Application study of non-alcoholic steatohepatitis based on radiomics of computed tomography
10.3969/j.issn.1673-9701.2024.04.014
- VernacularTitle:CT影像组学识别非酒精性脂肪性肝炎的应用研究
- Author:
Yan HU
1
,
2
;
Qiaowei SONG
3
,
4
Author Information
1. Department of Gastroenterology, Hangzhou Shangcheng District People&rsquo
2. s Hospital, Hangzhou 310009, Zhejiang, China
3. Department of Radiology, Zhejiang Provinical Pople&rsquo
4. s Hospital, Hangzhou 310014, Zhejiang, China
- Publication Type:Journal Article
- Keywords:
Computed tomography;
Non-alcoholic steatohepatitis;
Nonalcoholic fatty liver disease;
Radiomics;
Prediction model
- From:
China Modern Doctor
2024;62(4):56-59
- CountryChina
- Language:Chinese
-
Abstract:
Objective To construct and validate a whole liver radiomics model based on computed tomography(CT)to identify patients with non-alcoholic steatohepatitis(NASH).Methods A total of 122 patients with nonalcoholic fatty liver disease treated in Zhejiang Provincial People's Hospital from June 2018 to December 2022 were retrospectively selected,including 52 patients with NASH.They were randomly divided into training group(n=85)and test group(n=37)according to a ratio of 7:3,and the liver plain scan images of each patient were selected to extract the radiomics features.The extracted features of the training group were dimensioned down and the radiomics signature was established.After that,a joint model was constructed with relevant clinical features to identify NASH patients,and the diagnostic effectiveness of the model was evaluated using receiver operating characteristic curve and test group data.Results The joint model was constructed based on age and radiomics labels.The diagnostic efficiency of the model for identifying NASH patients in training group and test group were 0.899 and 0.880,the specificity were 91.2%and 88.1%,and the sensitivity were 86.7%and 88.2%,respectively.In addition,the calibration curve also showed good calibration performance in training group and test group.Conclusion The joint model based on liver CT can quantitatively evaluate NASH,and is expected to provide a non-invasive evaluation tool for clinical use.