Viz.ai and RapidAI—two leading companies developing artificial intelligence (AI)-powered technologies intended to optimise stroke care—have separately announced positive study data on their AI solutions.
Earlier this week, Viz.ai announced results from a clinical study recently published in The Neurohospitalist highlighting the “transformative impact” of the Viz Connect care pathway on post-stroke patient care. The study demonstrates that the incorporation of Viz Connect drove improvements in cardiac monitoring utilisation, care efficiency, healthcare provider (HCP) experience, and patient satisfaction, according to a company press release.
These data stem from a recent collaboration between Viz.ai and Medtronic to leverage AI-driven solutions to streamline communication between neurology and cardiology teams, and accelerate critical interventions for stroke patients.
Viz.ai claims that Viz Connect was designed to improve coordination between stroke specialists and electrophysiologists with the goal of enhancing post-stroke atrial fibrillation (AF) assessments. To evaluate the impact of Viz Connect in this space, lead author Brett Meyer (University of California San Diego [UCSD], San Diego, USA) spearheaded a single-centre study comparing Viz Connect to existing processes. The study assessed the platform’s ability to improve patient cardiac monitoring, reduce monitoring device placement times, and optimise HCP and patient experience.
“Viz Connect has helped us improve how we coordinate care for stroke patients at risk of atrial fibrillation,” Meyer commented. “By replacing a complex and fragmented system that relied on electronic medical record consults and direct physician-to-physician communication with Viz.ai’s streamlined platform, we achieved a robust increase in cardiac monitoring placements prior to discharge. The platform’s closed-loop communication, real-time alerts and ease of use have not only improved clinician workflows but also enabled more meaningful patient discussions, paving the way for better risk reduction strategies and improved outcomes.”
Specific results from the study are as follows:
- Increased access to care—the number of patients receiving cardiac monitoring overall increased by 1.3 times after adopting Viz Connect, while the number of patients receiving guideline-driven cardiac monitoring specifically increased by 8.4 times with Viz Connect, demonstrating its impact on advancing the standard of care.
- Improved efficiency—median time to device placement decreased by 97% (from 32 days to one day) for insertable cardiac monitors (ICMs) and decreased by 59% (from 22 days to nine days) for external event monitors (EMs). Viz.ai claims that improvements in median time to care were largely driven by Viz Connect’s ability to facilitate inpatient cardiac monitoring before discharge. Following Viz Connect implementation, the number of patients receiving appropriate ICM placement before discharge increased by 17 times as well.
- Enhanced HCP and patient experience—100% of surveyed clinicians preferred Viz Connect over the previous process, and patient satisfaction scores improved once Viz Connect was implemented.
Elsewhere, RapidAI noted in a press release of its own that data from a head-to-head study have found that RapidAI’s large vessel occlusion (LVO) detection technology was able to outperform Viz.ai’s equivalent solution by 33%, detecting more stroke cases and reducing the risk of missed diagnoses.
In what is said to be the largest direct comparison of the two companies’ technologies to date, involving 1,591 consecutive patients, the DUEL study evaluated the performance of Rapid LVO versus Viz LVO in the detection of LVOs on computed tomography (CT) angiography.
DUEL data demonstrate the superiority of Rapid LVO compared to Viz LVO, according to RapidAI’s release, with the company’s technology detecting 33% more LVO-positive cases (98% vs 74%). The release also details that Viz LVO failed to detect 26% of LVO-positive cases, while Rapid LVO correctly identified 94% of LVO-negative cases versus 91% with Viz LVO.
“Trusting the accuracy of clinical AI tools is critically important, especially for time-sensitive conditions like stroke,” said lead author Harmeet Sachdev (Good Samaritan Hospital, San Jose, USA). “By comparing two leading AI-driven tools, we now have a clearer understanding of performance differences, including important distinctions in sensitivity and specificity. Our findings raise the concern about missing LVO detection, and causing potential delays in diagnosis and treatment in a substantial number of patients.”
A second study highlighted in RapidAI’s release evaluated the performance of RAPID SDH for the detection of both acute and chronic subdural haematomas (SDHs), with the accuracy of Rapid SDH being compared to gold-standard expert readers (n=313). As noted by the company, the results indicate that Rapid SDH provides fast and accurate detection of acute and chronic SDH on non-contrast CT, demonstrating a sensitivity of 92.4% and a specificity of 98.7%, with a median processing time of approximately 45 seconds.
These findings on Rapid SDH, and data from DUEL, were delivered along with several other studies at the ongoing International Stroke Conference (ISC; 5–7 February, Los Angeles, USA), according to RapidAI’s release.
