Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph
Image-based teleconsultation using smartphones has become increasingly popular. In parallel, deep learning algorithms have been developed to detect radiological findings in chest X-rays (CXRs). However, the feasibility of using smartphones to automate this process has yet to be evaluated. This study...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Nature Publishing Group
2021
|
| Online Access: | https://hdl.handle.net/1721.1/130526 |
| _version_ | 1811075259496923136 |
|---|---|
| author | Kuo, Po-Chih Pollard, Tom Joseph Johnson, Alistair Edward William Celi, Leo Anthony G. |
| author2 | Massachusetts Institute of Technology. Institute for Medical Engineering & Science |
| author_facet | Massachusetts Institute of Technology. Institute for Medical Engineering & Science Kuo, Po-Chih Pollard, Tom Joseph Johnson, Alistair Edward William Celi, Leo Anthony G. |
| author_sort | Kuo, Po-Chih |
| collection | MIT |
| description | Image-based teleconsultation using smartphones has become increasingly popular. In parallel, deep learning algorithms have been developed to detect radiological findings in chest X-rays (CXRs). However, the feasibility of using smartphones to automate this process has yet to be evaluated. This study developed a recalibration method to build deep learning models to detect radiological findings on CXR photographs. Two publicly available databases (MIMIC-CXR and CheXpert) were used to build the models, and four derivative datasets containing 6453 CXR photographs were collected to evaluate model performance. After recalibration, the model achieved areas under the receiver operating characteristic curve of 0.80 (95% confidence interval: 0.78–0.82), 0.88 (0.86–0.90), 0.81 (0.79–0.84), 0.79 (0.77–0.81), 0.84 (0.80–0.88), and 0.90 (0.88–0.92), respectively, for detecting cardiomegaly, edema, consolidation, atelectasis, pneumothorax, and pleural effusion. The recalibration strategy, respectively, recovered 84.9%, 83.5%, 53.2%, 57.8%, 69.9%, and 83.0% of performance losses of the uncalibrated model. We conclude that the recalibration method can transfer models from digital CXRs to CXR photographs, which is expected to help physicians’ clinical works. |
| first_indexed | 2024-09-23T10:03:15Z |
| format | Article |
| id | mit-1721.1/130526 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T10:03:15Z |
| publishDate | 2021 |
| publisher | Nature Publishing Group |
| record_format | dspace |
| spelling | mit-1721.1/1305262022-09-26T15:25:09Z Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph Kuo, Po-Chih Pollard, Tom Joseph Johnson, Alistair Edward William Celi, Leo Anthony G. Massachusetts Institute of Technology. Institute for Medical Engineering & Science Image-based teleconsultation using smartphones has become increasingly popular. In parallel, deep learning algorithms have been developed to detect radiological findings in chest X-rays (CXRs). However, the feasibility of using smartphones to automate this process has yet to be evaluated. This study developed a recalibration method to build deep learning models to detect radiological findings on CXR photographs. Two publicly available databases (MIMIC-CXR and CheXpert) were used to build the models, and four derivative datasets containing 6453 CXR photographs were collected to evaluate model performance. After recalibration, the model achieved areas under the receiver operating characteristic curve of 0.80 (95% confidence interval: 0.78–0.82), 0.88 (0.86–0.90), 0.81 (0.79–0.84), 0.79 (0.77–0.81), 0.84 (0.80–0.88), and 0.90 (0.88–0.92), respectively, for detecting cardiomegaly, edema, consolidation, atelectasis, pneumothorax, and pleural effusion. The recalibration strategy, respectively, recovered 84.9%, 83.5%, 53.2%, 57.8%, 69.9%, and 83.0% of performance losses of the uncalibrated model. We conclude that the recalibration method can transfer models from digital CXRs to CXR photographs, which is expected to help physicians’ clinical works. Taiwan. Ministry of Science and Technology (Grant MOST109-2222-E-007-004-MY3) National Institutes of Health (U.S.) (R01 grant EB017205) 2021-04-27T15:28:05Z 2021-04-27T15:28:05Z 2021-02 2020-07 2021-04-06T15:54:37Z Article http://purl.org/eprint/type/JournalArticle 2398-6352 https://hdl.handle.net/1721.1/130526 Kuo, Po-Chih et al. “Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph.” npj Digital Medicine, 4, 1 (February 2021): 25 © 2021 The Author(s) en 10.1038/s41746-021-00393-9 npj Digital Medicine Creative Commons Attribution 4.0 International license https://creativecommons.org/licenses/by/4.0/ application/pdf Nature Publishing Group Nature |
| spellingShingle | Kuo, Po-Chih Pollard, Tom Joseph Johnson, Alistair Edward William Celi, Leo Anthony G. Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title | Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title_full | Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title_fullStr | Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title_full_unstemmed | Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title_short | Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph |
| title_sort | recalibration of deep learning models for abnormality detection in smartphone captured chest radiograph |
| url | https://hdl.handle.net/1721.1/130526 |
| work_keys_str_mv | AT kuopochih recalibrationofdeeplearningmodelsforabnormalitydetectioninsmartphonecapturedchestradiograph AT pollardtomjoseph recalibrationofdeeplearningmodelsforabnormalitydetectioninsmartphonecapturedchestradiograph AT johnsonalistairedwardwilliam recalibrationofdeeplearningmodelsforabnormalitydetectioninsmartphonecapturedchestradiograph AT celileoanthonyg recalibrationofdeeplearningmodelsforabnormalitydetectioninsmartphonecapturedchestradiograph |