A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images
Abstract Background Automatic segmentation of hepatocellular carcinoma (HCC) on computed tomography (CT) scans is in urgent need to assist diagnosis and radiomics analysis. The aim of this study is to develop a deep learning based network to detect HCC from dynamic CT images. Methods Dynamic CT imag...
Main Authors: | , , , , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2024-03-01
|
Series: | Cancer Imaging |
Subjects: | |
Online Access: | https://doi.org/10.1186/s40644-024-00686-8 |
_version_ | 1797233273295863808 |
---|---|
author | I-Cheng Lee Yung-Ping Tsai Yen-Cheng Lin Ting-Chun Chen Chia-Heng Yen Nai-Chi Chiu Hsuen-En Hwang Chien-An Liu Jia-Guan Huang Rheun-Chuan Lee Yee Chao Shinn-Ying Ho Yi-Hsiang Huang |
author_facet | I-Cheng Lee Yung-Ping Tsai Yen-Cheng Lin Ting-Chun Chen Chia-Heng Yen Nai-Chi Chiu Hsuen-En Hwang Chien-An Liu Jia-Guan Huang Rheun-Chuan Lee Yee Chao Shinn-Ying Ho Yi-Hsiang Huang |
author_sort | I-Cheng Lee |
collection | DOAJ |
description | Abstract Background Automatic segmentation of hepatocellular carcinoma (HCC) on computed tomography (CT) scans is in urgent need to assist diagnosis and radiomics analysis. The aim of this study is to develop a deep learning based network to detect HCC from dynamic CT images. Methods Dynamic CT images of 595 patients with HCC were used. Tumors in dynamic CT images were labeled by radiologists. Patients were randomly divided into training, validation and test sets in a ratio of 5:2:3, respectively. We developed a hierarchical fusion strategy of deep learning networks (HFS-Net). Global dice, sensitivity, precision and F1-score were used to measure performance of the HFS-Net model. Results The 2D DenseU-Net using dynamic CT images was more effective for segmenting small tumors, whereas the 2D U-Net using portal venous phase images was more effective for segmenting large tumors. The HFS-Net model performed better, compared with the single-strategy deep learning models in segmenting small and large tumors. In the test set, the HFS-Net model achieved good performance in identifying HCC on dynamic CT images with global dice of 82.8%. The overall sensitivity, precision and F1-score were 84.3%, 75.5% and 79.6% per slice, respectively, and 92.2%, 93.2% and 92.7% per patient, respectively. The sensitivity in tumors < 2 cm, 2–3, 3–5 cm and > 5 cm were 72.7%, 92.9%, 94.2% and 100% per patient, respectively. Conclusions The HFS-Net model achieved good performance in the detection and segmentation of HCC from dynamic CT images, which may support radiologic diagnosis and facilitate automatic radiomics analysis. |
first_indexed | 2024-04-24T16:13:33Z |
format | Article |
id | doaj.art-a92c0700a21943ae84df10a7de856154 |
institution | Directory Open Access Journal |
issn | 1470-7330 |
language | English |
last_indexed | 2024-04-24T16:13:33Z |
publishDate | 2024-03-01 |
publisher | BMC |
record_format | Article |
series | Cancer Imaging |
spelling | doaj.art-a92c0700a21943ae84df10a7de8561542024-03-31T11:33:51ZengBMCCancer Imaging1470-73302024-03-0124111010.1186/s40644-024-00686-8A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography imagesI-Cheng Lee0Yung-Ping Tsai1Yen-Cheng Lin2Ting-Chun Chen3Chia-Heng Yen4Nai-Chi Chiu5Hsuen-En Hwang6Chien-An Liu7Jia-Guan Huang8Rheun-Chuan Lee9Yee Chao10Shinn-Ying Ho11Yi-Hsiang Huang12Division of Gastroenterology and Hepatology, Department of Medicine, Taipei Veterans General HospitalInstitute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung UniversityInstitute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung UniversityInstitute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung UniversityInstitute of Computer Science and Engineering, National Yang Ming Chiao Tung UniversityDepartment of Radiology, Taipei Veterans General HospitalDepartment of Radiology, Taipei Veterans General HospitalDepartment of Radiology, Taipei Veterans General HospitalNational Taiwan University School of MedicineDepartment of Radiology, Taipei Veterans General HospitalCancer Center, Taipei Veterans General HospitalInstitute of Bioinformatics and Systems Biology, National Yang Ming Chiao Tung UniversityDivision of Gastroenterology and Hepatology, Department of Medicine, Taipei Veterans General HospitalAbstract Background Automatic segmentation of hepatocellular carcinoma (HCC) on computed tomography (CT) scans is in urgent need to assist diagnosis and radiomics analysis. The aim of this study is to develop a deep learning based network to detect HCC from dynamic CT images. Methods Dynamic CT images of 595 patients with HCC were used. Tumors in dynamic CT images were labeled by radiologists. Patients were randomly divided into training, validation and test sets in a ratio of 5:2:3, respectively. We developed a hierarchical fusion strategy of deep learning networks (HFS-Net). Global dice, sensitivity, precision and F1-score were used to measure performance of the HFS-Net model. Results The 2D DenseU-Net using dynamic CT images was more effective for segmenting small tumors, whereas the 2D U-Net using portal venous phase images was more effective for segmenting large tumors. The HFS-Net model performed better, compared with the single-strategy deep learning models in segmenting small and large tumors. In the test set, the HFS-Net model achieved good performance in identifying HCC on dynamic CT images with global dice of 82.8%. The overall sensitivity, precision and F1-score were 84.3%, 75.5% and 79.6% per slice, respectively, and 92.2%, 93.2% and 92.7% per patient, respectively. The sensitivity in tumors < 2 cm, 2–3, 3–5 cm and > 5 cm were 72.7%, 92.9%, 94.2% and 100% per patient, respectively. Conclusions The HFS-Net model achieved good performance in the detection and segmentation of HCC from dynamic CT images, which may support radiologic diagnosis and facilitate automatic radiomics analysis.https://doi.org/10.1186/s40644-024-00686-8Hepatocellular carcinomaDeep learningSegmentationDetectionComputed tomography |
spellingShingle | I-Cheng Lee Yung-Ping Tsai Yen-Cheng Lin Ting-Chun Chen Chia-Heng Yen Nai-Chi Chiu Hsuen-En Hwang Chien-An Liu Jia-Guan Huang Rheun-Chuan Lee Yee Chao Shinn-Ying Ho Yi-Hsiang Huang A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images Cancer Imaging Hepatocellular carcinoma Deep learning Segmentation Detection Computed tomography |
title | A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
title_full | A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
title_fullStr | A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
title_full_unstemmed | A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
title_short | A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
title_sort | hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from computed tomography images |
topic | Hepatocellular carcinoma Deep learning Segmentation Detection Computed tomography |
url | https://doi.org/10.1186/s40644-024-00686-8 |
work_keys_str_mv | AT ichenglee ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yungpingtsai ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yenchenglin ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT tingchunchen ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT chiahengyen ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT naichichiu ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT hsuenenhwang ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT chienanliu ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT jiaguanhuang ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT rheunchuanlee ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yeechao ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT shinnyingho ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yihsianghuang ahierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT ichenglee hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yungpingtsai hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yenchenglin hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT tingchunchen hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT chiahengyen hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT naichichiu hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT hsuenenhwang hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT chienanliu hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT jiaguanhuang hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT rheunchuanlee hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yeechao hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT shinnyingho hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages AT yihsianghuang hierarchicalfusionstrategyofdeeplearningnetworksfordetectionandsegmentationofhepatocellularcarcinomafromcomputedtomographyimages |