构建多参数MRI影像组学模型预测肝细胞癌组织分化程度

doi:10.3969/j.issn.1009-976X.2025.06.004

摘要/Abstract

摘要： 目的基于多参数MRI影像组学模型,评估其在术前无创预测肝细胞癌（HCC）组织学分化程度中的价值。方法回顾性分析2022年9月至2024年12月中山大学孙逸仙纪念医院206例术后病理证实为HCC的患者资料,按Edmondson-Steiner（E-S）分级分为高级别组（n=57）和低级别组（n=149）,并以7∶3比例随机分为训练集（n=144）和验证集（n=62）。临床预测因子通过单因素和多因素分析进行筛选,影像组学特征依次经单因素分析、最大相关最小冗余（mRMR）及最小绝对收缩和选择算子（LASSO）进行选择。采用Logistic回归建立预测模型,受试者工作特征曲线（ROC）和曲线下面积（AUC）评估模型性能,DeLong检验比较模型间AUC值差异,校准曲线和Hosmer-Lemeshow检验评估模型校准度。结果多参数影像组学联合模型（Comb_Rad）在训练集和验证集中的AUC分别为0.904、0.886,且在验证集上显著优于所有单一参数影像组学模型（DeLong检验：均P<0.05）。进一步整合临床指标的Rad_Clin模型表现最佳,训练集和验证集AUC分别为0.933、0.902。Rad_Clin与Comb_Rad模型在训练集中的差异具有统计学意义（P= 0.047）,且两者在训练集和验证集中均显著优于临床模型（Clin,均P<0.05）。Rad_Clin模型具有良好的拟合度（训练集P=0.905,验证集P=0.853）。结论基于多参数 MRI 的影像组学联合模型能够有效预测HCC组织分化程度,进一步融合临床指标可提升预测性能。

关键词: 肝细胞癌, 分化程度, 磁共振成像, 影像组学

Abstract: Objective To develop a multiparametric MRI-based radiomics model for the preoperative noninvasive prediction of histological differentiation in hepatocellular carcinoma （HCC）. Methods A total of 206 patients with pathologically confirmed HCC who underwent surgical resection at Sun Yat-sen Memorial Hospital between September 2022 and December 2024 were retrospectively enrolled. Patients were classified into high-grade （n=57） and low-grade （n=149） groups based on the Edmondson-Steiner grading system and were then randomly assigned to training （n=144） and validation （n=62） cohorts at a ratio of 7∶3. Clinical predictors were identified through univariate and multivariate logistic regression analysis. Radiomics features were extracted from multiparametric MRI images and selected by applying univariate logistic regression analysis, maximum relevance minimum redundancy （mRMR）, and least absolute shrinkage and selection operator （LASSO） algorithms. Predictive models were developed using logistic regression. Model performance was evaluated by the receiver operating characteristic （ROC） curves and area under the curve （AUC）. The DeLong test was employed to compare AUC values. Calibration curves and the Hosmer-Lemeshow test were used to evaluate the model's calibration. Results The combined radiomics model （Comb_Rad） achieved AUCs of 0.904 and 0.886 in the training and validation cohorts, significantly outperformed all single-parameter radiomics models in the validation cohort （all P<0.05, DeLong test）. The radiomics-clinical combined model （Rad_Clin） achieved the best performance, with AUCs of 0.933 in the training cohort and 0.902 in the validation cohort. The Rad_Clin model demonstrated significantly better performance than the Comb_Rad model in the training cohort （P=0.047）, while both models significantly outperformed the clinical model （Clin） in both cohorts （all P<0.05）. Additionally, the Rad_Clin model demonstrated good fit （training cohort P=0.905, test cohort P=0.853）. Conclusion A multiparametric MRI-based radiomics model can effectively predict the histological differentiation of HCC, and the integration of clinical factors further improves predictive performance.

Key words: hepatocellular carcinoma, differentiation, magnetic resonance imaging, radiomics

中图分类号:

曾佳乐, 奉鑫, 徐啟业, 王智慧. 构建多参数MRI影像组学模型预测肝细胞癌组织分化程度[J]. 岭南现代临床外科, 2025, 25(06): 367-374.

ZENG Jia-le, FENG Xin, XU Qi-ye, WANG Zhi-hui. A multiparametric MRI-based radiomics model for the preoperative prediction of histological differentiation in hepatocellular carcinoma[J]. Lingnan Modern Clinics In Surgery, 2025, 25(06): 367-374.

参考文献

[1] Sung H, Ferlay J, Siegel RL, et al.Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries[J]. CA Cancer J Clin, 2021, 71(3): 209-249.
[2] Brancato V, Garbino N, Salvatore M, et al.MRI-Based Radiomic Features Help Identify Lesions and Predict Histopathological Grade of Hepatocellular Carcinoma[J]. Diagnostics (Basel), 2022, 12(5): 1085.
[3] Li C, Xu J, Liu Y, et al.Kupffer Phase Radiomics Signature in Sonazoid-Enhanced Ultrasound is an Independent and Effective Predictor of the Pathologic Grade of Hepatocellular Carcinoma[J]. J Oncol, 2022, 2022: 6123242.
[4] Martins-Filho SN, Paiva C, Azevedo RS, et al.Histological Grading of Hepatocellular Carcinoma-A Systematic Review of Literature[J]. Front Med (Lausanne), 2017, 4: 193.
[5] Wang W, Wu S, Zhang J, et al.Preoperative Pathological Grading of Hepatocellular Carcinoma Using Ultrasomics of Contrast-Enhanced Ultrasound[J]. Acad Radiol, 2021, 28(8): 1094-1101.
[6] Wu M, Yu H, Pang S, et al.Application of CT-based radiomics combined with laboratory tests such as AFP and PIVKA-II in preoperative prediction of pathologic grade of hepatocellular carcinoma[J]. BMC Med Imaging, 2025, 25(1): 51.
[7] Liver EAFTSOT.EASL Clinical Practice Guidelines: Management of hepatocellular carcinoma[J]. J Hepatol, 2018, 69(1): 182-236.
[8] Mao B, Zhang L, Ning P, et al.Preoperative prediction for pathological grade of hepatocellular carcinoma via machine learning-based radiomics[J]. Eur Radiol, 2020, 30(12): 6924-6932.
[9] Costa AF, Thipphavong S, Arnason T, et al.Fat-Containing Liver Lesions on Imaging: Detection and Differential Diagnosis[J]. AJR Am J Roentgenol, 2018, 210(1): 68-77.
[10] Rong D, Liu W, Kuang S, et al.Preoperative prediction of pathologic grade of HCC on gadobenate dimeglumine-enhanced dynamic MRI[J]. Eur Radiol, 2021, 31(10): 7584-7593.
[11] Liu H, Lu Y, Wang Q, et al.Machine Learning-Based CEMRI Radiomics Integrating LI-RADS Features Achieves Optimal Evaluation of Hepatocellular Carcinoma Differentiation[J]. J Hepatocell Carcinoma, 2023, 10: 2103-2115.
[12] 陆煜杰, 顾文豪, 许大波, 等. 基于CEMRI瘤内瘤周影像组学预测肝细胞癌分化程度的研究[J]. 磁共振成像, 2025, 16(3): 51-57.
[13] Gillies RJ, Kinahan PE, Hricak H.Radiomics: Images Are More than Pictures, They Are Data[J]. Radiology, 2016, 278(2): 563-577.
[14] Zhou Z, Cao S, Chen C, et al.A Novel Nomogram for the Preoperative Prediction of Edmondson-Steiner Grade III-IV in Hepatocellular Carcinoma Patients[J]. J Hepatocell Carcinoma, 2023, 10: 1399-1409.
[15] Han YE, Cho Y, Kim MJ, et al.Hepatocellular carcinoma pathologic grade prediction using radiomics and machine learning models of gadoxetic acid-enhanced MRI: a two-center study[J]. Abdom Radiol (NY), 2023, 48(1): 244-256.
[16] Hu X, Li C, Wang Q, et al.Development and External Validation of a Radiomics Model Derived from Preoperative Gadoxetic Acid-Enhanced MRI for Predicting Histopathologic Grade of Hepatocellular Carcinoma[J]. Diagnostics (Basel), 2023, 13(3): 413.
[17] Buergy D, Wenz F, Groden C, et al.Tumor-platelet interaction in solid tumors[J]. Int J Cancer, 2012, 130(12): 2747-2760.
[18] Enomoto S, Tamai H, Shingaki N, et al.Assessment of hepatocellular carcinomas using conventional magnetic resonance imaging correlated with histological differentiation and a serum marker of poor prognosis[J]. Hepatol Int, 2011, 5(2): 730-737.
[19] Yoneda N, Matsui O, Kobayashi S, et al.Current status of imaging biomarkers predicting the biological nature of hepatocellular carcinoma[J]. Jpn J Radiol, 2019, 37(3): 191-208.
[20] Jiang T, Xu JH, Zou Y, et al.Diffusion-weighted imaging (DWI) of hepatocellular carcinomas: a retrospective analysis of the correlation between qualitative and quantitative DWI and tumour grade[J]. Clin Radiol, 2017, 72(6): 465-472.
[21] Chen W, Zhang T, Xu L, et al.Radiomics Analysis of Contrast-Enhanced CT for Hepatocellular Carcinoma Grading[J]. Front Oncol, 2021, 11: 660509.
[22] Huang H, Pan X, Zhang Y, et al.Triphasic CT Radiomics Model for Preoperative Prediction of Hepatocellular Carcinoma Pathological Grading[J]. J Hepatocell Carcinoma, 2025, 12: 1725-1742.