The Chang Gung Pleural Effusion Detection Software is a medical software that can automatically detect whether there is a pleural effusion in Chest X-Ray. The purpose of this study is to verify whether the Chang Gung Pleural Effusion Detection Software can correctly identify patients with pleural effusion in Chest X-Ray. The results of the software analysis will be used for the performance of the software on the primary and secondary outcomes.

This clinical trial is a retrospective study. DICOM images of de-identified Chest X-Ray were collected from 6 hospitals of Chang Gung Memorial Hospital from January 1, 2018 to December 31, 2020. After confirming that the Chest X-Ray that meet the inclusion and exclusion criteria are correct, 282 samples will be sampled for this test, including 141 images with pleural effusion and 141 images without pleural effusion. The image must be in DICOM format.

Then, 3 specialists physicians interpret 282 samples whether there were pleural effusion, and the result was the standard of this study (Reference standard). After determining the reference standard of each Chest X-Ray, the 282 samples were input into the Chang Gung Pleural Effusion Detection Software, and analyzed by the primary and secondary outcomes.

Inclusion Criteria:

• The case were over 20 years old and under 100 years old.
• DICOM format.
• Brand model of Canon.
• DICOM resolution length range (700-3400 pixels), width range (490-3600 pixels).
• poster-anterior view(PA-view) of chest X-ray

Exclusion Criteria:

• The chest X-ray contains items that affect interpretation, such as chest tubes, endotracheal tubes, and heart rhythm regulators, but does not include patches and circuits used in electrocardiograms.
• The chest X-ray that are difficult to interpret due to poor image quality.

This is a retrospective study, and the data comes from the Chang Gung Medical Research Database(CGRD) which was an database form 6 hospitals of Chang Gung Memorial hospital. According to inclusion and exclusion criteria, Chest X-ray data from the database during 2018.01.01~2020.12.31 was collected.