Radiologists reporting on CT with AI save time and reach higher agreement

New clinical research into the impact of computer-aided detection (CAD) in routine clinical practice has been published today in the European Journal of Radiology Open. The study with two radiologists shows that reporting pulmonary nodules on CT using Aidence’s Veye Lung Nodules reduces reading time by an average of ~ 40%. It also indicates a striking improvement in agreement when aided by Veye.

The retrospective study was performed at Spaarne Gasthuis, a large teaching hospital in Hoofddorp, the Netherlands. Two radiologists independently assessed 50 chest CT scans for incidental pulmonary nodules twice, with six months in between sessions.

Their readings were first aided, then unaided by Veye Lung Nodules, an AI solution that automatically detects, measures, classifies, and tracks the growth of pulmonary nodules.

Reading time

A significant reduction in reading time was observed for both readers: 33.4% and 42.6%. Overall, the radiologists reported fewer nodules when using the AI system. Yet even when an equal number of nodules were reported, the time-saving persisted. The volumetric measurements provided by Veye Lung Nodules, which include growth percentage and volume doubling time from the most recent prior, also contributed to the reduction.

Interobserver agreement

The study further showed a notably higher agreement in the patient management recommendations given by the radiologists when supported by Veye, namely by a linear weighted kappa of 0.84. The AI-based measurements (diameter and volume), volume doubling times, and composition may have mitigated some of the largest sources of reader disagreements.

Conclusions

The study suggests that Veye Lung Nodules could contribute to more uniform patient management recommendations. These would “allow for more robust and effective triaging in clinical practice and lung cancer screening programmes”, the authors wrote.

Due to the experimental setting, it is more challenging to extrapolate the time reduction benefits to clinical practice. The decrease in reporting time is, nonetheless, consistent with other published and anecdotal evidence from radiologists using Veye.

About Aidence

Founded in 2015, Aidence rallies over 80 data scientists, software engineers, medical, regulatory, and business professionals to provide intelligent software for the lung cancer pathway. The company is part of the AI division of RadNet, a US-based leading provider of diagnostic imaging services.

2048 1366 Partner Content
My name is HAL 9000, how can I assist you?
This website uses cookies to ensure the best possible experience. By clicking accept, you agree to our use of cookies and similar technologies.
Privacy Policy