Prediction of cerebral injury secondary to endovascular recanalization of acute ischemic stroke – PRECISE-STROKE

Target population
LVO-AIS patients successfully treated by EVT. They are enrolled in prospective cohorts and RCTS coordinated by PRECISE investigators. Endpoint is the occurrence of an HT within 24-48 h after stroke treatment following the ECASS-2 criteria: HI2, PH1 and 2 (Hacke et al., Lancet 1998 PMID 9788453)

Radiological Biomarkers: We have performed a complete review of the surrogate biomarkers of ischemia reperfusion injury predictive of the risk of HT. Among those, we have focused our research on hypoperfusion severity; BBB permeability and venous outflow from perfusion imaging collected before the achievement of a successful endovascular reperfusion. In addition, we are investigating other biomarkers such as specific DWI signal abnormality (Dark Diffusion Sign of the caudate nucleus). Multimodal MRI and CT Perfusion are systematically reprocessed in collaboration with our partner the Stanford Stroke Center (Pr GW Albers), and analyzed by our imaging WP group coordinated by Dr P Seners.

Deep Thrombo Immune profiling: As part of our collaboration with Stanford University IA and immunology department coordinated by Pr B Gaudillière we are performing systematic high dimensional profiling of the thrombo immune response using single cell cytometry to identify systemic thromboimmune activation profiles associated with HT after successful EVT using the Stabl machine learning method. We have finished a proof-of-concept study enrolling 10 patients who experienced HT after successful EVT and 10 who did not. We did perform systematic peripheral blood collection before thrombectomy (Groin puncture) at the end of the procedure and 24 h after. Stabl identified sparse neutrophil activation profiles in the samples collected before EVT associated with HT and we are replicating these analyses in a prospective cohort of radiologically phenotyped 50 patients. If validated we will then prospectively investigate the yield of a combined modelling between imaging and biology in the treatment and control arm of RCTs coordinated by our consortium members. They are all testing treatment complementary to EVT to mitigate the thromboinflammatory response. Some are evaluating innovative antithrombotic approaches (antiplatelets IA thrombolysis) other Dornase Alpha which is mitigating the consequences of PNN activation by dissolving the NETS. This will allow us to monitor the impact of this treatment on the relationships observed between radiological and biological profiles and HT. There has been substantial change in the list of studies previously mentioned in the grant application. FRAME LP and GREEN stopped for futility. But several patients enrolled in GREEN were sampled before study termination. MR Reperfusion has not started yet. But we are fortunate to expand our enrollment to the Net Target study that did enroll patients treated by IVT and EVT who did receive Dornase alpha, a treatment that reduces Netosis a strong trigger of the thrombo inflammatory response.

Imaging
We have assessed blood brain barrier permeability on the perfusion MRI performed before successful EVT in the FRAME study (PI JM Olivot) We have demonstrated for the first time that BBB damage can be detected before EVT and was predictive of HT occurrence after endovascular reperfusion. (Leigh et al. Annals Clin Transl Neurol 2024 PMID 39462241, IF=4). We also have coined a new biomarker, the Dark Diffusion Sign, on pretreatment DWI imaging that was associated with the occurrence HT after successful EVT in the same cohort. The paper has been accepted and is soon to be published in the review AJNR (IF 3). We are beginning new analyses investigating the severity of hypoperfusion and venous outflow.

Deep Thrombo Immune Profiling
In collaboration with our partner, Pr B Gaudillière from Stanford university, the department of AI and immunology has validated a machine learning processing of high dimensional omics such as single cell cytoF output that identify specific immunological activation profiles predictive of meaningful clinical endpoint such as infections rate after surgery, still birth etc… We have replicated this approach in patient undergoing an EVT and identified in the peripheral blood sample before EVT a specific Neutrophil activation profile associated with the occurrence of HT. We are reproducing the same type of analyses in a larger cohort of 50 patients. Analyses are ongoing. If the results are replicated a paper is expected for 2026.

Training
Over the past 2 years we have recruited 2 research fellow and 1 engineer who will work on the data of the PRECISE. We have organized with StrokeLink network a research seminar in September in Toulouse on the topic of cerebral ischemia and inflammation.

Valorization of PRECISE STUDY RESULTS

1- Creation of a network of excellence on the subject of thrombo-inflammatory response and ischemic reperfusion lesion. In 2024, we successfully obtained the MRSEI funding to support our applications. We used it to organize seminar and meeting with KOL in the field of neuroimmunology Pr J Kipnis, (St Louis, USA); Pr A Planas (Barcelona, Spain); and K Akassoglou (UCSF USA). We applied for the call for funding Network of Excellence from the Leducq Foundation. We were ranked within the first 20%. We did not get funded, but benefitted from a full review limited to the first 25% with important recommendations that will be used for our next application in 2026. Due to the current uncertainty regarding research funding in the USA, we are also exploring applications for ERC funding.
2- Creation of a clinical research platform: DEEP CARE PROJECT dedicated to high dimensional phenotyping of patients evaluated for an acute cerebral injury. It will be created at the interface with our CIC 1436 (PRECISE COORDINATING Structure), Toulouse Acute ICU and Stroke Unit and new INSERM team being created to host IA analyses. For this purpose, we obtained in 2024 3 relevant fundings.
a. LABCOM AI-NIGMA between our INSERM Team and the Start Up Surge. Coordination of the WP AI of the MESSIDORE SAVE BRAIN.
b. Coordination of the WP Molecular imaging of the RHU TIPICH.
c. We are now applying for Occitanie Region support as well as INSERM Program.
3- Deliverables
a. Point of Care Biological Panel to tailor LVO-AIS acute management, using combined OLINK and CytoF analyses
b. Identification of targetable pathways to be manipulated in validated animal model (eg. The SPAN project in the US as part of a collaboration with Pr P Lyden and the StrokeNet Network).
c. Expansion of this approach to other stroke subtypes: Subarachnoid hemorrhage, intracerebral hemorrhage, Post Cardiac Arrest Anoxic Brain Injury,

Deadline extension: We aim to finish the PRECISE project in 2028; 1 year after the planned end of the project. We will therefore request next year a 1-year expansion. This delay is caused by the termination of 2 studies for futility and the delay initiating others that will be detailed in our next application for extension.

Leigh et al. Annals Clin Transl Neurol 2024 PMID 39462241, IF 4
Seners et al. Stroke 2023 ; PMID : 37334709, IF= 7.8
Lyon et al 2025 AJNR In Press (IF 3.5)

Endovascular therapy (EVT) has revolutionized the management of acute ischemic stroke due to large vessel occlusion (LVO-AIS). However, despite the 90% recanalization rate, over 50% of patients suffer hemorrhagic transformation (HT), the major modifiable predictor of poor functional recovery. Accurate prediction of HT would transform management by allowing personalized treatment. Our network’s experimental data suggest persistent microvascular thrombosis and blood-brain-barrier (BBB) leakage triggering thrombo-inflammation are critical factors in the pathogenesis of HT. The goal of the PRECISE-Stroke network is to identify a predictive signature of HT after EVT-recanalization by pursuing three aims: 1) to determine the role of microcirculation thrombosis (Hypoperfusion and alteration of venous outflow) and BBB leakage in HT using pre- and post-EVT clinical perfusion imaging, 2) to determine the contribution of platelets and immune cell-specific events to HT pathogenesis using high-content, single-cell proteomic analyses of patient blood and clot samples retrieved during EVT and established preclinical stroke models, 3) to train and validate a predictive model of HT using high-dimensional statistical methods integrating clinical, imaging, and biological data. Key deliverables will be 1) a robust, non-invasive, actionable predictive signature of HT after EVT-recanalization and 2) new mechanistic biomarkers to guide the development of new therapeutic interventions. The proposed studies mobilize multidisciplinary, senior and junior investigators with combined expertise in stroke trials, preclinical stroke models, systems-immunology, and artificial intelligence (AI). The PRECISE-Stroke network will provide a unique opportunity to train and mentor budding stroke researchers in an emerging multidisciplinary field at the interface of imaging, systems-biology, and AI.