NON-INVASIVE PRESSURE GRADIENT ESTIMATION OF STENOTIC LESIONS IN RENAL MULTIFOCAL FIBROMUSCULAR DYSPLASIA: A PILOT STUDY

Tom Robberechts,Alessandro Giudici,Jacek Kadziela,Maria S Stoenoiu,Kim M. Donkers,Jojanne E. Van Leeuwen,Andrzej Januszewicz,Magdalena Januszewicz,Bart Spronck,Alexandre Persu
DOI: https://doi.org/10.1097/01.hjh.0001021568.82711.41
IF: 4.9
2024-05-01
Journal of Hypertension
Abstract:Objective: Currently, there is no validated method to assess the haemodynamic significance of renal artery stenosis in fibromuscular dysplasia (FMD) patients. A translesional pressure gradient >10 mmHg during catheter-based angiography has been proposed as a cut-off for considering angioplasty. However, for reasons of invasiveness, time, accessibility and cost, this approach is seldom used. Therefore, development of a non-invasive method to evaluate the hemodynamic impact of FMD lesions is urgently needed. Design and method: We developed a computer application to quantify FMD-related stenoses and estimate the resulting pressure gradient based on angiography images. Based on analysis of images of 55 patients from the FEIRI database, the application allows 1) binary segmentation (vessel vs. background) of the image, 2) identification of vessel centreline through skeletonisation, 3) tracking of vessel diameter, and 4) quantitative characterisation of beads and stenoses. We used empirical fluid dynamics laws and the obtained vessel geometry to estimate pressure drops across lesions, also accounting for turbulences and vessel tortuosity. Retrospective image analysis and subsequent gradient estimation were performed blinded for clinical data and in vivo pressure gradients. Calculated pressure gradients were validated against those measured in vivo through invasive catheterisation in 9 patients from the ARCADIA-POL study. Results: We analysed angiographic images from FMD-related lesions of n=9 female patients aged 27–59 years, in whom translesional pressure gradients were available. n=18 distinct images were obtained: thirteen pre-angioplasty (four patients with bilateral lesions) and five post-angioplasty. Predicted pressure gradients showed a strong correlation with in vivo pressure gradients (r=0.88, Figure 1). Furthermore, in 11/13 (85%) pre-angioplasty images, the software correctly discriminated between significant (>10 mmHg) vs. non-significant pressure gradients. Our software also reliably predicted the decrease in pressure gradients pre-to-post-angioplasty. Conclusions: We show for the first time the feasibility of pressure gradient prediction based on the anatomy of renal FMD lesions. The next steps will consist in expanding our study population and adapting our approach (currently depending on catheter-based angiography), to non-invasive imaging such as CT- or MR-angiography. Finally, we will correlate estimated pressure gradients with clinical outcomes such as blood pressure response to angioplasty.
peripheral vascular disease
What problem does this paper attempt to address?