Deep Learning System Boosts Radiologist Accuracy and Speed for Head CTs

By MedImaging International staff writers
Posted on 26 Sep 2023

Image: Annalise Enterprise CTB acts like a ‘second pair of eyes’ for radiologists (Photo courtesy of Annalise.ai)

Non-contrast computed tomography of the brain (NCCTB) is a commonly employed method for identifying intracranial pathology. Despite its frequent use, the complex scan outcomes are prone to being misunderstood. Now, a deep learning system acts like a second pair of eyes for radiologists in interpreting these scans and identifies as many as 130 different findings on unenhanced CT scans of the brain in less than two minutes. This is crucial for quick diagnosis and treatment of conditions like strokes or internal bleeding in the brain that require rapid intervention.

The Annalise Enterprise CTB solution from Annalise.ai (Sydney, Australia) is a state-of-the-art radiology AI solution for non-contrast head CT studies. This AI solution is designed to quickly identify up to 130 different radiological findings on non-contrast head CT scans, including conditions that need time sensitive interventions. Now, a pioneering clinical study has demonstrated that this state-of-the-art diagnostic support technology can effectively enhance the performance levels of radiologists.

The multi-reader, multi-case study compared the effectiveness of 32 radiologists in diagnosing clinical issues on non-contrast CT brain scans. Initially, these professionals assessed the cases without the aid of the Annalise Enterprise CTB technology. After a break of at least four months, the same scans were reviewed again, but this time with the support of Annalise's technology. The findings showed that the use of Annalise Enterprise CTB significantly boosted the accuracy of detection and also sped up the interpretation process compared to when the radiologists worked unassisted.

The study demonstrated that the accuracy of radiologists improved by 32% when aided by Annalise Enterprise CTB, and the total time taken for interpretation decreased by 11%. For 91 specific findings, the system significantly increased the accuracy of radiologists, while for the rest, the results were comparable to those of the radiologists working without the Annalise solution. This implies that integrating Annalise Enterprise CTB into practice could offer several advantages, such as lowering the rate of errors, improving efficiency, and enhancing patient care by enabling faster and more accurate diagnoses.

“This important clinical study indicates that radiologist performance improves with diagnostic support from a comprehensive, multi-finding AI solution,” said Annalise Chief Medical Officer Rick Abramson. “The results have important implications for patients and clinicians. We look forward to further measuring and validating them in the clinical trials we continue to conduct around the world.”

Related Links:
Annalise.ai