Identification of vessels encapsulating tumor clusters in solitary hepatocellular carcinoma via imaging biomarkers in preoperative contrast-enhanced magnetic resonance imaging

Abstract

Background: The value of Liver Imaging Reporting and Data System (LI-RADS) radiological features and tumor three-dimensional volumetric quantification in preoperative magnetic resonance imaging (MRI) for predicting the vessels encapsulating tumor clusters (VETC) pattern of solitary hepatocellular carcinoma (HCC) is unknown. This study aimed to assess the value of these indicators for predicting the VETC pattern of solitary HCC.

Methods: In total, 36 patients with HCC were selected from a cohort containing 126 patients for further data evaluation. VETC was evaluated by histopathologists, and the three-dimensional tumor volume (TV) was analyzed in the arterial phase (AP) and portal venous phase. LI-RADS radiological characteristics were defined on the basis of LI-RADS version 2018. Quantitative parameters were derived from multiparametric MRI data. Significant MRI biomarkers for predicting VETC-positive status in solitary HCC were ascertained via logistic regression analysis. A nomogram was accordingly constructed and evaluated in terms of discrimination, calibration, clinical utility, and accuracy.

Results: A total of 15 cases were VETC positive, while 21 cases were VETC negative. The values for nodule-in-nodule architecture, mosaic architecture, total liver volume, TV, necrosis tumor volumetric percentage, necrosis tumor burden, and tumor-to-liver signal intensity (SI) ratio on AP images were higher in VETC-positive HCCs than in VETC-negative HCCs (P<0.05). Multivariate logistic analysis indicated that necrosis tumor volumetric percentage, tumor-to-liver SI ratio on AP images, and nodule-in-nodule architecture were independent predictive factors of VETC status (P<0.05). The calibration and discrimination performance of the nomogram were good, with an area under curve of 0.942, and the prediction accuracy was a satisfactory 88.89%, indicating that the nomogram possessed potential clinical benefits.

Conclusions: Preoperative MRI features possess the potential to identify VETC pattern in solitary HCC.

Keywords: Carcinoma; biomarkers; hepatocellular; magnetic resonance imaging (MRI); nomograms.

Figure 1

Flowchart of the inclusion and exclusion of patients. HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; VETC, vessels encapsulating tumor clusters.

Figure 2

NTV measurement. (A) The PVP images of a 60-year-old male who had a solitary HCC lesion with internal necrosis in the liver left lobe. (B,C) The contours of the necrosis region were automatically identified via the “organ” task and then semiautomatically adjusted by the observer in a slice-by-slice manner to obtain a three-dimensional image of necrosis. (D) This HCC had high NTVP, which is the ratio of the NTV to the TV and was confirmed to be VETC positive by CD34 immunohistochemical staining (original magnification 100×). The VETC-positive HCC showed clusters of tumor cells bordered by a complete rim of CD34 positive endothelial cells. NTV, necrosis tumor volume; PVP, portal venous phase; HCC, hepatocellular carcinoma; NTVP, necrosis tumor volumetric percentage; TV, tumor volume; VETC, vessels encapsulating tumor clusters; CD, cluster of differentiation.

Figure 3

Representative images of quantitative parameters. One HCC lesion showed (A) arterial phase hyperenhancement in contrast to the background liver and (B) had an AP-TLSI of 1.483. (C) This HCC was confirmed as VETC positive by CD34 immunohistochemical staining (original magnification 50×). The-VETC positive HCC showed clusters of tumor cells bordered by a complete rim of CD34-positive endothelial cells. HCC, hepatocellular carcinoma; AP-TLSI, tumor-to-liver signal intensity ratio on arterial phase images; VETC, vessels encapsulating tumor clusters; CD, cluster of differentiation.

Figure 4

HCC histologically confirmed as VETC positive with a nodule-in-nodule architecture. (A-C) The fat-suppressed T2-weighted image, arterial-phase image, and portal venous-phase image of one HCC lesion with nodule in nodule architecture, respectively (white arrows). (D) HCC was confirmed as VETC positive by CD34 immunohistochemical staining (original magnification 100×). The VETC-positive HCC showed clusters of tumor cells bordered by a complete rim of CD34-positive endothelial cells. HCC, hepatocellular carcinoma; VETC, vessels encapsulating tumor clusters; CD, cluster of differentiation.

Figure 5

Nomogram for predicting the probability of VETC in a patient with solitary HCC. First, points for each predictor are assigned according to the corresponding values from the uppermost point scale. Second, the total number of points is obtained by summing up points of all predictors. Third, a vertical line is drawn down the total points to indicate the probability of VETC positivity. NTVP, necrosis tumor volumetric percentage; AP-TLSI, tumor-to-liver signal intensity ratio on arterial phase images; HCC, hepatocellular carcinoma; VETC, vessels encapsulating tumor clusters.

Figure 6

Calibration and decisions curves of the nomogram for predicting VETC-positive HCC. (A) The calibration curve showed the goodness of fit of the nomogram. The x-axis represents the nomogram-predicted probability, and the y-axis represents the actual probability of VETC-positive HCCs. Perfect prediction would correspond to the black dashed line. The red solid line represents the entire cohort (n=36), and the green solid line is the bias-corrected by bootstrapping (B =500 repetitions). A solid line closer to the diagonal dotted line indicates a better prediction ability. (B) The DCA showed that the nomogram had a higher overall net benefit across the full range of threshold probabilities. The y-axis represents the net benefit, and the x-axis represents the threshold probability. VETC, vessels encapsulating tumor clusters; HCC, hepatocellular carcinoma; DCA, decision curve analysis.