Septal curvature as a robust and reproducible marker for basal septal hypertrophy

Marciniak, Maciej; Gilbert, Andrew; Loncaric, Filip; Fernandes, João, 1964-; Bijnens, Bart; Sitges, Marta; King, Andrew; Crispi Brillas, Fàtima; Lamata, Pablo

doi:http://dx.doi.org/10.1097/HJH.0000000000002813

Septal curvature as a robust and reproducible marker for basal septal hypertrophy

License

Citation

Marciniak M, Gilbert A, Loncaric F, Fernandes JF, Bijnens B, Sitges M, et al. Septal curvature as a robust and reproducible marker for basal septal hypertrophy. J Hypertens. 2021 Jul;39(7):1421-8. DOI: 10.1097/HJH.0000000000002813

Abstract

Background: Basal septal hypertrophy (BSH) is an asymmetric, localized thickening of the upper interventricular septum and constitutes a marker of an early remodelling in patients with hypertension. This morphological trait has been extensively researched because of its prevalence in hypertension, yet its clinical and prognostic value for individual patients remains undetermined. One of the reasons is the lack of a reliable and reproducible metric to quantify the presence and the extent of BSH. This article proposes the use of the curvature of the left ventricular endocardium as a robust feature for BSH characterization, and as an objective criterion to quantify current subjective ‘visual assessment’ of the presence of sigmoidal septum. The proposed marker, called average septal curvature, is defined as the inverse of the radius adjacent to each point of the endocardial contour along the basal and mid inferoseptal segments of the left ventricle. Method: Robustness and reproducibility were assessed on a cohort of 220 patients, including 161 hypertensive patients (32 with BSH) and 59 healthy controls. Results: The results show that compared with the conventionally used wall thickness metrics, the new marker is more reproducible (relative standard deviation of errors of 7 vs. 13%, and 8 vs. 38% for intra-observer and inter-observer variability, respectively) and better correlates to the functional parameters related to BSH, with main difference (absolute rank correlation 0.417 vs. 0.341) in local deformation changes assessed by longitudinal strain. Conclusion: Average septal curvature is a more precisely defined and reproducible metric than thickness ratios, it can be fully automated, and better infers the functional remodelling related to hypertension.