Abstract
We aimed to develop a deep-learning algorithm detecting 10 common abnormalities (DLAD-10) on chest radiographs and to evaluate its impact in diagnostic accuracy, timeliness of reporting, and workflow efficacy.
DLAD-10 was trained with 146 717 radiographs from 108 053 patients using a ResNet34-based neural network with lesion-specific channels for 10 common radiologic abnormalities (pneumothorax, mediastinal widening, pneumoperitoneum, nodule/mass, consolidation, pleural effusion, linear atelectasis, fibrosis, calcification, and cardiomegaly). For external validation, the performance of DLAD-10 on a same-day CT-confirmed dataset (normal:abnormal, 53:147) and an open-source dataset (PadChest; normal:abnormal, 339:334) was compared to that of three radiologists. Separate simulated reading tests were conducted on another dataset adjusted to real-world disease prevalence in the emergency department, consisting of four critical, 52 urgent, and 146 non-urgent cases. Six radiologists participated in the simulated reading sessions with and without DLAD-10.
DLAD-10 exhibited areas under the receiver-operating characteristic curves (AUROCs) of 0.895–1.00 in the CT-confirmed dataset and 0.913–0.997 in the PadChest dataset. DLAD-10 correctly classified significantly more critical abnormalities (95.0% [57/60]) than pooled radiologists (84.4% [152/180]; p=0.01). In simulated reading tests for emergency department patients, pooled readers detected significantly more critical (70.8% [17/24] versus 29.2% [7/24]; p=0.006) and urgent (82.7% [258/312] versus 78.2% [244/312]; p=0.04) abnormalities when aided by DLAD-10. DLAD-10 assistance shortened the mean time-to-report critical and urgent radiographs (640.5±466.3 versus 3371.0±1352.5 s and 1840.3±1141.1 versus 2127.1±1468.2, respectively; p-values<0.01) and reduced the mean interpretation time (20.5±22.8 versus 23.5±23.7 s; p<0.001).
DLAD-10 showed excellent performance, improving radiologists' performance and shortening the reporting time for critical and urgent cases.
Footnotes
This manuscript has recently been accepted for publication in the European Respiratory Journal. It is published here in its accepted form prior to copyediting and typesetting by our production team. After these production processes are complete and the authors have approved the resulting proofs, the article will move to the latest issue of the ERJ online. Please open or download the PDF to view this article.
Conflict of interest: Dr. NAM reports grants from National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (grant number: NRF-2018R1A5A1060031), grants from Seoul National University Hospital Research Fund (grant number: 03-2019-0190), during the conduct of the study.
Conflict of interest: Dr. Kim reports other from Employee of Lunit Incorporated, during the conduct of the study.
Conflict of interest: Dr. Park reports grants from National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (grant number: NRF-2018R1A5A1060031), grants from Seoul National University Hospital Research Fund (grant number: 03-2019-0190), during the conduct of the study.
Conflict of interest: Dr. Hwang has nothing to disclose.
Conflict of interest: Dr. Lee has nothing to disclose.
Conflict of interest: Dr. Hong has nothing to disclose.
Conflict of interest: Dr. Goo has nothing to disclose.
Conflict of interest: Dr. Park reports grants from National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (grant number: NRF-2018R1A5A1060031), grants from Seoul National University Hospital Research Fund (grant number: 03-2019-0190), during the conduct of the study.
This is a PDF-only article. Please click on the PDF link above to read it.
- Received August 7, 2020.
- Accepted November 3, 2020.
- Copyright ©ERS 2020
