T-Xnet is intended to mark regions of a chest X-ray with specific abnormalities associated with tuberculosis (TB) or pneumonia and alert radiologists and clinicians to these regions during image interpretation.

Certification

Pending

Development Stage

Validation

Deployment

Offline only

Intended Age Group

18+ years

Target Setting

Primary health centers, general hospital (above primary level), teleradiology companies, government/public sector, e.g. national TB program, private sector

Current Market

Currently not on the market

Input

»» Can be used to read images from any kind of chest X-ray machine (vendor neutral)

»» X-ray image format: JPEG, PNG, DICOM, TIFF

»» Chest X-ray type: posterior-anterior chest X-ray, anterior-posterior chest X-ray, Lateral chest X-ray, portable

»» Other requirements: none

Output

Structured report including: 

»» Heat map, 

»» Probability score for TB, 

»» Probability score for the following pulmonary abnormalities: atelectasis, cavity, mass, nodule, pleural effusion, pneumothorax

Server

Minimum 2 GB RAM and CPU greater than or equal to 2.3 GHz

Integration

It is possible to integrate the product with the client’s legacy picture archiving and communication system (PACS).

Software

No external dependencies

Processing Time

Maximum 10 seconds

Hardware

A computer with minimum 2 GB RAM and CPU with at least 2.3 GHz (small stick computer is sufficient).

Validation

Optional: Fine tuning may be useful in certain scenarios. For this, a dataset with a particular machine will need to be prepared.

Data Sharing & Privacy

»» Data is not shared with the manufacturer

»» There is an option to de-identify data

Software Updates

»» Software is updated every 6 months

»» Two options for upgrades: if there is no internet connection available, the manufacturer will ship a replacement device to the client; or they will require a momentary internet connection to run an online upgrade

»» Extra costs: none

Price

»» Flexible pricing models are available.

»» Please contact company for quote: Rajarajeshwari K. at

Hardware

A computer with minimum 2 GB RAM and CPU with at least 2.3 GHz (small stick computer is sufficient).

Product Development Method

Supervised deep learning (CNN, RNN)

Training

The product was trained on 120 000 chest X-rays from USA and China

Reference Standard

Human reader

Publications

»» Peer-reviewed publications are not yet available.