Last Updated:
6 September 2021
InferRead DR Chest is a screening tool that analyses chest radiograph DICOM images from medical imaging storage devices and automatically evaluates the probability of an abnormality suggestive of tuberculosis (TB). It shows the abnormal cases in a study list. InferRead DR Chest is used to assist doctors in the diagnostic process, but should not be used as a confirmatory diagnostic tool on its own. The InferRead DR Chest product offers different deployment configurations that adapt to different use cases.
Certification
CE-marked Class IIa
Development Stage
On the market
Deployment
Online & offline
Intended Age Group
12+ years
Target Setting
Primary health centres, general hospital (above primary level), teleradiology companies, government/public sector e.g. national TB programme, private sector
Current Market
As of August 2021, Infervision has expanded its global strategic layout to America, Asia-Pacific and Europe regions, with presence in nearly 20 countries worldwide.
Input
»» Can be used to read images from any kind of chest X-ray machine (vendor neutral)
»» Chest X-ray image format: JPEG, PNG, DICOM
»» Chest X-ray type: posterior-anterior chest X-ray, anterior-posterior chest X-ray
Output
Structured report including:
»» Heat map
»» Dichotomous output indicating the presence or absence of the following abnormalities: TB, abscess, airfluid level, atelectasis, blunted costophrenic angle, bronchiectasis, calcification, cavity, chest wall invasion/destruction, consolidation, fibrosis, honeycombing, hyperinflation, interstitial markings, loculated pleural effusion, mass, nodule, opacity, pleural effusion, prominence in hilar region, pneumothorax
»» Probability score for TB
»» Probability score for each abnormality
»» Location of each abnormality
Hardware
The computer in place should meet the following requirements:
»» CPU: Intel Core i3 and above
»» Memory: above 4 GB
Validation
Validation is required when the product is installed on the X-Ray machine for the first time. The data transmission needs to be tested to guarantee the workflow can operate appropriately.
Hardware
The computer in place should meet the following requirements:
»» CPU: Intel Core i3 and above
»» Memory: above 4 GB
Server
Requirements for server hardware:
»» CPU: Intel i7 6850K processor and above
»» GPU: NVIDIA GeForce 1080 and above or V100 and above
»» Memory: DDR4 is recommended with a capacity of at least 16 GB
»» Hard disk: for system disk, choose solid state hard drives with capacity of at least 120 GB; a high-speed mechanical hard disk with a capacity of at least 3 TB and rotation speed of 7200 RPM is required
»» Network card: 1000M PCI-E network card
Integration
It is possible to integrate the product with the client’s legacy picture archiving and communication system (PACS).
Software
The computer in place should meet the requirements:
»» Operating system: Windows XP and above
»» Browser: Google Chrome 49.0 and above
Requirements for server software:
»» Operating System: Ubuntu 18.04 LTS and above
»» Browser: Google Chrome 49.0 and above
Processing Time
In the case of the minimum client configuration, and with gigabit broadband, it takes less than 5 seconds to process 1 DICOM image.
Data Sharing & Privacy
»» Server Location: located worldwide (Google Cloud) and/or a local or national server can be set up if required
»» Data are not shared with the developer
»» De-identification option can be provided if required by customer
Price
»» Flexible pricing models are available
»» Perpetual license for InferRead software and related accessories and services are available on the Stop TB Partnership’s Global Drug Facility (GDF) catalogue. Please see the Diagnostics Catalogue of GDF for more details.
»» Please contact the company for quote (contact@infervision.com)
Product Development Method
Supervised deep learning (CNN, RNN)
Training
Please contact the company for this information (contact@infervision.com)
Reference Standard
Culture, sputum smear microscopy, and human reader
Publications
»» Qin, Z.Z., Ahmed, S., Sarker, M.S. et al. Tuberculosis detection from chest x-rays for triaging in a high tuberculosis-burden setting: an evaluation of five artificial intelligence algorithms. Lancet Digital Health. 2021;3:e543–5437: 2113–31. https://doi.org/10.1016/S2589-7500(21)00116-3
This website works best with browsers other than Internet Explorer.