Revamping Delano’s classification of optic nerve variants using AI powered grading

This study aims to evaluate the diagnostic accuracy of an artificial intelligence based application in classifying the course of the optic nerve according to Delano classification on computed tomography imaging of paranasal sinuses. Identification of optic nerve course is important before functional endoscopic sinus surgery because certain anatomical variations increase the risk of surgical injury and permanent visual loss. This observational cross sectional study will be conducted in the Department of Radiodiagnosis at Saveetha Medical College and Hospital Chennai among adult patients undergoing CT scan of paranasal sinuses. A total of 500 CT scans will be included. Each CT scan will be reviewed by experienced radiologists and classified according to Delano classification which will serve as the reference standard. The same CT images will be analyzed using an artificial intelligence based mobile application which uses convolutional neural network technology for automated classification of optic nerve variants. The diagnostic performance of the artificial intelligence application will be assessed by calculating sensitivity specificity positive predictive value negative predictive value and area under receiver operating characteristic curve. The agreement between artificial intelligence based classification and radiologist interpretation will be evaluated. The study will also assess the distribution of Delano types and associated anatomical variations such as optic nerve dehiscence Onodi cell and pneumatized anterior clinoid process on CT imaging. The results of this study may help in improving preoperative radiological assessment and surgical planning in patients undergoing endoscopic sinus surgery.