Automated evaluation of pulmonary lesion changes on chest radiograph during follow-up using semantic segmentation

Sep 29, 2026

Uploaded by Aditya

Automated evaluation model uses lesion-specific semantic segmentation to determine whether consolidation and pleural effusion remain changed or unchanged on paired follow-up chest radiographs. Training used 5,178 radiographs from a single institution, while paired emergency department and ICU studies set change thresholds and enabled temporal validation. Performance was measured with receiver operating characteristic AUC and compared with a thoracic radiologist across settings, showing generally comparable accuracy.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

PURPOSE

METHODS

RESULTS

CONCLUSION

CLINICAL SIGNIFICANCE

KEYWORDS

Copyright @ 2026 Author(s) -[Available online at dirjournal.org.](Available online at dirjournal.org. ORIGINAL)[ ORIGINAL](Available online at dirjournal.org. ORIGINAL) ARTICLE Content of this journal is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Automated evaluation of pulmonary lesion changes on chest radiograph during follow-up using semantic segmentation Youngjae Kim 1,2* Yura Ahn3* Sang Min Lee3 Han Na Noh4 Jongjun Won2 Chaewon Kim2 Hyunna Lee5 1University of Ulsan Faculty of Medicine, Department of Biomedical Engineering, AMIST, Asan Medical Center, Seoul, Republic of Korea 2University of Ulsan Faculty of Medicine, Department of Convergence Medicine, Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of Korea 3University of Ulsan Faculty of Medicine, Department of Radiology and Research Institute of Radiology, Asan Medical Center, Seoul, Republic of Korea 4University of Ulsan Faculty of Medicine, Health Screening and Promotion Center, Asan Medical Center, Seoul, Republic of Korea 5Bigdata Research Center, Asan Institute for Life Science, Asan Medical Center, Seoul, Republic of Korea *Joint first authors Corresponding author: Sang Min Lee E-mail: [sangmin.lee.md@gmail.com](sangmin.lee.md@gmail.com) Received 22 July 2025; revision requested 24 August 2025; accepted 23 September 2025. Publication date: 02.01.2026 DOI: 10.4274/dir.2025.253567 PURPOSE To develop and validate a deep learning-based model utilizing lesion-specific segmentation to determine the changed/unchanged status of consolidation and pleural effusion in paired chest radiographs (CRs) . METHODS The model was trained using 5.178 CRs from a single institution for lesion segmentation. Paired CRs from the emergency department (ED) and intensive care unit (ICU) were used to determine the thresholds for change and temporal validation. Model performance was evaluated through the area under the receiver operating characteristic curve (AUC), and its accuracy was compared with that of a thoracic radiologist. RESULTS In the ED, the model achieved AUCs of 0.988 and 0.883 for consolidation and pleural effusion, respectively, with accuracies of 0.900 (36/40) and 0.825 (33/40). The radiologist showed accuracies of 0.975 (39/40) and 0.950 (38/40), respectively. In the ICU, model AUCs were 0.970 (consolidation) and 0.955 (pleural effusion), with accuracies of 0.875 (35/40) and 0.800 (32/40), respectively. Radiologist performance was 0.975 (39/40) for consolidation and 1.000 (40/40) for pleural effusion. No significant accuracy differences were observed between the model and radiologist for consolidation in the ICU or both targets in the ED (all P > 0.05), except for pleural effusion in the ICU (P = 0.01) . CONCLUSION The lesion-specific deep learning model was feasible for identifying interval changes in consolidation and pleural effusion on follow-up CRs. CLINICAL SIGNIFICANCE It could potentially be utilized for prioritizing interpretation, generating alerts, and extracting time-series data from multiple follow-up CRs. KEYWORDS Radiography, thoracic, follow-up studies, diagnosis, computer-assisted, artificial intelligence, segmentation Chest radiography is a widely used medical imaging modality due to its cost-effective ness and low radiation exposure. Chest radiographs (CRs) detect thoracic abnormalities and track changes during follow-ups. Monitoring abnormalities such as pleural effusion or consolidation is crucial for evaluating disease progression and treatment response.1-4 However, frequent follow-up CRs increase workload. For example, in intensive care units (ICUs), CRis often performed daily for patients who are critically ill or after device adjustments, generating millions of ICU CRs annually in the United States.5,6 Consequently, the timely and accurate interpretation of follow-up CRs is becoming more

FAQs

What is the study’s main goal for follow-up chest radiographs?

The goal is to develop and validate a deep learning model that uses lesion-specific segmentation to classify consolidation and pleural effusion as changed or unchanged on paired chest radiographs during follow-up.

How was the model trained and validated?

The model was trained on lesion segmentation using 5,178 chest radiographs from a single institution. Paired radiographs from the emergency department and ICU were used to determine change thresholds and perform temporal validation.

How did the model perform compared with a thoracic radiologist?

In both emergency and ICU settings, the model achieved high AUCs for consolidation and pleural effusion. Accuracy results were generally comparable to the radiologist, with no significant differences for consolidation in the ICU or both targets in the ED, except for pleural effusion in the ICU.

Sep 29, 2026

Uploaded by Aditya