CT-Based Quantitative Analysis for Pathological Features Associated With Postoperative Recurrence and Potential Application Upon Artificial Intelligence: A Narrative Review With a Focus on Chronic Subdural Hematomas

Abstract

Chronic subdural hematomas (CSDHs) frequently affect the elderly population. The postoperative recurrence rate of CSDHs is high, ranging from 3% to 20%. Both qualitative and quantitative analyses have been explored to investigate the mechanisms underlying postoperative recurrence. We surveyed the pathophysiology of CSDHs and analyzed the relative factors influencing postoperative recurrence. Here, we summarize various qualitative methods documented in the literature and present our unique computer-assisted quantitative method, published previously, to assess postoperative recurrence. Imaging features of CSDHs, based on qualitative analysis related to postoperative high recurrence rate, such as abundant vascularity, neomembrane formation, and patent subdural space, could be clearly observed using the proposed quantitative analysis methods in terms of mean hematoma density, brain re-expansion rate, hematoma volume, average distance of subdural space, and brain shifting. Finally, artificial intelligence (AI) device types and applications in current health care are briefly outlined. We conclude that the potential applications of AI techniques can be integrated to the proposed quantitative analysis method to accomplish speedy execution and accurate prediction for postoperative outcomes in the management of CSDHs.

Endoscopically observed outer membrane of chronic subdural hematoma after endovascular embolization of middle meningeal artery

“…It is difficult to expect the atrophic brain to re-expand itself for clearance of the subdural space. [ 14 ] The stabilization of CSDH from the intracranial space into a subdural fluid capsule is another treatment goal. Although intracranial hypotension in elderly patients cannot be compensated by MMA embolization, the recurrence rate should be reduced if the inflammation resolved.…”

Machine Learning in Neuroimaging of Traumatic Brain Injury: Current Landscape, Research Gaps, and Future Directions

“…Kellogg (2021) trained CNN models for both pre-and post-operative cSDHs, achieving a DICE score of 0.83 in predicting cSDH volumes [92]. Moreover, insights from a study by Kung et al highlight the capabilities of machine learning in predicting post-operative recurrence of SDHs by analyzing specific pathological features [93].…”

“…The status of the cisterns on CT was among the most reliable features for predicting in-hospital mortality, demonstrating the utility of incorporating independent imaging features into these models [91]. Further research is warranted in areas like chronic subdural hematomas (cSDHs) [92,93] and the application of structural MRI for TBI detection [99].…”