A recent study has demonstrated that a novel approach using advanced optical imaging combined with an artificial intelligence algorithm can produce accurate, real-time intraoperative diagnoses of brain tumors.
The accuracy of this innovative imaging and machine-learning technique was compared with that of pathologist interpretation using conventional images for tumor diagnosis. Both methods displayed similar results, with the artificial intelligence-based diagnosis being 94.6% accurate, compared to 93.9% for the pathologist-based interpretation.
The artificial intelligence system was refined using a deep convolutional neural network with more than 2.5 million samples from 415 patients. The system is able to produce a predicted brain tumor diagnosis in under 2.5 minutes.
Senior author Daniel Orringer (New York University Grossman School of Medicine, NY, USA) commented: “As surgeons, we’re limited to acting on what we can see; this technology allows us to see what would otherwise be invisible, to improve speed and accuracy in the [operating room], and reduce the risk of misdiagnosis.”