DeepTek CXR Analyzer
Identification of suspicious regions of interest (ROIs) in the lungs, pleura, cardiac, and hardware.

Overview
The AI application detects suspicious ROIs by analysing frontal chest radiographs using deep learning algorithms and provides relevant annotations to assist radiologists with their interpretations.

Features
Detects suspicious regions of interest (ROIs) in frontal (AP/PA) Chest Radiographs and provides:
- A bounding box surrounding the ROI
- A category label representing the category of the ROI.
- Lungs: such as TB, Fibrosis, Pneumonia, Edema, Nodules, Opacity
- Pleura: such as Pneumothorax, Pleural thickening, Pleural calcification, Pleural effusion
- Cardiac: such as Cardiomegaly, Pericardial Effusion
- Hardware: such as sternal sutures, chest leads, pacemaker, implants, lines, tubes, spinal implants



Benefits
- Classifies chest abnormalities in radiographs with high accuracy, potentially improving clinical workflows by enabling faster and more efficient diagnoses, and even outperforming human readers in some cases.¹
- The device helps detect suspicious ROIs in Lungs, Pleura, Cardiac, and Hardware categories.
- Reader performance in detecting and localizing suspicious ROIs in Lungs, Pleura, Cardiac, and Hardware regions, are likely to improve when aided by DeepTek CXR Analyzer.
- Earlier detection of suspicious ROIs will allow earlier intervention, and increase diagnostic accuracy, thereby minimizing the risk of delayed diagnosis.
- Reader performance in finalizing non-suspicious scans may improve when aided by DeepTek CXR Analyzer.
Image examples
To find out more about this solution or any of the other 140+ applications on the Blackford Platform, please book a discovery call with our team.

Book a meeting
We’d welcome the opportunity to learn more about your AI needs and to explain how partnering with Blackford can drive efficiency and provide ongoing value.
Book a Meeting1. Ajmera P, Onkar P, Desai S, Pant R, Seth J, Gupte T, et al. Validation of a Deep Learning Model for Detecting Chest Pathologies from Digital Chest Radiographs. Diagnostics. 2023 Feb 2;13(3):557.