Introduction: The purpose of this study is to construct a software using artificial neural network (ANN) to predict the safe time period of temporary artery occlusion (TAO) during intracranial aneurysm surgery.
Methods: Retrospective data of 105 patients with confirmed aneurysm were used for “learning” process of ANN in anterior circulation area (AcomA and MCA). Pre, intra and post-operative variables used for the input layer of ANN were age, pre and postoperative neurological deficit, Fisher grading score, aneurysm location, diameter of arteries in circle of Willis, clipping time, mean velocity of arteries in Willis circle, and CT and MRI findings. After “learning”, the software was prospectively “trained” (10 patients) and “tested” (10 patients) using the aforementioned variables to predict the safe TAO time in aneurysm surgeries.
Results: In the retrospective analysis of 105 patients, 32 (30.4%) were male and 73 (69.6%) were female (age range 30-72 years) with 80 (76.1%) aneurysms located in MCA and 25 (23.9%) in AcomA. Temporary clipping time varied from 120s-950s. Among 20 patients in prospective dataset analysis, 9 (45%) were male and 11 (55%) were female (age range 30-76 years). Fourteen (70%) had MCA and 6 (30%) had AcomA aneurysms with temporary clipping time ranging from 75s-930s. Safe time evaluated to be higher in AcomA aneurysm surgery rather than MCA aneurysm surgery. Characteristic validity of software was assessed 87% in test cases with 13% relative risk.
Conclusions: The concept of ANN modeling for TAO is introduced for a neurosurgical audience. Authors suggest that the validity (87%) of this software should be tested in larger patient population. Also, with advanced ANN modeling, the clinician may expect to build developed models with more powerful prediction abilities.
Patient Care: Temporary artery occlusion (TAO) during intracranial aneurysm surgery is a fundamental component in facilitating aneurysm dissection and reduces detrimental aneurysm rupture. This procedure can cause ischemia and stroke in certain cases. Using a safe strategy for predicting the safe time-point in TAO is crucial to avoid the unfavorable outcome. For this reason, we have developed a software using artificial neural network to predict TAO safe time-point to overcome the adverse side effects of this procedure.
Learning Objectives: By the conclusion of this session, participants should be able to describe the importance and accuracy of artificial neural network (ANN) in predicting a safe time point for temporary artery occlusion in aneurysm surgery.