EFAI Neurosuite CT Midline Shift Assessment System (MLS-CT-100)

The EFAI Neurosuite CT Midline Shift Assessment System is a deep learning-based radiological computer-aided triage software designed to automatically identify features suggestive of midline shift on non-contrast head CT scans. Its purpose is to provide passive notifications for worklist prioritization, effectively enhancing diagnostic workflow without replacing clinician judgment, as validated by robust clinical performance data.

The EFAI Neurosuite CT Midline Shift Assessment System (MLS-CT-100) from Ever Fortune.AI, Co., Ltd. was reviewed by the U.S. Food & Drug Administration and determined to be substantially equivalent for its intended uses. This system functions as radiological computer-aided triage and notification software, designed to streamline diagnostic workflows by identifying critical cases. This system employs deep learning algorithms to automatically analyze non-contrast head CT scans in adults (18+) for features indicative of midline shift (MLS). It provides passive, case-level notifications to PACS/workstations to facilitate worklist prioritization or triage, without being intended for stand-alone clinical decision-making or directing attention to specific image areas. Extensive performance data, including a multi-site clinical validation study, demonstrated high sensitivity and specificity for MLS detection, supporting its regulatory clearance despite minor differences from predicate devices.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. What specific deep learning architecture and training data were utilized to achieve high sensitivity and specificity for midline shift detection?; How does the system's performance for midline shift detection vary across different etiologies of intracranial pathology (e.g., hemorrhage, mass effect, edema)?; Given its role in worklist prioritization, what is the measured impact of the system on diagnostic turnaround times and clinician workflow efficiency in a real-world setting?; Are there any specific limitations or edge cases where the system's performance might degrade, such as in the presence of severe artifacts or complex anatomical variations?; What post-market surveillance or continuous improvement strategies are planned to adapt the algorithm to evolving clinical data or imaging standards?