RapidAI has announced the publication of TIMELESS—a randomised trial evaluating the use of tenecteplase in acute ischaemic stroke treatment—in the New England Journal of Medicine (NEJM), in addition to multiple other data from research studies unveiled at this year’s International Stroke Conference (ISC; 7–9 February, Phoenix, USA).
As per a press release from the company, these data “underscore and reinforce the pivotal role of the RapidAI platform in advancing stroke care”.
“In the world of AI [artificial intelligence], clinical validation is paramount to instil confidence in clinician decision-making,” said Karim Karti, chief executive officer of RapidAI. “Securing another NEJM publication and sharing results from more than a dozen studies at ISC 2024 reaffirms our lasting and ongoing contribution to enhancing stroke care.”
TIMELESS study
Having presented preliminary findings from the Genentech-sponsored TIMELESS trial at the 2023 European Stroke Organisation Conference (ESOC; 24–26 May, Munich, Germany), Gregory Albers (Stanford Medicine, Stanford, USA) delivered subgroup analyses from the trial last week—which were simultaneously published in NEJM.
The TIMELESS trial compared tenecteplase with placebo in large vessel occlusion (LVO) acute ischaemic stroke patients presenting in a later time window (4.5–24 hours from symptom onset) since last known well.
TIMELESS is the first study to show that an intravenous thrombolytic drug can be given up to 24 hours after time last known well in LVO stroke patients without causing an increase in brain haemorrhage, as per RapidAI’s release. And, while clinical outcomes in the trial were not significantly different with tenecteplase compared to placebo, encouraging subgroups discussed by Albers at ISC 2024—such as patients with proximal middle cerebral artery (MCA) occlusions or those treated at ‘spoke sites’ prior to transfer for thrombectomy—have been identified for future studies.
Key technologies from RapidAI, including Rapid Computed Tomography Perfusion (CTP) and Rapid Magnetic Resonance Imaging (MRI) Perfusion Maps, were leveraged for all patient selection in the trial. According to the company’s recent release, key imaging criteria for study inclusion—including target mismatch ratio of ≥1.8, mismatch volume of ≥15ml, and ischaemic core volume <70ml—were assessed based on RapidAI software.
Additional research
Another piece of research presented at ISC related to the detection of medium vessel occlusions (MeVOs) and collateral assessment with a multimodality AI approach from Mohammad Mahdi Sowlat (Medical University of South Carolina, Charleston, USA) et al.
Of 149 patients whose scans were retrospectively analysed, 27.3% of MeVOs were successfully identified by Rapid LVO technology alone, and another 25.3% by Rapid Vessel Density Maps, meaning a total of 52.6% of MeVOs were identified by the two combined. This signals the potential benefit of using Vessel Density Maps to flag MeVO patients for review at centres without CTP capabilities, according to RapidAI. In addition, 93.3% of MeVOs were identified when utilising Rapid LVO, Rapid CT Angiography (CTA), and Rapid CTP.
Elsewhere, as per an abstract from Tim Malisch (Ascension Alexian Brothers, Elk Grove Village, USA) et al entitled, “Improved stroke outcomes following implementation of RapidAI platform at Ascension-Illinois”, 1,732 acute ischaemic stroke patients were analysed based on the Get With The Guidelines Stroke database.
This research sought to compare findings before and after the implementation of RapidAI, and data showed a median reduction of 30 minutes per patient from door-to-mechanical thrombectomy after the integration of RapidAI. Post-RapidAI integration, there was also an increase in thrombectomy procedure volume from 9.6% to 14.5%, and Rapid technologies were associated with optimised thrombectomy workflow as well as improved modified Rankin scale (mRS) scores at discharge.
Finally, 2023 findings were presented from a national database demonstrating “substantial delays” between non-contrast CT (NCCT) and CTA in suspected stroke patients from Christina Mijalski Sells (Stanford University Medical Center, Palo Alto, USA) et al.
Data on 23,935 patients from 717 US hospitals using CT and CTA were processed with RapidAI technology to determine time delays between the modalities. Some 18% of all patients had a delay of more than 15 minutes between NCCT and CTA, and 6% had a delay of more than 30 minutes. As per RapidAI’s release, even among high-performing hospitals, many patients had a >15-minute delay between modalities—reaffirming the potential operational and clinical value of Rapid’s NCCT suite in identifying LVOs on NCCT scans, and substantially reducing door-to-diagnosis time.
