Objective To investigate the risk factor of hepatocellular carcinoma (HCC) with vessels encapsulating tumor clusters (VETC) and the application value of its risk scoring model.

Methods The retrospective cross‑sectional study was conducted. The clinicopathological data of 149 patients with HCC who were admitted to two medical centers, including 97 cases in the Jiangnan University Medical Center and 52 cases in the Affiliated Xingtai People′s Hospital of Hebei Medical University, from January 2017 to April 2020 were collected. There were 116 males and 33 females, aged (58±12)years. There were 74 cases with VETC and 75 cases without VETC. Observation indica-tors: (1) clinical characteristics of patients with and without VETC; (2) imaging features of patients with and without VETC; (3) multivariable analysis of HCC patients with VETC; (4) construction of VETC related risk scoring model and its performance evaluation; (5) postoperative early tumor recurrence of patients with and without VETC who were confirmed by risk scoring model and histopathological examination. Measurement data with normal distribution were represented as Mean±SD, and comparison between groups was conducted using the t test. Count data were described as absolutes, and comparison between groups was conducted using the chi‑square test and continuous correction chi-square test. Variables of clinical and imaging characteristics with statistically signifi-cant were included in the multivariate analysis. Multivariate analysis was conducted using the Logistic regression model of backward stepwise selection. VETC related risk scoring model was constructed based on the results of Logistic regression model. The receiver operating characteristic (ROC) curve was drawn, and the area under curve (AUC), the sensitivity, specificity, accuracy and their 95% confidence interval (CI) were calculated. The maximizing Youden index was the optimal cutoff value for VETC prediction. The Hosmer Lemeshow goodness of fit test was used to assess the consistency between VETC risk scoring model predicted VTEC status and the true VETC status. The Kaplan‑Meier method was used to calculate survival rates and draw survival curves. The Log‑rank test was used for survival analysis.

Results (1) Clinical characteristics of patients with and without VETC. Cases with postoperative albumin <36 g/L were 57 in patients with VETC, versus 68 in patients without VETC, respectively, showing a significant difference between them (χ²=5.13, P<0.05). (2) Imaging features of patients with and without VETC. Cases with lesion imaging presence as nonperipheral washout, cases with lesion imaging presence as mosaic architecture, cases with lesion imaging presence as intratumoral hemorrhage, cases with lesion imaging presence as corona enhancement, cases with lesion imaging presence as non‑smooth tumor margin, cases with lesion imaging presence as peritumoral enhancement in arterial phase, cases with lesion imaging presence as intratumoral arteries, cases with lesion imaging presence as peritumoral hypointensity in hepatobiliary phase, cases with lesion imaging enhancement type as uniform low enhancement, uniform high enhance-ment, heterogeneous enhancement with septations and heterogeneous enhancement with irregular ring-like structures, cases with intratumoral necrosis or ischemic, cases with tumor diameter >5 cm were 73, 35, 33, 15, 39, 28, 42, 27, 4, 5, 27, 38, 45, 46 in patients with VETC, versus 64, 16, 13, 3, 19, 15, 9, 13, 9, 35, 5, 26, 10, 10 in patients without VETC, respectively, showing significant differences in the above indicators between them (χ²=8.92, 11.15, 12.97, 9.28, 11.74, 5.77, 33.14, 6.96, 41.79, 36.05, 37.86, P<0.05). (3) Multivariable analysis of patients with VETC. Results of multivariable analysis showed that lesion imaging enhancement as heterogeneous enhancement with septations, lesion imaging enhancement as heterogeneous enhancement with irregular ring‑like structures, intratumoral necrosis or ischemia and tumor diameter >5 cm were independent risk factors influen-cing patients with VETC (odds ratio=4.18, 7.62, 4.23, 4.08, 95%CI as 1.60‒11.60, 2.00‒31.70, 1.71‒10.90, 1.60‒10.80), P<0.05). (4) Construction of VETC related risk scoring model and its performance evaluation. The VETC related risk scoring model was constructed as (heterogeneous enhancement with septations, presence: 1.0, absence: 0)+(heterogeneous enhancement with irregular ring‑like structures, presence: 1.5, absence: 0)+(intratumoral necrosis or ischemia, presence: 1.0, absence: 0)+(main tumor diameter >5 cm, presence: 1.0, absence: 0). The AUC, sensitivity, specificity, and accuracy of VETC related risk scoring model were 0.86 (95%CI as 0.80‒0.92), 79.7% (95%CI as 69.2%‒87.3%), 80.0% (95%CI as 69.6%‒87.5%) and 79.9% (95%CI as 72.7%‒85.5%), respectively. Results of Hosmer⁃Lemeshow goodness of fit test showed a good consistency between VETC risk scoring model predicted VETC status and true VETC status (P>0.05). (5) Postoperative early tumor recurrence of patients with and without VETC who were confirmed by risk scoring model and histopathological examination. All 149 patients were followed up for 29(range, 26‒35)months. The time to tumor recurrence and 2‑year cumulative tumor recurrence rate of 149 patients were 29(range, 24‒33)months and 43.0%, respectively. The 2‑year tumor cumulative recurrence rate of patients with and without VETC predicted by risk scoring model was 47.8% and 37.9%, respectively, showing a significant difference between (χ²=3.90, P<0.05). The 2‑year cumulative tumor recurrence rate of patients with and without VETC confirmed by postoperative histopathological examination was 47.4% and 38.1%, respectively, showing a significant difference between (χ²=4.20, P<0.05).

Conclusions Lesion imaging enhancement as heterogeneous enhancement with septations or irregular ring-like structures, intratumoral necrosis or ischemia and tumor diameter >5 cm are independent risk factors influen-cing HCC patients with VETC. The proposed risk scoring model based on those three risk factors achieves an optimal preoperative diagnostic performance.