Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging

<b><i>Background:</i></b> The exact focus of computed tomography (CT)-based artificial intelligence techniques when staging liver fibrosis is still not exactly known. This study aimed to determine both the added value of splenic information to hepatic information, and the cor...

Full description

Bibliographic Details
Main Authors: Yunchao Yin, Derya Yakar, Rudi A. J. O. Dierckx, Kim B. Mouridsen, Thomas C. Kwee, Robbert J. de Haas
Format: Article
Language:English
Published: MDPI AG 2022-02-01
Series:Diagnostics
Subjects:
Online Access:https://www.mdpi.com/2075-4418/12/2/550
_version_ 1797481103433400320
author Yunchao Yin
Derya Yakar
Rudi A. J. O. Dierckx
Kim B. Mouridsen
Thomas C. Kwee
Robbert J. de Haas
author_facet Yunchao Yin
Derya Yakar
Rudi A. J. O. Dierckx
Kim B. Mouridsen
Thomas C. Kwee
Robbert J. de Haas
author_sort Yunchao Yin
collection DOAJ
description <b><i>Background:</i></b> The exact focus of computed tomography (CT)-based artificial intelligence techniques when staging liver fibrosis is still not exactly known. This study aimed to determine both the added value of splenic information to hepatic information, and the correlation between important radiomic features and information exploited by deep learning models for liver fibrosis staging by CT-based radiomics. <b><i>Methods</i>:</b> The study design is retrospective. Radiomic features were extracted from both liver and spleen on portal venous phase CT images of 252 consecutive patients with histologically proven liver fibrosis stages between 2006 and 2018. The radiomics analyses for liver fibrosis staging were done by hepatic and hepatic–splenic features, respectively. The most predictive radiomic features were automatically selected by machine learning models. <b><i>Results</i></b>: When using splenic–hepatic features in the CT-based radiomics analysis, the average accuracy rates for significant fibrosis, advanced fibrosis, and cirrhosis were 88%, 82%, and 86%, and area under the receiver operating characteristic curves (AUCs) were 0.92, 0.81, and 0.85. The AUC of hepatic–splenic-based radiomics analysis with the ensemble classifier was 7% larger than that of hepatic-based analysis (<i>p</i> < 0.05). The most important features selected by machine learning models included both hepatic and splenic features, and they were consistent with the location maps indicating the focus of deep learning when predicting liver fibrosis stage. <b><i>Conclusions:</i></b> Adding CT-based splenic radiomic features to hepatic radiomic features increases radiomics analysis performance for liver fibrosis staging. The most important features of the radiomics analysis were consistent with the information exploited by deep learning.
first_indexed 2024-03-09T22:10:43Z
format Article
id doaj.art-bf61090f883149cd891ab68a5e9b5a33
institution Directory Open Access Journal
issn 2075-4418
language English
last_indexed 2024-03-09T22:10:43Z
publishDate 2022-02-01
publisher MDPI AG
record_format Article
series Diagnostics
spelling doaj.art-bf61090f883149cd891ab68a5e9b5a332023-11-23T19:34:01ZengMDPI AGDiagnostics2075-44182022-02-0112255010.3390/diagnostics12020550Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis StagingYunchao Yin0Derya Yakar1Rudi A. J. O. Dierckx2Kim B. Mouridsen3Thomas C. Kwee4Robbert J. de Haas5Department of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The NetherlandsDepartment of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The NetherlandsDepartment of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The NetherlandsDepartment of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The NetherlandsDepartment of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The NetherlandsDepartment of Radiology, Medical Imaging Center Groningen, University of Groningen, University Medical Center Groningen, P.O. Box 30001, 9700 RB Groningen, The Netherlands<b><i>Background:</i></b> The exact focus of computed tomography (CT)-based artificial intelligence techniques when staging liver fibrosis is still not exactly known. This study aimed to determine both the added value of splenic information to hepatic information, and the correlation between important radiomic features and information exploited by deep learning models for liver fibrosis staging by CT-based radiomics. <b><i>Methods</i>:</b> The study design is retrospective. Radiomic features were extracted from both liver and spleen on portal venous phase CT images of 252 consecutive patients with histologically proven liver fibrosis stages between 2006 and 2018. The radiomics analyses for liver fibrosis staging were done by hepatic and hepatic–splenic features, respectively. The most predictive radiomic features were automatically selected by machine learning models. <b><i>Results</i></b>: When using splenic–hepatic features in the CT-based radiomics analysis, the average accuracy rates for significant fibrosis, advanced fibrosis, and cirrhosis were 88%, 82%, and 86%, and area under the receiver operating characteristic curves (AUCs) were 0.92, 0.81, and 0.85. The AUC of hepatic–splenic-based radiomics analysis with the ensemble classifier was 7% larger than that of hepatic-based analysis (<i>p</i> < 0.05). The most important features selected by machine learning models included both hepatic and splenic features, and they were consistent with the location maps indicating the focus of deep learning when predicting liver fibrosis stage. <b><i>Conclusions:</i></b> Adding CT-based splenic radiomic features to hepatic radiomic features increases radiomics analysis performance for liver fibrosis staging. The most important features of the radiomics analysis were consistent with the information exploited by deep learning.https://www.mdpi.com/2075-4418/12/2/550liverartificial intelligencemachine learningradiomicsmultidetector computed tomography
spellingShingle Yunchao Yin
Derya Yakar
Rudi A. J. O. Dierckx
Kim B. Mouridsen
Thomas C. Kwee
Robbert J. de Haas
Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
Diagnostics
liver
artificial intelligence
machine learning
radiomics
multidetector computed tomography
title Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
title_full Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
title_fullStr Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
title_full_unstemmed Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
title_short Combining Hepatic and Splenic CT Radiomic Features Improves Radiomic Analysis Performance for Liver Fibrosis Staging
title_sort combining hepatic and splenic ct radiomic features improves radiomic analysis performance for liver fibrosis staging
topic liver
artificial intelligence
machine learning
radiomics
multidetector computed tomography
url https://www.mdpi.com/2075-4418/12/2/550
work_keys_str_mv AT yunchaoyin combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging
AT deryayakar combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging
AT rudiajodierckx combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging
AT kimbmouridsen combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging
AT thomasckwee combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging
AT robbertjdehaas combininghepaticandsplenicctradiomicfeaturesimprovesradiomicanalysisperformanceforliverfibrosisstaging