We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Analysis of Sonazoid contrast-enhanced ultrasound for predicting the risk of microvascular invasion in hepatocellular carcinoma: a prospective multicenter study.
- Authors
Yao, Jundong; Li, Kaiyan; Yang, Hong; Lu, Shichun; Ding, Hong; Luo, Yan; Li, Kai; Xie, Xiaoyan; Wu, Wei; Jing, Xiang; Liu, Fangyi; Yu, Jie; Cheng, Zhigang; Tan, Shuilian; Dou, Jianping; Dong, XueJuan; Wang, Shuo; Zhang, Yiqiong; Li, Yunlin; Qi, Erpeng
- Abstract
Objectives: The aim of this study was to evaluate the potential of Sonazoid contrast-enhanced ultrasound (SNZ-CEUS) as an imaging biomarker for preoperative prediction of microvascular invasion (MVI) in hepatocellular carcinoma (HCC). Methods: From August 2020 to March 2021, we conducted a prospective multicenter study on the clinical application of Sonazoid in liver tumor; a MVI prediction model was developed and validated by integrating clinical and imaging variables. Multivariate logistic regression analysis was used to establish the MVI prediction model; three models were developed: a clinical model, a SNZ-CEUS model, and a combined model and conduct external validation. We conducted subgroup analysis to investigate the performance of the SNZ-CEUS model in non-invasive prediction of MVI. Results: Overall, 211 patients were evaluated. All patients were split into derivation (n = 170) and external validation (n = 41) cohorts. Patients who had MVI accounted for 89 of 211 (42.2%) patients. Multivariate analysis revealed that tumor size (> 49.2 mm), pathology differentiation, arterial phase heterogeneous enhancement pattern, non-single nodular gross morphology, washout time (< 90 s), and gray value ratio (≤ 0.50) were significantly associated with MVI. Combining these factors, the area under the receiver operating characteristic (AUROC) of the combined model in the derivation and external validation cohorts was 0.859 (95% confidence interval (CI): 0.803–0.914) and 0.812 (95% CI: 0.691–0.915), respectively. In subgroup analysis, the AUROC of the SNZ-CEUS model in diameter ≤ 30 mm and ˃ 30 mm cohorts were 0.819 (95% CI: 0.698–0.941) and 0.747 (95% CI: 0.670–0.824). Conclusions: Our model predicted the risk of MVI in HCC patients with high accuracy preoperatively. Clinical relevance statement: Sonazoid, a novel second-generation ultrasound contrast agent, can accumulate in the endothelial network and form a unique Kupffer phase in liver imaging. The preoperative non-invasive prediction model based on Sonazoid for MVI is helpful for clinicians to make individualized treatment decisions. Key Points: • This is the first prospective multicenter study to analyze the possibility of SNZ-CEUS preoperatively predicting MVI. • The model established by combining SNZ-CEUS image features and clinical features has high predictive performance in both derivation cohort and external validation cohort. • The findings can help clinicians predict MVI in HCC patients before surgery and provide a basis for optimizing surgical management and monitoring strategies for HCC patients.
- Subjects
CLINICAL prediction rules; CONTRAST-enhanced ultrasound; ULTRASOUND contrast media; LONGITUDINAL method; RECEIVER operating characteristic curves; LOGISTIC regression analysis
- Publication
European Radiology, 2023, Vol 33, Issue 10, p7066
- ISSN
0938-7994
- Publication type
Article
- DOI
10.1007/s00330-023-09656-3