Florent Le Ven1, Helena Tizón-Marcos1, Christina Fuchs1, Patrick Mathieu1, Philippe Pibarot1, Eric Larose2. 1. Département Multidisciplinaire De Cardiologie, Institut Universitaire de Cardiologie et de pneumologie de Québec, and Faculté de Médecine de l'Université Laval, Québec, Québec, Canada. 2. Département Multidisciplinaire De Cardiologie, Institut Universitaire de Cardiologie et de pneumologie de Québec, and Faculté de Médecine de l'Université Laval, Québec, Québec, Canada. Electronic address: eric.larose@criucpq.ulaval.ca.
Abstract
BACKGROUND: Calcific aortic valve disease affects 10%-15% of the elderly population, causing considerable morbidity and mortality. There is no imaging technique that allows for the assessment of tissue composition of the valve in vivo. We thus investigated whether multiparametric magnetic resonance imaging (MRI) could characterize and quantify lipid, fibrous, and mineralized tissues within aortic valve (AV) cusps. METHODS: AV leaflets were explanted from patients with severe aortic stenosis at the time of valve replacement surgery. Aortic cusps were imaged ex vivo using 1.5 T MRI using 3 gradient-echo sequences with T1, moderate T2, and proton density weightings (T1w, T2w, and PDw). Histopathologic analysis was performed on coregistered slices to identify and measure mineralized tissue, fibrous tissue, and lipid-rich tissue. Area and mean grey values were measured in all 3 weightings by standardized software. RESULTS: Four hundred ninety-two regions of interest from 30 AV leaflets were studied. Total leaflet surface and the areas of mineralized (P < 0.0001), fibrous (P = 0.002), and lipid-rich (P = 0.0001) tissues measured by MRI matched closely those measured by histopathologic examination. All 3 weightings provided significant discrimination between median grey values for mineralized, fibrous, and lipid-rich tissues (P < 0.0001 for T1w, moderate T2w, and PDw). A best-fit equation integrating the grey value data from all 3 weightings allowed multiparametric MRI to identify valve leaflet components with areas under the receiver operating characteristic curve of 0.92, 0.81, and 0.72, respectively. CONCLUSIONS: AV leaflet characteristics, including tissue composition, distribution, and area, may be successfully measured by multiparametric MRI with good to excellent accuracy.
BACKGROUND:Calcific aortic valve disease affects 10%-15% of the elderly population, causing considerable morbidity and mortality. There is no imaging technique that allows for the assessment of tissue composition of the valve in vivo. We thus investigated whether multiparametric magnetic resonance imaging (MRI) could characterize and quantify lipid, fibrous, and mineralized tissues within aortic valve (AV) cusps. METHODS: AV leaflets were explanted from patients with severe aortic stenosis at the time of valve replacement surgery. Aortic cusps were imaged ex vivo using 1.5 T MRI using 3 gradient-echo sequences with T1, moderate T2, and proton density weightings (T1w, T2w, and PDw). Histopathologic analysis was performed on coregistered slices to identify and measure mineralized tissue, fibrous tissue, and lipid-rich tissue. Area and mean grey values were measured in all 3 weightings by standardized software. RESULTS: Four hundred ninety-two regions of interest from 30 AV leaflets were studied. Total leaflet surface and the areas of mineralized (P < 0.0001), fibrous (P = 0.002), and lipid-rich (P = 0.0001) tissues measured by MRI matched closely those measured by histopathologic examination. All 3 weightings provided significant discrimination between median grey values for mineralized, fibrous, and lipid-rich tissues (P < 0.0001 for T1w, moderate T2w, and PDw). A best-fit equation integrating the grey value data from all 3 weightings allowed multiparametric MRI to identify valve leaflet components with areas under the receiver operating characteristic curve of 0.92, 0.81, and 0.72, respectively. CONCLUSIONS: AV leaflet characteristics, including tissue composition, distribution, and area, may be successfully measured by multiparametric MRI with good to excellent accuracy.
Authors: Brian R Lindman; Marie-Annick Clavel; Patrick Mathieu; Bernard Iung; Patrizio Lancellotti; Catherine M Otto; Philippe Pibarot Journal: Nat Rev Dis Primers Date: 2016-03-03 Impact factor: 52.329