qXR is a chest x-ray interpretation software which analyses chest radiographs. The device integrates with the radiology workflow, using image viewing and reporting tools. The software can detect, localize and provide the position of suspected abnormalities and provides results as clinically relevant tags. The software can :

Identify normal chest X-rays.
Detect the presence or absence of 23 abnormalities:
Atelectasis, Blunted Costophrenic Angle, Non-Aortic Calcification, Degenerative Spine Conditions, Hyperinflation/ Emphysema, Reticulonodular Pattern, Scoliosis, Tracheal shift, Elevated Hemidiaphragm, Cardiomegaly, Cavity, Consolidation, Fibrosis, Hilar Enlargement, Nodule, Opacity, Pleural Effusion, Pneumothorax, Rib Fractures, Lung Nodule Malignancy Score, Radiological signs of Tuberculosis and Mass.
Measure the distance between the tip of the breathing tube (endotracheal or tracheostomy tube) and carina.¹

Images shown for illustrative purposes only.

Scanning Parameters

AP or PA chest x-ray

Resolution minimum of 1440 x 1440

Gray level at least 10 bits

AI Output

Identify normal chest X-rays.

Detect presence/absence of > 23 abnormalities: Atelectasis, blunted costophrenic angle, non-aortic calcification, degenerative spine conditions, hyperinflation/emphysema, reticulonodular pattern, tracheal shift, elevated hemidiaphragm, cardiomegaly, cavity, consolidation, scoliosis, fibrosis, hilar enlargement, nodule, opacity, pleural effusion, pneumothorax, rib fractures, lung nodule malignancy score, tuberculosis, mass.

Measure distance between the tip of the breathing tube (endotracheal or tracheostomy tube) and carina.

DICOM format outputs: A single series in secondary capture and/or GSPS format is returned to the PACS, without lines and labels of the target abnormalities. The target abnormality outlines and labels are marked out on image series.

PDF formatted text reports: A text report (pdf format) is returned to the PACS indicating the names of all the target abnormalities detected. The .pdf report also contains text indicating the approximate location in the chest for any abnormalities on X-rays. For convenience, two type of PDF reports are presented to the user (free text and structured).

Other Relevant Information

Currently not made available with the Calantic Viewer. Separately distributed by Bayer.

Vendor
Qure.ai Technologies Pvt Ltd.

EU risk class and CE marking
qXR has CE marking (CE0123) and risk class IIb.

Reimbursement status
Not reimbursed.

Contraindications
Not applicable.

Patient Target Group

Tuberculosis: 3 years of age or older.
Normal, Atelectasis, Blunted Costophrenic Angle, Non- Aortic Calcification, Degenerative Spine Conditions, Hyperinflation/Emphysema, Reticulonodular Pattern, Scoliosis, Tracheal shift, Elevated Hemidiaphragm, Cardiomegaly, Cavity, Consolidation, Fibrosis, Hilar Enlargement, Nodule, Opacity, Pleural Effusion, Pneumothorax and Mass: 15 years of age or older.
Rib Fractures, Lung Nodule Malignancy score, Endotracheal and Tracheostomy tubes: > 18 years old.

Intended user
Any physician trained and experienced in the independent review and interpretation of chest X-ray scans. This user may be a radiologist, surgeon, emergency care physician, intensivist or other clinical specialist with the appropriate training.

Limitations

qXR is intended to be used by certified medical specialists trained and experienced to review and report Chest X-ray scans.
qXR is limited to analysis of imaging data and should not be used in-lieu of full patient evaluation or relied upon to make or confirm diagnosis.
The clinician is responsible for reviewing the device findings with the original image for any treatment initiation.
The device shall not be used for scans outside of the compatibility as defined in the instructions of use.
The localization and highlighting of abnormalities are a feature of the product and these markings appear on the image for clinician review, on X-ray/scans where these abnormalities are detected. However, clinical data supporting the accuracy of the localization for each target abnormality is limited.

References

Instructions for Use qXR, software version 4.0

Kaviani, Parisa, Subba R. Digumarthy, Bernardo C. Bizzo, Bhargava Reddy, Manoj Tadepalli, Preetham Putha, Ammar Jagirdar, Shadi Ebrahimian, Mannudeep K. Kalra, and Keith J. Dreyer. 2022. "Performance of a Chest Radiography AI Algorithm for Detection of Missed or Mislabeled Findings: A Multicenter Study" Diagnostics 12, no. 9: 2086. https://doi.org/10.3390/diagnostics12092086