IMPROVE score

Individualised MRI based PRediction scOre using plaque Vulnerability for symptomatic carotid artEry disease patients (IMPROVE) is a novel decision rule for symptomatic patients with 30 – 99% carotid stenosis that may assist in selecting patients for revascularisation.¹ IMPROVE incorporates IPH, stenosis severity, age, sex, and classification of last event (cerebral vs. ocular) to predict the overall risk of ipsilateral ischaemic stroke recurrence within 3 years. IMPROVE is the first model that incorporates plaque vulnerability to predict ipsilateral ischaemic stroke risk on contemporary optimal medical therapy with adequate power (≥ 10 events per variable). Bootstrapping based internal validation demonstrated strong predictive performance and stability, with an optimism corrected C statistic of 0.82.¹ 

Previous qualitative research, based on semi-structured interviews with 40 Dutch clinicians involved in decision-making for carotid revascularization, examined clinicians’ perspectives on the IMPROVE decision rule and found that the tool was generally well received, while underscoring the need for clinical evaluation.²

A model based impact study illustrated the potential benefits of IMPROVE based selection of patients for revascularisation, compared with stratification based on current guidelines. Stratification selection for revascularisation of patients with an IMPROVE calculated 3 year ipsilateral ischaemic stroke risk of ≥ 10% led to a cost effective 35% reduction in the composite of peri-operative death and stroke and ipsilateral ischaemic stroke recurrence within 3 years compared with stratification based on current guidelines.³ Even at the lower bound of the 95% CI of the IMPROVE model, probabilistic analyses showed IMPROVE’s superiority over standard care.³

Application of IMPROVE in clinical practice may require additional carotid plaque MRI. In approximately half of the patients, MRI is required for the decision rule, while in the other half, it is unnecessary because the stroke risk is already determined to be either high or low based on the other predictors.

1. Nies KPH, Smits LJM, van Kuijk SMJ, Hosseini AA, van Dam-Nolen DHK, Kwee RM, et al. Individualized MRI-Based Stroke Prediction Score Using Plaque Vulnerability for Symptomatic Carotid Artery Disease Patients (IMPROVE). Stroke. 2025.

    

2. Bierens J, Ament SMC, Truijman MTB, de Borst GJ, Nederkoorn PJ, Bos D, et al. Plaque Magnetic Resonance Imaging Based Decision Rule for the Selection of Symptomatic Patients for Carotid Revascularisation: Clinician Perspectives on Acceptability and Implementation Barriers in the Netherlands. Eur J Vasc Endovasc Surg. 2025.

    

3. Nies K, Ramaekers B, Auer D, Schindler A, Saam T, Bos D, et al. Model-based cost-effectiveness of the IMPROVE score versus care as usual for risk-based selection of symptomatic patients for carotid revascularization. Res Sq. 2025.