Have a personal or library account? Click to login
Evaluating the severity of microvascular invasion in hepatocellular carcinoma, by probing the combination of enhancement modes and growth patterns through magnetic resonance imaging Cover

Evaluating the severity of microvascular invasion in hepatocellular carcinoma, by probing the combination of enhancement modes and growth patterns through magnetic resonance imaging

By: Yanzhuo Li,  Sijie Li,  Yan Lei,  Lianlian Liu and  Bin Song  
Open Access
|Apr 2025

Figures & Tables

Figure 1.

A 34-year-old male with a clearly bounded solitary HCC and M0 status, (A-D) exhibits persistent enhancement in AP, PVP, DP, and a smooth tumor margin (red arrow). A 52-year-old male with HCC and M1 grade, which illustrates a typical no/mini enhancement (red arrow) (E-G). Vague margin is visible at the superior edge of the tumor (black arrowhead) (H). A 65-year-old male who had HCC with M2 status was detected. The mass has a Max/Min-R of 1.59 and heterogeneous moderate hyperintensity on T2WI (I), which displays no or minimal enhancement (red arrow) on AP, PVP, and DP(J-L). The lesion appears as extranodular nodule (black arrowhead) (L).
A 34-year-old male with a clearly bounded solitary HCC and M0 status, (A-D) exhibits persistent enhancement in AP, PVP, DP, and a smooth tumor margin (red arrow). A 52-year-old male with HCC and M1 grade, which illustrates a typical no/mini enhancement (red arrow) (E-G). Vague margin is visible at the superior edge of the tumor (black arrowhead) (H). A 65-year-old male who had HCC with M2 status was detected. The mass has a Max/Min-R of 1.59 and heterogeneous moderate hyperintensity on T2WI (I), which displays no or minimal enhancement (red arrow) on AP, PVP, and DP(J-L). The lesion appears as extranodular nodule (black arrowhead) (L).

Figure 2.

(A,B) The MVI nomogram was built by incorporating four variables, and among the MVI-positive cases, three variables were used to establish another nomogram for predicting M2 grade. (C,D) Calibration curves of the nomogram in predicting MVI and its M2 grade. X-axis, nomogram-predicted probability of MVI or M2; Y-axis, observed MVI or M2. (E,F) Nomogram model (blue line) outperforms all (red line) and none (horizontal green line) across all reasonable threshold probabilities in predicting MVI and its M2 grade. (G,H) The ROC curves demonstrate the discriminatory ability of the two nomograms of MVI positive and its M2 grade.
(A,B) The MVI nomogram was built by incorporating four variables, and among the MVI-positive cases, three variables were used to establish another nomogram for predicting M2 grade. (C,D) Calibration curves of the nomogram in predicting MVI and its M2 grade. X-axis, nomogram-predicted probability of MVI or M2; Y-axis, observed MVI or M2. (E,F) Nomogram model (blue line) outperforms all (red line) and none (horizontal green line) across all reasonable threshold probabilities in predicting MVI and its M2 grade. (G,H) The ROC curves demonstrate the discriminatory ability of the two nomograms of MVI positive and its M2 grade.

The diagnostic performance of nomogram models for MVI-positive and M2 grade in HCC

ModelAUC (95%CI)Accuracy (%)Sensitivity (%)Specificity (%)PPVNPV
MVI-positive0.855(0.833–0.937)84.088.981.174.292.2
M2 grade0.805(0.703–0.908)79.074.282.071.983.7

Univariate and multivariate logistic regression analysis for predicting M2 grade

VariableUnivariable analysisMultivariable analysis
Odds ratio (95 % CI)P-valueOdds ratio (95 % CI)P-value
Clinical features
Age (≥ 58 years)1.68 (0.68–4.14)0.262
Sex (male)1.27 (0.35–4.69)0.703
HBeAg(+)0.32 (0.09–1.05)0.060
AFP lg101.86 (0.75–4.63)0.183
PIVKA-II lg101.09 (0.64–1.88)0.745
Pathological data
Satellite nodule0.46 (0.18–1.16)0.098
MRI findings
Tumor Max-D (≥ 4.3cm)1.39 (0.56–3.42)0.480
Tumor Min-D (≥ 3.7cm)1.07 (0.44–2.61)0.888
Max/Min-R (≥ 1.22)3.53 (1.38–9.04)0.009*2.91 (1.07–7.92)0.036*
Irregular tumor shape1.32 (0.52–3.32)0.560
Non-smooth tumor margin3.43 (1.33–8.82)0.011*
Confluent multinodule growth3.29 (1.19–9.08)0.021*3.92 (1.25–12.25)0.019*
Washin/washout enhanced mode0.34 (0.13–0.88)0.026*0.31 (0.11–0.92)0.035*
Beak sign1.18 (0.48–2.91)0.721
Arterial Rim-enhancement2.15 (0.73–6.34)0.167
Peritumoral enhancement on AP1.99 (0.79–5.01)0.143
Peritumoral hypointensity on HBP1.77 (0.63–4.92)0.227

Univariate and multivariate logistic regression analysis for predicting MVI-positive HCCs

VariableUnivariable analysisMultivariable analysis
Odds ratio (95 % CI)P-valueOdds ratio (95 % CI)P-value
Clinical features
Age (≥ 57 years)1.11 (0.64–1.94)0.702
Sex (male)1.28 (0.60–2.72)0.525
Etiology (hepatitis B virus)1.02 (0.54–1.93)0.960
AFP (≥ 20ng/mL)1.98 (1.12–3.49)0.018*
PIVKA-II lg101.18 (0.84–1.66)0.342
Pathological data
Edmondson-Steiner (III-IV)2.35 (1.05–5.26)0.037*
Satellite nodule2.53 (1.43–4.48)0.001*
MRI findings
Tumor Max-D (≥ 3.9cm)2.14 (1.22–3.77)0.008*
Tumor Min-D (≥ 3.1cm)2.04 (1.16–3.59)0.014*
Irregular tumor shape4.59 (2.54–8.33)<0.001*
Non-smooth tumor margin2.93 (1.63–5.27)<0.001*
Solitary nodule growth0.18 (0.10–0.34)<0.001*0.25 (0.12–0.52)< 0.001*
No/mini enhanced mode2.76 (1.31–5.85)0.008*3.24 (1.19–8.88)0.022*
Beak sign4.07 (2.25–7.38)<0.001*
Arterial Rim-enhancement2.07 (0.97–4.42)0.060
Peritumoral enhancement on AP5.23 (2.56–10.66)<0.001*5.19 (2.15–12.53)< 0.001*
Peritumoral hypointensity on HBP11.26 (5.84–21.68)<0.001*10.74 (5.07–22.75)< 0.001*
DOI: https://doi.org/10.2478/raon-2025-0021 | Journal eISSN: 1581-3207 | Journal ISSN: 1318-2099
Language: English
Page range: 183 - 192
Submitted on: Oct 12, 2025
|
Accepted on: Jan 27, 2025
|
Published on: Apr 11, 2025
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2025 Yanzhuo Li, Sijie Li, Yan Lei, Lianlian Liu, Bin Song, published by Association of Radiology and Oncology
This work is licensed under the Creative Commons Attribution 4.0 License.