AUTHOR=Yang Xue , Shao Guoqing , Liu Jiaojiao , Liu Bin , Cai Chao , Zeng Daobing , Li Hongjun 

TITLE=Predictive machine learning model for microvascular invasion identification in hepatocellular carcinoma based on the LI-RADS system

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.1021570

DOI=10.3389/fonc.2022.1021570

ISSN=2234-943X

ABSTRACT=Purposes: This study aimed to establish a predictive model of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) by contrast enhanced computed tomography (CT), which relied on a combination of machine learning approach and imaging features covering liver imaging and reporting and data system (LI-RADS) features.
Methods: The retrospective study included 279 patients with surgery who underwent preoperative enhanced CT. They were randomly allocated to training set, and validation set, and test set (167 patients vs 56 patients vs 56 patients, respectively). Significant imaging findings for predicting MVI were identified through the least absolute shrinkage and selection operator (LASSO) logistic regression method. Predictive model were performed by machine learning algorithms, support vector machine (SVM), in training set and validation set, and evaluated in test set. Further, a combined model adding clinical findings to radiologic model was developed. Based on LI-RADS category, subgroup analyses were conducted.  
Results: We included 116 patients with MVI which were diagnosed through pathological confirmation. Six imaging features were selected about MVI prediction: four LI-RADS features (corona enhancement, enhancing capsule, non-rim aterial phase hyperehancement, tumor size), two non LI-RADS features (internal arteries, non-smooth tumor margin). The radiological feature with the best accuracy was corona enhancement which followed by internal arteries and tumor size. Accuracy of radilogical model and combined model were 0.725-0.714 and 0.802-0.732 in training set, validation set and test set, respectively.In LR-4/5 subgroup, sensitivity of 100% and NPV of 100%  were obtained by the high sensitivity threshold. Specificity of 100% and PPV of 100% were acquired through the high specificity threshold. 
Conclusion: A combination of LI-RADS features and none LI-RADS features, and serum alpha-    -fetoprotein value could be applied as preoperative biomarker for predicting MVI by the machine learning apporach. Furthermore, its good performance in the subgroup by LI-RADS category may help optimize management of HCC patients.