0.01	0.80–1.73	1.19	0.37	0.74–1.68	1.13	0.56
RVI	0.42–1.60	0.83	0.59	0.44–1.45	0.81	0.49	0.37–1.36	0.73	0.33
TTPVI	0.48–1.77	0.91	0.76	0.42–1.33	0.74	0.30	0.39–1.33	0.71	0.27
TCGA-LIHC Cohort
	Overall Survival			Progression-Free Survival
	95% CI	HR	p-value	95% CI	HR	p-value
MVI	0.58–4.20	1.57	0.36	1.28–8.72	3.29	0.01
RVI	0.72–6.89	2.19	0.16	0.67–5.93	2.05	0.18
TTPVI	0.36–4.14	1.12	0.85	0.67–5.99	1.87	0.25

Note: Bold print p-values indicate statistical differences in the regression analyses (p<0.1).

Abbreviations: iBT, interstitial brachytherapy; cTACE, conventional transarterial chemoembolization; TCGA-LIHC, The Cancer Genome Atlas – Liver Hepatocellular Carcinoma (public external database); MVI, microvascular invasion; RVI, Radiogenomic Venous Invasion; TTPVI, Two Trait Predictor of Venous Invasion; HR, hazard ratio; 95% CI, 95% confidence interval.

Table 5 Correlation of Semantic Imaging Features with Radiogenomic Venous Invasion (RVI) and Two-Trait Predictor of Venous Invasion (TTPVI)

Semantic Imaging Feature	Imaging Biomarker	iBT Cohort		cTACE/iBT Cohort
		Pearson’s r	p-value	Pearson’s r	p-value
Vessel in lesion (art.)	TTPVI	0.97	<0.01	0.30	0.03
	RVI	0.57	<0.01	0.32	0.03
Vessel in lesion (pv.)	TTPVI	0.79	<0.01	0.43	<0.01
	RVI	0.71	<0.01	0.04	0.77
Enhancement pattern (art.)	TTPVI	−0.22	0.10	0.43	<0.01
	RVI	−0.23	0.08	0.04	0.77
Perilesional enhancement (art.)	TTPVI	0.27	0.04	0.09	0.49
	RVI	0.16	0.23	−0.02	0.90
Perilesional hypointensity (HBP)	TTPVI	0.24	0.08	0.23	0.11
	RVI	0.11	0.41	0.07	0.62
Tumor margin (ven.)	TTPVI	−0.16	0.23	−0.03	0.84
	RVI	−0.17	0.20	0.17	0.26
Tumor capsule (ven.)	TTPVI	0.13	0.32	−0.10	0.50
	RVI	0.03	0.82	−0.01	0.93
Relative tumoral enhancement (art.)	TTPVI	0.01	0.92	0.07	0.65
	RVI	−0.10	0.12	−0.23	0.12
Relative tumoral enhancement (pv.)	TTPVI	−0.16	0.23	0.07	0.61
	RVI	−0.33	0.01	−0.04	0.76
Tumor liver ratio (HBP)	TTPVI	0.01	0.98	−0.04	0.78
	RVI	0.28	0.03	−0.01	0.96
Tumor liver ratio (DWI)	TTPVI	−0.01	0.98	−0.35	0.02
	RVI	−0.31	0.02	0.07	0.67
Tumor liver ratio (ADC)	TTPVI	0.07	0.63	−0.24	0.10
	RVI	−0.04	0.74	0.11	0.44

Note: Bold p-values (and respective Pearson’s r) indicate statistical significance (p<0.05).

Abbreviations: TTPVI, two-trait predictor of venous invasion; RVI, radiogenomic venous invasion; Art, arterial phase; pv, portal venous phase; ven, venous phase; HBP, hepatobiliary phase of T1-weighted images; DWI, diffusion weighted imaging; ADC, apparent diffusion coefficient maps; iBT, interstitial brachytherapy; cTACE, conventional transarterial chemoembolization.

Segal E, Sirlin CB, Ooi C, et al. Decoding global gene expression programs in liver cancer by noninvasive imaging. Nat Biotechnol. 2007;25(6):675–680.

 PubMed Web of Science ®Google Scholar

Banerjee S, Wang DS, Kim HJ, et al. A computed tomography radiogenomic biomarker predicts microvascular invasion and clinical outcomes in hepatocellular carcinoma. Hepatology. 2015;62(3):792–800.

 PubMed Web of Science ®Google Scholar

Data Sharing Statement

Additional data are available to readers upon reasonable request to the corresponding author.