PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging
Abstract Pulmonary embolism (PE) is a life-threatening clinical problem and computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosis. Prompt diagnosis and immediate treatment are critical to avoid high morbidity and mortality rates, yet PE remains among the diagnoses most...
| Main Authors: | , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2020-04-01
|
| Series: | npj Digital Medicine |
| Online Access: | https://doi.org/10.1038/s41746-020-0266-y |
| _version_ | 1797430198500589568 |
|---|---|
| author | Shih-Cheng Huang Tanay Kothari Imon Banerjee Chris Chute Robyn L. Ball Norah Borus Andrew Huang Bhavik N. Patel Pranav Rajpurkar Jeremy Irvin Jared Dunnmon Joseph Bledsoe Katie Shpanskaya Abhay Dhaliwal Roham Zamanian Andrew Y. Ng Matthew P. Lungren |
| author_facet | Shih-Cheng Huang Tanay Kothari Imon Banerjee Chris Chute Robyn L. Ball Norah Borus Andrew Huang Bhavik N. Patel Pranav Rajpurkar Jeremy Irvin Jared Dunnmon Joseph Bledsoe Katie Shpanskaya Abhay Dhaliwal Roham Zamanian Andrew Y. Ng Matthew P. Lungren |
| author_sort | Shih-Cheng Huang |
| collection | DOAJ |
| description | Abstract Pulmonary embolism (PE) is a life-threatening clinical problem and computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosis. Prompt diagnosis and immediate treatment are critical to avoid high morbidity and mortality rates, yet PE remains among the diagnoses most frequently missed or delayed. In this study, we developed a deep learning model—PENet, to automatically detect PE on volumetric CTPA scans as an end-to-end solution for this purpose. The PENet is a 77-layer 3D convolutional neural network (CNN) pretrained on the Kinetics-600 dataset and fine-tuned on a retrospective CTPA dataset collected from a single academic institution. The PENet model performance was evaluated in detecting PE on data from two different institutions: one as a hold-out dataset from the same institution as the training data and a second collected from an external institution to evaluate model generalizability to an unrelated population dataset. PENet achieved an AUROC of 0.84 [0.82–0.87] on detecting PE on the hold out internal test set and 0.85 [0.81–0.88] on external dataset. PENet also outperformed current state-of-the-art 3D CNN models. The results represent successful application of an end-to-end 3D CNN model for the complex task of PE diagnosis without requiring computationally intensive and time consuming preprocessing and demonstrates sustained performance on data from an external institution. Our model could be applied as a triage tool to automatically identify clinically important PEs allowing for prioritization for diagnostic radiology interpretation and improved care pathways via more efficient diagnosis. |
| first_indexed | 2024-03-09T09:24:11Z |
| format | Article |
| id | doaj.art-8ceb4d87e32c41efa595cc53fbb017ec |
| institution | Directory Open Access Journal |
| issn | 2398-6352 |
| language | English |
| last_indexed | 2024-03-09T09:24:11Z |
| publishDate | 2020-04-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Digital Medicine |
| spelling | doaj.art-8ceb4d87e32c41efa595cc53fbb017ec2023-12-02T06:35:48ZengNature Portfolionpj Digital Medicine2398-63522020-04-01311910.1038/s41746-020-0266-yPENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imagingShih-Cheng Huang0Tanay Kothari1Imon Banerjee2Chris Chute3Robyn L. Ball4Norah Borus5Andrew Huang6Bhavik N. Patel7Pranav Rajpurkar8Jeremy Irvin9Jared Dunnmon10Joseph Bledsoe11Katie Shpanskaya12Abhay Dhaliwal13Roham Zamanian14Andrew Y. Ng15Matthew P. Lungren16Department of Biomedical Data Science, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Biomedical Data Science, Stanford UniversityDepartment of Computer Science, Stanford UniversityCenter for Artificial Intelligence in Medicine & Imaging, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Radiology, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Radiology, Stanford UniversityDepartment of Emergency Medicine, Intermountain Medical CenterDepartment of Radiology, Stanford UniversityMichigan State University, College of Human MedicineDepartment of Pulmonary Critical Care Medicine, Stanford UniversityDepartment of Computer Science, Stanford UniversityDepartment of Biomedical Data Science, Stanford UniversityAbstract Pulmonary embolism (PE) is a life-threatening clinical problem and computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosis. Prompt diagnosis and immediate treatment are critical to avoid high morbidity and mortality rates, yet PE remains among the diagnoses most frequently missed or delayed. In this study, we developed a deep learning model—PENet, to automatically detect PE on volumetric CTPA scans as an end-to-end solution for this purpose. The PENet is a 77-layer 3D convolutional neural network (CNN) pretrained on the Kinetics-600 dataset and fine-tuned on a retrospective CTPA dataset collected from a single academic institution. The PENet model performance was evaluated in detecting PE on data from two different institutions: one as a hold-out dataset from the same institution as the training data and a second collected from an external institution to evaluate model generalizability to an unrelated population dataset. PENet achieved an AUROC of 0.84 [0.82–0.87] on detecting PE on the hold out internal test set and 0.85 [0.81–0.88] on external dataset. PENet also outperformed current state-of-the-art 3D CNN models. The results represent successful application of an end-to-end 3D CNN model for the complex task of PE diagnosis without requiring computationally intensive and time consuming preprocessing and demonstrates sustained performance on data from an external institution. Our model could be applied as a triage tool to automatically identify clinically important PEs allowing for prioritization for diagnostic radiology interpretation and improved care pathways via more efficient diagnosis.https://doi.org/10.1038/s41746-020-0266-y |
| spellingShingle | Shih-Cheng Huang Tanay Kothari Imon Banerjee Chris Chute Robyn L. Ball Norah Borus Andrew Huang Bhavik N. Patel Pranav Rajpurkar Jeremy Irvin Jared Dunnmon Joseph Bledsoe Katie Shpanskaya Abhay Dhaliwal Roham Zamanian Andrew Y. Ng Matthew P. Lungren PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging npj Digital Medicine |
| title | PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging |
| title_full | PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging |
| title_fullStr | PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging |
| title_full_unstemmed | PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging |
| title_short | PENet—a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging |
| title_sort | penet a scalable deep learning model for automated diagnosis of pulmonary embolism using volumetric ct imaging |
| url | https://doi.org/10.1038/s41746-020-0266-y |
| work_keys_str_mv | AT shihchenghuang penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT tanaykothari penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT imonbanerjee penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT chrischute penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT robynlball penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT norahborus penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT andrewhuang penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT bhaviknpatel penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT pranavrajpurkar penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT jeremyirvin penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT jareddunnmon penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT josephbledsoe penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT katieshpanskaya penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT abhaydhaliwal penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT rohamzamanian penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT andrewyng penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging AT matthewplungren penetascalabledeeplearningmodelforautomateddiagnosisofpulmonaryembolismusingvolumetricctimaging |