Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics
Abstract Background In clinical work, there are difficulties in distinguishing pulmonary contusion(PC) from bacterial pneumonia(BP) on CT images by the naked eye alone when the history of trauma is unknown. Artificial intelligence is widely used in medical imaging, but its diagnostic performance for...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-07-01
|
| Series: | BMC Medical Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12880-025-01802-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849767233552646144 |
|---|---|
| author | Tie Deng Junbang Feng Xingyan Le Yuwei Xia Feng Shi Fei Yu Yiqiang Zhan Xinghua Liu Chuanming Li |
| author_facet | Tie Deng Junbang Feng Xingyan Le Yuwei Xia Feng Shi Fei Yu Yiqiang Zhan Xinghua Liu Chuanming Li |
| author_sort | Tie Deng |
| collection | DOAJ |
| description | Abstract Background In clinical work, there are difficulties in distinguishing pulmonary contusion(PC) from bacterial pneumonia(BP) on CT images by the naked eye alone when the history of trauma is unknown. Artificial intelligence is widely used in medical imaging, but its diagnostic performance for pulmonary contusion is unclear. In this study, artificial intelligence was used for the first time to identify lung contusion and bacterial pneumonia, and its diagnostic performance was compared with that of manual. Methods In this retrospective study, 2179 patients between April 2016 and July 2022 from two hospitals were collected and divided into a training set, an internal validation set, an external validation set. PC and BP were automatically recognized, segmented using VB-net and radiomics features were automatically extracted. Four machine learning algorithms including Decision Trees, Logistic Regression, Random Forests and Support Vector Machines(SVM) were using to built the models. De-long test was used to compare the performance among models. The best performing model and four radiologists diagnosed the external validation set, and compare the diagnostic efficacy of human and artificial intelligence. Results VB-net automatically detected and segmented PC and BP. Among the four machine learning models we’ve built, De-long test showed that SVM model had the best performance, with AUC, accuracy, sensitivity, and specificity of 0.998 (95% CI: 0.995-1), 0.980, 0.979, 0.982 in the training set, 0.891 (95% CI: 0.854–0.928), 0.979, 0.750, 0.860 in the internal validation set, 0.885 (95% CI: 0.850–0.920), 0.903, 0.976, 0.794 in the external validation set. The diagnostic ability of the SVM model was superior to that of human (P < 0.05). Conclusion Our VB-net automatically recognizes and segments PC and BP in chest CT images. SVM model based on radiomics features can quickly and accurately differentiate between them with higher accuracy than experienced radiologist. |
| format | Article |
| id | doaj-art-a358403c0af947d2b2bb318fea75c84b |
| institution | DOAJ |
| issn | 1471-2342 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Medical Imaging |
| spelling | doaj-art-a358403c0af947d2b2bb318fea75c84b2025-08-20T03:04:17ZengBMCBMC Medical Imaging1471-23422025-07-012511910.1186/s12880-025-01802-1Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomicsTie Deng0Junbang Feng1Xingyan Le2Yuwei Xia3Feng Shi4Fei Yu5Yiqiang Zhan6Xinghua Liu7Chuanming Li8Medical Imaging Department, Chongqing Emergency Medical Center, School of Medicine, Chongqing University Central Hospital, Chongqing UniversityMedical Imaging Department, Chongqing Emergency Medical Center, School of Medicine, Chongqing University Central Hospital, Chongqing UniversityMedical Imaging Department, Chongqing Emergency Medical Center, School of Medicine, Chongqing University Central Hospital, Chongqing UniversityShanghai United Imaging Intelligence, Co., Ltd.Shanghai United Imaging Intelligence, Co., Ltd.Medical Imaging Department, Chongqing Emergency Medical Center, School of Medicine, Chongqing University Central Hospital, Chongqing UniversityShanghai United Imaging Intelligence, Co., Ltd.Department of Radiology, Chongqing University Three Gorges HospitalMedical Imaging Department, Chongqing Emergency Medical Center, School of Medicine, Chongqing University Central Hospital, Chongqing UniversityAbstract Background In clinical work, there are difficulties in distinguishing pulmonary contusion(PC) from bacterial pneumonia(BP) on CT images by the naked eye alone when the history of trauma is unknown. Artificial intelligence is widely used in medical imaging, but its diagnostic performance for pulmonary contusion is unclear. In this study, artificial intelligence was used for the first time to identify lung contusion and bacterial pneumonia, and its diagnostic performance was compared with that of manual. Methods In this retrospective study, 2179 patients between April 2016 and July 2022 from two hospitals were collected and divided into a training set, an internal validation set, an external validation set. PC and BP were automatically recognized, segmented using VB-net and radiomics features were automatically extracted. Four machine learning algorithms including Decision Trees, Logistic Regression, Random Forests and Support Vector Machines(SVM) were using to built the models. De-long test was used to compare the performance among models. The best performing model and four radiologists diagnosed the external validation set, and compare the diagnostic efficacy of human and artificial intelligence. Results VB-net automatically detected and segmented PC and BP. Among the four machine learning models we’ve built, De-long test showed that SVM model had the best performance, with AUC, accuracy, sensitivity, and specificity of 0.998 (95% CI: 0.995-1), 0.980, 0.979, 0.982 in the training set, 0.891 (95% CI: 0.854–0.928), 0.979, 0.750, 0.860 in the internal validation set, 0.885 (95% CI: 0.850–0.920), 0.903, 0.976, 0.794 in the external validation set. The diagnostic ability of the SVM model was superior to that of human (P < 0.05). Conclusion Our VB-net automatically recognizes and segments PC and BP in chest CT images. SVM model based on radiomics features can quickly and accurately differentiate between them with higher accuracy than experienced radiologist.https://doi.org/10.1186/s12880-025-01802-1Pulmonary contusionBacterial pneumoniaDeep learningRadiomicsTomographyX-ray computed |
| spellingShingle | Tie Deng Junbang Feng Xingyan Le Yuwei Xia Feng Shi Fei Yu Yiqiang Zhan Xinghua Liu Chuanming Li Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics BMC Medical Imaging Pulmonary contusion Bacterial pneumonia Deep learning Radiomics Tomography X-ray computed |
| title | Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| title_full | Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| title_fullStr | Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| title_full_unstemmed | Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| title_short | Automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| title_sort | automatic recognition and differentiation of pulmonary contusion and bacterial pneumonia based on deep learning and radiomics |
| topic | Pulmonary contusion Bacterial pneumonia Deep learning Radiomics Tomography X-ray computed |
| url | https://doi.org/10.1186/s12880-025-01802-1 |
| work_keys_str_mv | AT tiedeng automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT junbangfeng automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT xingyanle automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT yuweixia automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT fengshi automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT feiyu automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT yiqiangzhan automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT xinghualiu automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics AT chuanmingli automaticrecognitionanddifferentiationofpulmonarycontusionandbacterialpneumoniabasedondeeplearningandradiomics |