A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms
Abstract Background Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed t...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2022-05-01
|
| Series: | EJNMMI Research |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13550-022-00901-2 |
| _version_ | 1828348145789042688 |
|---|---|
| author | Esben Andreas Carlsen Kristian Lindholm Amalie Hindsholm Mathias Gæde Claes Nøhr Ladefoged Mathias Loft Camilla Bardram Johnbeck Seppo Wang Langer Peter Oturai Ulrich Knigge Andreas Kjaer Flemming Littrup Andersen |
| author_facet | Esben Andreas Carlsen Kristian Lindholm Amalie Hindsholm Mathias Gæde Claes Nøhr Ladefoged Mathias Loft Camilla Bardram Johnbeck Seppo Wang Langer Peter Oturai Ulrich Knigge Andreas Kjaer Flemming Littrup Andersen |
| author_sort | Esben Andreas Carlsen |
| collection | DOAJ |
| description | Abstract Background Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed to implement and test an artificial intelligence (AI) network for tumor segmentation. Patients with gastroenteropancreatic or lung NEN with [64Cu]Cu-DOTATATE PET/CT performed were included in our training (n = 117) and test cohort (n = 41). Further, 10 patients with no signs of NEN were included as negative controls. Ground truth segmentations were obtained by a standardized semiautomatic method for tumor segmentation by a physician. The nnU-Net framework was used to set up a deep learning U-net architecture. Dice score, sensitivity and precision were used for selection of the final model. AI segmentations were implemented in a clinical imaging viewer where a physician evaluated performance and performed manual adjustments. Results Cross-validation training was used to generate models and an ensemble model. The ensemble model performed best overall with a lesion-wise dice of 0.850 and pixel-wise dice, precision and sensitivity of 0.801, 0.786 and 0.872, respectively. Performance of the ensemble model was acceptable with some degree of manual adjustment in 35/41 (85%) patients. Final tumor segmentation could be obtained from the AI model with manual adjustments in 5 min versus 17 min for ground truth method, p < 0.01. Conclusion We implemented and validated an AI model that achieved a high similarity with ground truth segmentation and resulted in faster tumor segmentation. With AI, total tumor segmentation may become feasible in the clinical routine. |
| first_indexed | 2024-04-14T00:50:14Z |
| format | Article |
| id | doaj.art-eeb514d8235f486b8e309f1663897bd7 |
| institution | Directory Open Access Journal |
| issn | 2191-219X |
| language | English |
| last_indexed | 2024-04-14T00:50:14Z |
| publishDate | 2022-05-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | EJNMMI Research |
| spelling | doaj.art-eeb514d8235f486b8e309f1663897bd72022-12-22T02:21:49ZengSpringerOpenEJNMMI Research2191-219X2022-05-0112111010.1186/s13550-022-00901-2A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasmsEsben Andreas Carlsen0Kristian Lindholm1Amalie Hindsholm2Mathias Gæde3Claes Nøhr Ladefoged4Mathias Loft5Camilla Bardram Johnbeck6Seppo Wang Langer7Peter Oturai8Ulrich Knigge9Andreas Kjaer10Flemming Littrup Andersen11Department of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenENETS Neuroendocrine Tumor Center of Excellence, Copenhagen University Hospital – RigshospitaletDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenENETS Neuroendocrine Tumor Center of Excellence, Copenhagen University Hospital – RigshospitaletDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenDepartment of Clinical Physiology and Nuclear Medicine & Cluster for Molecular Imaging, Copenhagen University Hospital – Rigshospitalet & Department of Biomedical Sciences, University of CopenhagenAbstract Background Segmentation of neuroendocrine neoplasms (NENs) in [64Cu]Cu-DOTATATE positron emission tomography makes it possible to extract quantitative measures useable for prognostication of patients. However, manual tumor segmentation is cumbersome and time-consuming. Therefore, we aimed to implement and test an artificial intelligence (AI) network for tumor segmentation. Patients with gastroenteropancreatic or lung NEN with [64Cu]Cu-DOTATATE PET/CT performed were included in our training (n = 117) and test cohort (n = 41). Further, 10 patients with no signs of NEN were included as negative controls. Ground truth segmentations were obtained by a standardized semiautomatic method for tumor segmentation by a physician. The nnU-Net framework was used to set up a deep learning U-net architecture. Dice score, sensitivity and precision were used for selection of the final model. AI segmentations were implemented in a clinical imaging viewer where a physician evaluated performance and performed manual adjustments. Results Cross-validation training was used to generate models and an ensemble model. The ensemble model performed best overall with a lesion-wise dice of 0.850 and pixel-wise dice, precision and sensitivity of 0.801, 0.786 and 0.872, respectively. Performance of the ensemble model was acceptable with some degree of manual adjustment in 35/41 (85%) patients. Final tumor segmentation could be obtained from the AI model with manual adjustments in 5 min versus 17 min for ground truth method, p < 0.01. Conclusion We implemented and validated an AI model that achieved a high similarity with ground truth segmentation and resulted in faster tumor segmentation. With AI, total tumor segmentation may become feasible in the clinical routine.https://doi.org/10.1186/s13550-022-00901-2Tumor segmentationArtificial intelligenceNeuroendocrine neoplasms[64Cu]Cu-DOTATATE PETPrognostication |
| spellingShingle | Esben Andreas Carlsen Kristian Lindholm Amalie Hindsholm Mathias Gæde Claes Nøhr Ladefoged Mathias Loft Camilla Bardram Johnbeck Seppo Wang Langer Peter Oturai Ulrich Knigge Andreas Kjaer Flemming Littrup Andersen A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms EJNMMI Research Tumor segmentation Artificial intelligence Neuroendocrine neoplasms [64Cu]Cu-DOTATATE PET Prognostication |
| title | A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms |
| title_full | A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms |
| title_fullStr | A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms |
| title_full_unstemmed | A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms |
| title_short | A convolutional neural network for total tumor segmentation in [64Cu]Cu-DOTATATE PET/CT of patients with neuroendocrine neoplasms |
| title_sort | convolutional neural network for total tumor segmentation in 64cu cu dotatate pet ct of patients with neuroendocrine neoplasms |
| topic | Tumor segmentation Artificial intelligence Neuroendocrine neoplasms [64Cu]Cu-DOTATATE PET Prognostication |
| url | https://doi.org/10.1186/s13550-022-00901-2 |
| work_keys_str_mv | AT esbenandreascarlsen aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT kristianlindholm aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT amaliehindsholm aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT mathiasgæde aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT claesnøhrladefoged aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT mathiasloft aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT camillabardramjohnbeck aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT seppowanglanger aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT peteroturai aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT ulrichknigge aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT andreaskjaer aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT flemminglittrupandersen aconvolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT esbenandreascarlsen convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT kristianlindholm convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT amaliehindsholm convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT mathiasgæde convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT claesnøhrladefoged convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT mathiasloft convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT camillabardramjohnbeck convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT seppowanglanger convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT peteroturai convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT ulrichknigge convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT andreaskjaer convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms AT flemminglittrupandersen convolutionalneuralnetworkfortotaltumorsegmentationin64cucudotatatepetctofpatientswithneuroendocrineneoplasms |