ABC | Volume 112, Nº4, April 2019

Original Article Ciuffo et al LA Remodeling and Dyssynchrony Arq Bras Cardiol. 2019; 112(4):441-450 in those with paroxysmal AF. To test these hypotheses, we conducted a cross-sectional study to evaluate LA intra-atrial dyssynchrony and LA-LGE in patients with either paroxysmal or persistent AF. We also quantified the amount of time required for the postprocessing, and the inter-reader and intra-reader reproducibility of LA intra-atrial dyssynchrony. Methods Study population The study included 146 consecutive patients with symptomatic, drug-refractory AF referred for catheter ablation at the Johns Hopkins Hospital between June 2010 and December 2015 who underwent pre-procedural cardiac magnetic resonance (CMR). Patients with prior AF ablation or surgical procedure in the LA were excluded. Based on Heart Rhythm Society most recent guidelines, paroxysmal AF was defined as AF that terminates spontaneously or with intervention within 7 days of onset. Persistent AF is defined as continuous AF that is sustained beyond 7 days. 7 Patients in AF at the time of CMR were also excluded. The protocol was approved by the Institutional Review Board of The Johns Hopkins Hospital, and all the patients provided written informed consent. CMR protocol CMR was performed with a 1.5-Tesla scanner (Avanto; Siemens Medical Systems, Erlangen, Germany), a 6-channel phased-array body coil in combination with a 6-channel spine matrix coil. Electrocardiogram (ECG)-gated, breath‑holding cine CMR images were acquired in the long-axis two- and four‑chamber views by true fast imaging with steady‑state precession (TrueFISP) sequence with the following parameters: TE/TR3.0/1.5ms; flip angle 78°; in-plane pixel size 1.5×1.5mm 2 ; slice thickness 8 mm; slice spacing 2 mm; 30 frames per ECG R-R interval with a temporal resolution of 20-40 ms. The patients also underwent respiratory-navigated, ECG-gated LGE for quantification of LA fibrosis (Figure 3). LGE images were acquired within 15‑25 minutes following the injection of gadopentetate dimeglumine (0.2 mmol/kg; Bayer Healthcare Pharmaceuticals, Montville, NJ, USA) using a fat-saturated 3D inversion recovery‑prepared fast spoiled gradient-recalled echo sequence with the following parameters: TE/TR 1.52/3.8 ms; flip angle 10°; in-plane pixel size 1.3×1.3 mm 2 ; slice thickness 2.0mm. The trigger time for three‑dimensional (3D) LGE images was optimized to acquire imaging data during LA diastole as determined by the cine CMR images. The optimal inversion time was determined by an inversion time scout scan (median 270 ms, range 240‑290 ms) to maximize nulling of the LA myocardium. The image intensity ratio (IIR) 11 was measured to quantify LA‑LGE using QMass MR (version 7.2; Medis Medical Imaging Systems bv, Leiden, the Netherlands) on axial images from 3D axial image data. Briefly, IIR is a signal intensity of LA‑LGE normalized by the mean signal intensity of the LA blood pool. The IIR threshold of 1.22 that corresponds to bipolar voltage 0.3 mV on intracardiac electrogram was used to define myocardial fibrosis. 12,13 Preprocedural CMR scans were acquired within a range of 15–25 minutes (mean 18.8 ± 2.4 minutes). Magnetic resonance imaging Analysis Left atrial intra-atrial dyssynchrony Multimodality Tissue Tracking software (MTT, version 6.1, Toshiba, Japan) was used to quantify the LA longitudinal strains and strain rates in two-chamber and four-chamber views. The accuracy and reproducibility of MTT have been validated previously. 9,14 Briefly, an experienced operator, blinded to the type of AF, defined the LA endocardial and epicardial borders at the LA end diastole (Figure 1). The confluence of the pulmonary veins and LA appendage were excluded as appropriate. The software automatically propagates endocardial/epicardial borders over the entire cardiac cycle using a template matching algorithm. 14 Finally, the operator verified the quality of the tracking generated by MTT. The software automatically divides the LA into six equal-length segments in each of the two- and four-chamber views, creating a total of 12 segments (Figure 1). Longitudinal strain and strain rate were calculated within each of the 12 segments (Figure 2). Based on those curves, we defined five indices of LA intra-atrial dyssynchrony as follows: 15,16 • SD-time to peak strain (SD-TPS): Standard deviation of the time to peak longitudinal strain in 12 segments. This index quantifies intra-atrial dyssynchrony of the LA reservoir function. • SD-time to peak pre-atrial contraction strain (SD‑TPS preA ): Standard deviation of the time to the peak pre-atrial contraction strain in 12 segments. This index quantifies intra-atrial dyssynchrony of the LA reservoir and conduit function. A higher value of each index reflects a greater degree of intra-atrial dyssynchrony. We also presente the values of LA dyssynchrony as percentage (SD, %) of R-R’ interval. A similar assessment of LA dyssynchrony has been published and validated before using 3D echocardiography against standard two-dimensional (2D) echocardiography, in a population of individuals with paroxysmal and persistent AF against healthy subjects. 17,18 Out of a total of 1,752 segments, 34 (1.94%) were excluded from analysis because these segments lacked well-defined peaks in the strain/strain rate curves. A total of 22 subjects had at least one segment that was not analyzable, of whom 15 were in the persistent AF group and 7 were in the paroxysmal AF group (p = 0.02). LA function LA functional analysis was described previously. 16 The LA longitudinal strain and strain rate were calculated by averaging strain values in all 12 segments obtained in long-axis two‑ and four-chamber views. A positive and negative strain value indicates stretch and shortening, respectively, with respect to the reference configuration at the ventricular end diastole, defined as the peak of R wave on surface ECG. Maximum LA longitudinal strain (S max ) and pre-atrial contraction strain (S preA ) were identified from the strain curve (Figure 2); the strain rates in left ventricular (LV) systole (SR s ), LV early diastole (SR e ), and LA contraction (SR a ) were obtained from the strain rate curve. The LA volume curve was generated by the biplane modified 442