This is a special version of VUNO’s Chest X-ray model designed to screen for pneumonia, TB, malignancy, and interstitial lung disease based on a combination of four major thoracic findings - nodule/mass, consolidation, interstitial opacity, and pleural effusion detected by the model.

Intended Use

To screen for pneumonia, TB, malignancy, and interstitial lung disease based on a combination of four major thoracic findings - nodule/mass, consolidation, interstitial opacity, and pleural effusion detected by the model.

Limitations

This model provides binary results (Abnormal or Normal) without naming the lesion that triggers such results. Thus, the “Abnormal” result may be caused by one or a combination of those four findings or other patterns similar to them.

Information on training data

A total of 20,000-100% CT confirmed chest PA X-ray images

Model performance metrics

[Image basis]
AUROC : 0.985
[Lesion basis]
JAFROC : 0.943.

Performance Curves

curve2

curve1

curve3

AUROC(classification) and JAFROC(detection/segmentation) of Vuno's AI model. 'Pooled observer' is average performance of 9 radiologists.

Model History

Stability and Performance improvements.

Vuno

Vuno Inc.

VUNO’s AI solutions are reshaping the medical industry presenting a whole new level of experience to medical practitioners in their day-to-day workflow, enabling them to make diagnostic decisions faster and more accurately and to provide quality patient care. We are on a mission to “View the Invisible, Know the Unknown”, and will never rest until we discover the invisible and unknown territories of medicine by navigating the ambiguity with the help of artificial intelligence.

Woong Bae

Woong Bae

Email

Beomhee Park

Beomhee Park

Email

Minki Chung

Minki Chung

Email

Seo Taek Kong

Seo Taek Kong

Email

Learn more