Rapid ASPECTS

Rapid ASPECTS is a computer-aided diagnosis software that assists clinicians by automatically analyzing non-contrast CT images of the brain to generate an Alberta Stroke Program Early CT (ASPECT) Score for patients with suspected acute stroke due to MCA or ICA occlusion. The system employs a random forest machine learning algorithm to perform image registration, segmentation of regions of interest, and analysis of morphological characteristics to detect early ischemic brain injury within a six-hour window.

Rapid ASPECTS is a computer-aided diagnosis (CADx) software designed to assist clinicians in the assessment and characterization of brain tissue abnormalities using non-contrast computed tomography (NCCT) image data. Its primary purpose is to evaluate patients presenting for diagnostic imaging workup with known Middle Cerebral Artery (MCA) or Internal Carotid Artery (ICA) occlusion by automatically generating an Alberta Stroke Program Early CT (ASPECT) Score. The software registers images, segments and analyzes predefined ASPECTS Regions of Interest (ROIs), and extracts image data based on morphological characteristics to provide computer analytics. This process, utilizing an artificial intelligence algorithm, synthesizes imaging features into a single ASPECT score to help characterize early ischemic brain tissue injury (within 6 hours) and the extent of disease, with the understanding that it is an assistive tool and not for primary image interpretation. The system's methodology relies on a machine learning implementation, specifically a random forest technique, and a processing pipeline that includes image orientation and spatial normalization, delineation of ROIs with computation of voxel values, identification of previous/old stroke areas alongside early ischemic changes, and labeling of these regions to provide a summary score as per ASPECTS guidelines. This score can be reviewed and adjusted by the clinician via the Rapid Platform User Interface before being sent to PACS. Performance data from a reader study involving 50 cases and 8 readers indicated that Rapid ASPECTS significantly improved reader agreement with expert consensus (P<0.0001), with readers, on average, agreeing with almost half a region (0.425) more per scan. Notably, non-neuroradiologists' agreement improved from 73.6% to 79.8%, reaching levels comparable to expert neuroradiologists. The device has specific limitations, including its validation for use with known MCA/ICA occlusions, within 6 hours of symptom onset, and on GE Lightspeed VCT Scanners.

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. How does the 'random forest machine learning technique' specifically contribute to the accuracy of ASPECT scoring compared to other AI algorithms, and what is its robustness against variations in image quality?; What is the performance variability and reliability of Rapid ASPECTS when used with CT scanners from manufacturers other than the validated GE Lightspeed VCT Scanners, or with different scanning protocols?; Beyond the initial 6-hour window for early ischemic brain tissue injury, how rapidly does the diagnostic accuracy of Rapid ASPECTS decline, and are there any time-dependent biases in its scoring?; What specific 'morphological characteristics' are extracted and weighted by the AI algorithm to determine the ASPECT score, and how sensitive is the system to subtle early ischemic changes versus chronic lesions or artifacts?; How does the device's performance in identifying and highlighting 'previous/old stroke areas' versus 'areas of early ischemic change' impact the final ASPECT score, particularly in patients with complex cerebrovascular histories?