ContaCT

Viz ContaCT is an AI-powered stroke triage tool that utilizes deep learning to analyze CTA scans in real time for the detection of large vessel occlusions. By segmenting cerebral vessels and identifying absent blood flow, it rapidly notifies specialists to expedite evaluation for thrombectomy, thus improving the efficiency of acute stroke care.

Viz ContaCT (Viz LVO) is an advanced artificial intelligence tool specifically engineered for rapid stroke triage. Its core function is to analyze Computed Tomography Angiography (CTA) scans in real time, employing a sophisticated deep-learning model to meticulously segment cerebral vessels and identify potential large vessel occlusions (LVOs) by detecting areas characterized by an absence of blood flow, which are indicative of a stroke. This AI-driven approach aims to streamline the critical initial stages of stroke diagnosis by automating the detection process. Upon the detection of a suspected LVO, Viz ContaCT instantly alerts neurovascular specialists through a dedicated mobile application, the Viz app, providing a low-resolution scan preview to facilitate an accelerated evaluation for potential thrombectomy. The system's clinical utility is supported by validation studies demonstrating approximately 97% sensitivity and 74% specificity in identifying anterior circulation LVOs. By seamlessly integrating into the acute stroke workflow as a crucial notification aid, Viz ContaCT effectively reduces the time from scan to specialist assessment, thereby expediting the initiation of life-saving treatment for stroke patients and improving overall care efficiency.

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 Viz ContaCT's performance (sensitivity and specificity) vary when analyzing posterior circulation occlusions compared to its validated performance for anterior circulation LVOs?; What are the typical characteristics of false positive and false negative cases identified by Viz ContaCT, and what strategies are being implemented to further refine the deep-learning model's accuracy?; Beyond reducing time to specialist assessment, are there studies quantifying the impact of Viz ContaCT implementation on long-term patient functional outcomes (e.g., modified Rankin Scale scores at 90 days)?; How robust is the Viz ContaCT algorithm to variations in CTA image quality, scanner types, and contrast administration protocols across different medical institutions?; What is the learning curve for neurovascular specialists using the Viz app, and how does the low-resolution preview compare to full DICOM viewers in terms of diagnostic confidence for thrombectomy decisions?