Computational Medical Imaging and Prediction of Diffusion/FLAIR Mismatch in Stroke Patients

Active, not recruiting

• Conditions: Stroke

• Registration Number: NCT05192161
• Lead Sponsor: Fondation Hôpital Saint-Joseph

Brief Summary

Stroke is a public health issue and a priority for our institution. MRI plays an essential role in the management of stroke. In this context, the contribution of MRI is diagnostic, etiological and prognostic.

Among the MRI parameters evaluated in the acute phase of the stroke, the evaluation of the mismatch between the DIFFUSION and FLAIR sequences is crucial as it will directly contribute to the therapeutic decision. A FLAIR-diffusion mismatch, i.e., a lesion with a diffusion but not a FLAIR hypersignal, identifies patients whose time of onset of symptoms is probably less than 4.5 hours. It is therefore understandable that the main arterial recanalization techniques performed in the acute phase are primarily reserved for patients with a positive mismatch.

In current practice, mismatch assessment is performed subjectively, by visually comparing the two sequences, which is known to be the cause of a lack of reproducibility and diagnostic performance.

Computational medical imaging techniques ("radiomics") have recently gained momentum and offer the prospect of automated and therefore more reproducible analysis of medical imaging data. In stroke patients, radiomics extracted from FLAIR imaging could thus contribute to describe the "diffusion flair" mismatch in a continuous and objective way.

For the time being, data analysis cannot be performed in real time due to technical constraints. If it is proven that radiomics can reliably analyze the mismatch on the FLAIR sequence alone, the next step will be to make the analysis feasible in clinical routine (i.e. in a time frame adapted to the therapeutic management).

Detailed Description

Not available

Study Timeline

• Study First Submitted: 2021-12-30
• Study First Submitted QC: 2021-12-30
• Study First Posted: 2022-01-14
• Study Start: 2023-04-05
• Primary Completion: 2023-05-05
• Status Verified: 2023-09
• Last Update Submitted: 2023-09-11
• Last Update Posted: 2023-09-13
• Study Completion: 2023-12-31

Recruitment & Eligibility

• Status: ACTIVE_NOT_RECRUITING
• Sex: All
• Target Recruitment: 35

Inclusion Criteria

• Patient with age ≥ 18 years
• Patient with supra-tentorial stroke treated by mechanical thrombectomy and who had MRI imaging on admission before endovascular treatment at Paris Saint-Joseph Hospital
• Complete MRI protocol including Diffusion and FLAIR sequences
• French-speaking patient

Exclusion Criteria

• Patient under guardianship or curatorship
• Patient deprived of liberty
• Patient under court protection
• Patient objecting to the use of his data for this research

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
ability of a computational medical imaging technique ("radiomics"), performed on the FLAIR sequence alone, to predict the FLAIR diffusion mismatch in patients with ischemic stroke	Day 1	This outcome corresponds to FLAIR / Diffusion mismatch scores estimated with both techniques: automated and visual analysis.

Secondary Outcome Measures

Name	Time	Method
Inter-technique comparison of mismatch assessment between automated "radiomics" analysis and subjective visual analysis (currently performed in clinical routine)	Day 1	This outcome corresponds to the reproducibility and inter-technical agreement on the extent of the mismatch.
Evaluation of the prognostic value of the scores obtained with the two methods	Month 3	This outcome corresponds to the correlation of the scores obtained with the Rankin score at 3 months.

Trial Locations

Locations (2)

CHRU Lille; 🇫🇷 Lille, France

Groupe Hospitalier Paris Saint-Joseph; 🇫🇷 Paris, France