Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning
Abstract Objectives The goal of this study is to demonstrate the performance of radiomics and CNN-based classifiers in determining the primary origin of gastrointestinal liver metastases for visually indistinguishable lesions. Methods In this retrospective, IRB-approved study, 31 pancreatic cancer p...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2023-10-01
|
| Series: | Cancer Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s40644-023-00612-4 |
| _version_ | 1797556797518643200 |
|---|---|
| author | Hishan Tharmaseelan Abhinay K. Vellala Alexander Hertel Fabian Tollens Lukas T. Rotkopf Johann Rink Piotr Woźnicki Isabelle Ayx Sönke Bartling Dominik Nörenberg Stefan O. Schoenberg Matthias F. Froelich |
| author_facet | Hishan Tharmaseelan Abhinay K. Vellala Alexander Hertel Fabian Tollens Lukas T. Rotkopf Johann Rink Piotr Woźnicki Isabelle Ayx Sönke Bartling Dominik Nörenberg Stefan O. Schoenberg Matthias F. Froelich |
| author_sort | Hishan Tharmaseelan |
| collection | DOAJ |
| description | Abstract Objectives The goal of this study is to demonstrate the performance of radiomics and CNN-based classifiers in determining the primary origin of gastrointestinal liver metastases for visually indistinguishable lesions. Methods In this retrospective, IRB-approved study, 31 pancreatic cancer patients with 861 lesions (median age [IQR]: 65.39 [56.87, 75.08], 48.4% male) and 47 colorectal cancer patients with 435 lesions (median age [IQR]: 65.79 [56.99, 74.62], 63.8% male) were enrolled. A pretrained nnU-Net performed automated segmentation of 1296 liver lesions. Radiomics features for each lesion were extracted using pyradiomics. The performance of several radiomics-based machine-learning classifiers was investigated for the lesions and compared to an image-based deep-learning approach using a DenseNet-121. The performance was evaluated by AUC/ROC analysis. Results The radiomics-based K-nearest neighbor classifier showed the best performance on an independent test set with AUC values of 0.87 and an accuracy of 0.67. In comparison, the image-based DenseNet-121-classifier reached an AUC of 0.80 and an accuracy of 0.83. Conclusions CT-based radiomics and deep learning can distinguish the etiology of liver metastases from gastrointestinal primary tumors. Compared to deep learning, radiomics based models showed a varying generalizability in distinguishing liver metastases from colorectal cancer and pancreatic adenocarcinoma. |
| first_indexed | 2024-03-10T17:08:03Z |
| format | Article |
| id | doaj.art-9feae19fae0840dbb5615465790843e2 |
| institution | Directory Open Access Journal |
| issn | 1470-7330 |
| language | English |
| last_indexed | 2024-03-10T17:08:03Z |
| publishDate | 2023-10-01 |
| publisher | BMC |
| record_format | Article |
| series | Cancer Imaging |
| spelling | doaj.art-9feae19fae0840dbb5615465790843e22023-11-20T10:46:00ZengBMCCancer Imaging1470-73302023-10-012311910.1186/s40644-023-00612-4Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learningHishan Tharmaseelan0Abhinay K. Vellala1Alexander Hertel2Fabian Tollens3Lukas T. Rotkopf4Johann Rink5Piotr Woźnicki6Isabelle Ayx7Sönke Bartling8Dominik Nörenberg9Stefan O. Schoenberg10Matthias F. Froelich11Department of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityDepartment of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg UniversityAbstract Objectives The goal of this study is to demonstrate the performance of radiomics and CNN-based classifiers in determining the primary origin of gastrointestinal liver metastases for visually indistinguishable lesions. Methods In this retrospective, IRB-approved study, 31 pancreatic cancer patients with 861 lesions (median age [IQR]: 65.39 [56.87, 75.08], 48.4% male) and 47 colorectal cancer patients with 435 lesions (median age [IQR]: 65.79 [56.99, 74.62], 63.8% male) were enrolled. A pretrained nnU-Net performed automated segmentation of 1296 liver lesions. Radiomics features for each lesion were extracted using pyradiomics. The performance of several radiomics-based machine-learning classifiers was investigated for the lesions and compared to an image-based deep-learning approach using a DenseNet-121. The performance was evaluated by AUC/ROC analysis. Results The radiomics-based K-nearest neighbor classifier showed the best performance on an independent test set with AUC values of 0.87 and an accuracy of 0.67. In comparison, the image-based DenseNet-121-classifier reached an AUC of 0.80 and an accuracy of 0.83. Conclusions CT-based radiomics and deep learning can distinguish the etiology of liver metastases from gastrointestinal primary tumors. Compared to deep learning, radiomics based models showed a varying generalizability in distinguishing liver metastases from colorectal cancer and pancreatic adenocarcinoma.https://doi.org/10.1186/s40644-023-00612-4Deep learningRadiomicsMachine learningMetastasesGastrointestinal |
| spellingShingle | Hishan Tharmaseelan Abhinay K. Vellala Alexander Hertel Fabian Tollens Lukas T. Rotkopf Johann Rink Piotr Woźnicki Isabelle Ayx Sönke Bartling Dominik Nörenberg Stefan O. Schoenberg Matthias F. Froelich Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning Cancer Imaging Deep learning Radiomics Machine learning Metastases Gastrointestinal |
| title | Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning |
| title_full | Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning |
| title_fullStr | Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning |
| title_full_unstemmed | Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning |
| title_short | Tumor classification of gastrointestinal liver metastases using CT-based radiomics and deep learning |
| title_sort | tumor classification of gastrointestinal liver metastases using ct based radiomics and deep learning |
| topic | Deep learning Radiomics Machine learning Metastases Gastrointestinal |
| url | https://doi.org/10.1186/s40644-023-00612-4 |
| work_keys_str_mv | AT hishantharmaseelan tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT abhinaykvellala tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT alexanderhertel tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT fabiantollens tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT lukastrotkopf tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT johannrink tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT piotrwoznicki tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT isabelleayx tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT sonkebartling tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT dominiknorenberg tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT stefanoschoenberg tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning AT matthiasffroelich tumorclassificationofgastrointestinallivermetastasesusingctbasedradiomicsanddeeplearning |