MRI of the left atrium at 3T: evaluation of measurement reproducibility in healthy volunteers and patients with cardiovascular disease

Abstract

Background

Left ventricular (LV) function has traditionally been the focus for cardiac magnetic resonance imaging (MRI) investigations, but similar methods can also be applied to the left atrium (LA). Previous studies elsewhere have almost entirely involved the use of 1.5T systems, but 3T MRI can provide faster data acquisition with thinner image slices, and may be more suitable for quantifying the structure and function of the LA.

Purpose

To evaluate 3T-MRI for LA volume assessments in: (i) healthy volunteers (HV); (ii) patients with LV-hypertrophy and ischemia (LVHI); and (iii) patients with LV-hypertrophy and diabetes (LVHD).

Material and Methods

Participants were imaged using a balanced steady-state free precession sequence. Healthy volunteers were scanned twice and patients were scanned on one occasion. Volumes were segmented by two observers, and coefficients of repeatability (CoR) were derived.

Results

For LA volumes (indexed to body surface area), CoRs were in the range of 1.3–4.6 mL/m². The LVHI patients had enlarged LA volumes (diastolic, 46.4 mL/m²; systolic, 25.9 mL/m²) and reduced ejection fraction (EF) (44.9%) relative to the HV (diastolic, 39.0 mL/m²; systolic, 17.8 mL/m²; EF, 54.5%) and LVHD groups (diastolic, 41.4 mL/m²; systolic, 20.2 mL/m²; EF, 50.7%). LA volumes were moderately correlated with LV mass in the HV group (R²= 0.59 for LA end-systolic volume), but became weaker (R²≤ 0.17) for patient groups.

Conclusion

3T-MRI derived LA volume measurements are simple and repeatable, and can elicit clear differences between LVHI patients and HVs. These MRI endpoints provide scope for improved radiological interpretation of LA structure and function, and the high degree of repeatability validates their use for longitudinal investigations where precision work is essential.

Keywords

Cardiac magnetic resonance imaging (MRI)heart left atrium segmentation

Introduction

Left atrial (LA) size and function are known to be important prognostic factors for cardiovascular disease (CVD). An enlarged LA may be indicative of pathology such as atrial fibrillation (AF) (1), cerebral ischemic events (2), and various cardiomyopathies (3 –5) leading to increased risk of morbidity and mortality (6).

Magnetic resonance imaging (MRI) offers many advantages for cardiac imaging, including flexible field of view (FOV), excellent blood-tissue contrast, and the ability to apply precise geometric coverage to any imaging plane. The image resolution permits clearer border detection of the LA compared to other techniques such as ultrasound (7), and the use of CINE steady-state free precession (SSFP) sequences has evolved as the sequence of choice since it can provide excellent depiction of the LA (8).

In order to measure LA structure and function, simple assessments of length or area can be acquired, and model-based assumptions may be applied where the volume is calculated by “bi-plane” area-length measures (9) or by ellipsoid geometry (10). These assessments are quick, but shape modeling assumptions can be susceptible to errors since the LA is an asymmetric cavity. Further, LA dilation is unlikely to be evenly distributed in all planes and measurements of simple dimensions may be further insensitive to change for this reason (11).

A technical comparison of selected “volume-based” MRI methods is presented in Table 1 (12 –34). It is apparent that a “standard” MRI method for LA evaluation has yet to be defined, and it can also be seen that there is very little reported use of 3T machines for cardiac magnetic resonance (CMR) of LA structure and function. The use of a 3T system may offer added utility by virtue of the additional signal-to-noise ratio (SNR), which can be utilized to implement faster scanning (using increased parallel imaging) or for clearer delineation of the LA structures over thinner slices.

Table 1.

Previous studies where Simpson’s rule has been used to define LA volume parameters.

Author (year)	Ref	Field (T)	Pulse seq (gating)	Slice/Gap (mm)	Matrix	FOV (cm)	View	LAA/PV	Cohort (n)	LA volume (diastolic) (Abs (mL)/Norm (mL/m²)
Jarvinen (1996)	12	1.0	GRE (P)	10/N	256	34	VLA	I/E	HV (n = 8)	Abs 106 ± 21
	12	1.0	GRE (P)	10/N	256	34	VLA	I/E	HCM (n = 8)	Abs 142 ± 44
	12	1.0	GRE (P)	10/N	256	34	VLA	I/E	DCM (n = 8)	Abs 141 ± 37
Poutanen (2000)	13	1.5	SPGR (Y)	4–8/N	256	40	HLA	E/E	Juveniles (n = 30)	Abs 26.7 ± 7.4, Norm 22.1 ± 5.3
Sievers (2004)	14	1.5	SSFP (P)	7/3	N	N	SA	N	HV (n = 15)	Abs 61.3 ± 19
	14	1.5	SSFP (P)	7/3	N	N	SA	N	AF (n = 18)	Abs 112.7 ± 41.3
Hauser (2004)	15	1.5	SSFP (P)	10/N	256	48	HLA	E/E	AF (n = 18)	Abs 118 ± 39
Raman (2005)	16	1.5	SSFP/fGRE	N/N	N	N	HLA	E/N	HV (n = 15)	Norm 27 ± 9
	16	1.5	SSFP/fGRE	N/N	N	N	HLA	N	Paroxysmal AF (n = 20)	Norm 31 ± 8
	16	1.5	SSFP/fGRE	N/N	N	N	HLA	N	HCCMD (n = 12)	Norm 28 ± 8
Therkelsen (2005)	17	1.5	SPGR (Y)	6/0	256	35	VLA	N/E	HV (n = 19)	Norm 49.7 ± 6.0
	17	1.5	SPGR (Y)	6/0	N	35	VLA	N	Persistent AF (n = 58)	Norm 78.5 ± 18.5
	17	1.5	SPGR (Y)	6/0	N	35	VLA	N	Permanent AF (n = 19)	Norm 77.4 ± 19.1
Sievers (2005)	18	1.5	SSFP (P)	6/4	256	38	SA	N	HV (n = 18)	Abs 66.4 ± 20.1
	18	1.5	SSFP (R)	6/4	192	38	SA	N	HV (n = 18)	Abs 66.8 ± 19.8
Germans (2007)	19	1.5	SSFP (R)	5/5	N	N	VLA	I/E	HV (n = 40)	Norm (Old, 50 ± 6.7; Young, 47 ± 9.3)
Hudsmith (2007)	20	Var	SSFP (R)	7/3	N	35	SA	I/E	HV (n = 10)	Abs (1.5T, 81 ± 27; 3.0T, 78 ± 25)
	20	Var	SPGR (R)	7/3	N	35	SA	I/E	HV (n = 10)	Abs (1.5T, 82 ± 26; 3.0T, 77 ± 26)
Marui (2008)	21	1.5	SPGR (P)	8/2	N	N	Axial	N/N	AF (n = 74)	Abs 361 ± 74
Patel (2008)	22	1.5	SSFP (R)	6/N	N	N	Axial	E/E	Patients (n = 30)	Abs 107 ± 36 (range, 51–188)
Nori (2009)	23	1.5	SSFP (R)	6/0	256	40	HLA	N/N	Paroxysmal AF (n = 16)	Norm 37.0 ± 6.4
	23	1.5	SSFP (R)	6/0	256	40	HLA	N/N	Permanent AF (n = 13)	Norm 41.4 ± 8,7
	23	1.5	SSFP (R)	6/0	256	40	HLA	N/N	HV (n = 12)	Norm 28.8 ± 6.1
Bastarrika (2010)	24	1.5	SSFP (R)	5/0	192	32–38	Axial	I/E	HTP (n = 29)	Abs 158.2 ± 72.5
Maciera (2010)	25	1.5	SSFP (R)	5/0	N	N	SA	I/E	HV - (n = 60 men, 60 women)	Abs 73.0 ± 14.9, Norm 40.0 ± 6.7
Wen (2010)	26	3.0	SSFP (R)	5/N	224	35	VLA	E/E	Patients (n = 49)	Norm 46.2 ± 9.7
Sarikouch (2011)	27	1.5	SSFP (Y)	5-6/N	168	N	Axial	I/E	HC & A (n = 115)	Abs 64.7 ± 26.1
von Bary (2012)	28	1.5	3D SPGR	1.2/0	384	35	Axial	E/E	Paroxysmal AF (n = 213)	Abs 85.0 ± 28
Kuhl (2012)	29	1.5	SSFP (Y)	8/0	192	30	SA	I/E	Post STEMI (n = 54)	Norm 47.4 ± 9.8
Ahtarovski (2012)	30	1.5	SSFP (R)	7/0	192	30–36	SA	I/E	HV (n = 40)	Norm (Old, 47 ± 7; Young, 44 ± 9)
Nacif (2013)	31	1.5	SSFP (R)	8/2	192	35–42	SA	N/E	CMR patients (n = 81)	Abs 78.6 ± 31.1
Fredersdorf (2014)	32	1.5	3D SPGR	1.2/0	384	35	Axial	E/E	AF (n = 60)	Abs 89.3 ± 3.9
Mosen (2014)	33	1.5	SSFP (R)	8/0	N	N	SA	I/E	HV - M (n = 32)	Abs 93 ± 19; Norm 47 ± 8
	33	1.5	SSFP (R)	8/0	N	N	SA	I/E	Athletes - M (n = 42)	Abs 116 ± 19; 56 ± 8
	33	1.5	SSFP (R)	8/0	N	N	SA	I/E	HV - F (n = 21)	Abs 83 ± 17; Norm 46 ± 8
	33	1.5	SSFP (R)	8/0	N	N	SA	I/E	Athletes - F (n = 33)	Abs 90 ± 15; Norm 51 ± 7
Robertson (2014)	34	1.5	SSFP (R)	8/0	256	32	SA	I/E	HV (n = 20)	CC Abs 47 ± 15; CC Norm 24 ± 6
	34	1.5	SSFP (R)	8/0	256	32	SA	I/E	AF (n = 15)	CC Abs 40 ± 6; CC Norm 20 ± 3

Abs, absolute value; AF, atrial fibrillation; CC, cyclic volume change; DCM, dilated cardiomyopathy; E, excluded; F, female; fGRE, fast gradient echo; FOV, field of view; GRE, gradient echo; HC & A, healthy children and adolescents; HCCMD, heritable cardiac conduction and myocardial disease; HCM, hypertrophic cardiomyopathy; HLA, horizontal long axis; HTP, heart transplant patient; HV, healthy volunteers; I, included; LAA, left atrial appendage; M, male; N, data not available; Norm, normalized value; Old, older group; P, prospective gating; Pulse seq, pulse sequence; PV, pulmonary vein; R, retrospective gating; Ref, reference as numbered in manuscript; SA, short axis; SPGR, spoiled gradient echo; SSFP, steady-state free precession; STEMI, ST segment elevation myocardial infarction; T, Tesla; Var, variable field strength; VLA, vertical long axis; Y, yes; Young, younger group.

The primary aim of this pilot study was to implement a suitable 3T MRI protocol for optimized scanning of LA structure and function. For the technique to be considered useful, the resulting LA assessments should be: (i) sensitive to variation between different clinical groups; (ii) repeatable over successive scans within a short timescale; and (iii) consistent whether single or multiple observers undertake the postprocessing. Therefore, the specific objectives were to address these research questions by scanning a group of healthy volunteers at two time points over the course of a single day (“AM” and “PM”), and two groups of patients with CVD on a single occasion. The former would enable “single time-point”, “scan-to-scan”, “intra-observer”, and “inter-observer” measurement repeatability to be examined, while the latter would enable LA structural and functional differences between the study groups to be identified.

Material and Methods

The following three fully-informed and consenting study groups were recruited: 1, healthy volunteer (HV) group (8 men, 7 women; age range, 24–39 years; mean age, 31 years), with no prior history of cardiovascular disease and no contraindication to MRI. This group was included to derive a young “baseline” range for LA volumes and to provide a “hardest case” (smallest LA volumes) test for repeatability testing of the analysis software; 2, patients with LV-hypertrophy and ischemia (LVHI) group (13 men, 6 women; age range, 50–79 years; mean age, 66 years), with known history of ischemic heart disease or angina, LVH by M-Mode echocardiography as defined by the American Society of Echocardiography (115 g/m² for men, 95 g/m² for women), blood pressure (BP) <150/90, and no contraindication to MRI; 3, patients with LV-hypertrophy and diabetes (LVHD) group (11 men, 8 women; age range, 40–81 years; mean age, 64 years), with clinical type 2 diabetes, LVH (as defined above), BP <150/90, and no contraindication to MRI.

Healthy volunteers were scanned at two time-points (“AM” and “PM” on the same day) and patient volunteers were scanned at a single time-point. MRI was performed on a 3T Magnetom Trio (Siemens, Erlangen, Germany) using body array and spine matrix coils.

For LA assessment, a two-dimensional (2D) ECG-gated segmented CINE True Fast Imaging with Steady-State Free Precession (TrueFISP) sequence with retrospective gating was used to acquire a stack of “2ch” images (typically 10–14 slices) from the lateral side of the LA to the atrial septum (Fig. 1). Two slices (each 5 mm) were acquired per breath-hold at end expiration. Other imaging parameters were: TR/TE, 3.4/1.5 ms; flip angle (FA), 50°; parallel imaging (GRAPPA) factor, 3; and in-plane resolution, 173 × 256 over a targeted FOV of 360–420 mm. The same sequence was also used for LV assessment, where a short axis (SA) stack of data was acquired from the atrio-ventricular ring to the apex. The imaging parameters were: TR, 3.4 ms; TE, 1.5 ms; and FA, 50°. Two slices per breath-hold were obtained, with a parallel imaging (GRAPPA) factor of 2 and slice thickness of 6 mm (4 mm gap). The in-plane resolution was 128 × 192 pixels, with a FOV in the range of 340–450 mm.

Fig. 1.

Image acquisition and segmentation of quantitative left atrial parameters.

Segmentation of the LA (Fig. 2) for all participants was performed twice by an MRI physicist observer (with 10 years of CMR experience) using commercial software (“Argus”, Multi-Modality Work Platform, version VB17, Siemens). Contours were defined at atrial diastole and systole by measuring the blood volume contained by the atrial wall and the mitral valve. The LA appendage was included, and pulmonary vein structure was excluded wherever possible. Repeated segmentations were completed with at least 1 month between each analysis to ensure that “learning bias” was minimized. Additionally, a second MRI physicist observer (with 7 years of CMR experience) performed the segmentation once for every participant in order to establish inter-observer variability. The LA ejection fraction (LA-EF), end-diastolic volume (LA-EDV), end-systolic volume (LA-ESV), and stroke volume (LA-SV) were derived and indexed to body surface area (BSA) (35). Segmentation of the LV SA data was also undertaken in order to derive LV ejection fraction (LV-EF), end-diastolic volume (LV-EDV), end-systolic volume (LV-ESV), stroke volume (LV-SV), and LV mass. The LV endocardial and epicardial borders were defined on images at end-diastole (ED) and end-systole (ES) by each observer. Image slices were included if more than 50% full-thickness myocardium was visible, and papillary muscles were included within the LV mass if they were visibly indistinguishable from the endocardial wall.

Statistical testing for normality of all data was examined using the Shapiro-Wilks test (SPSS, IBM Corporation, Armonk, NY, USA) and analysis was performed using one-way analysis of variance (ANOVA) to test the null hypothesis that there were no structural or functional differences in the mean LA parameters between any of the groups. Intra- and inter-observer repeatability of LA measurements was examined by use of Bland–Altman statistics and described by the coefficient of repeatability (CoR). Correlations between LA volumes and the LV parameters were also investigated.

Results

All MR data were acquired successfully except for one healthy volunteer who “moved” during the PM LA scan. This dataset was removed from the subsequent analysis phase.

An example of a LA (with contour placement) in the 2ch plane is shown in Fig. 2. Visual inspection revealed that the LA volumes in the LVHI group appeared larger than for the other groups. An example is shown (Fig. 2), where images at LA end-diastole and LA end-systole are illustrated side-by-side for LVHI and HV participants.

Fig. 2.

Example left atrial segmentation in (a) a healthy volunteer (HV) and (b) a patient volunteer with LVH and angina (LVHI).

Quantitative LA parameters are presented in Table 2, along with intra- and inter-observer coefficient of repeatability (CoR) measurements. For volumetric data the CoRs were in the range of 1.3–4.6 mL/m², and for functional (LA-EF) data the CoRs were in the range of 3.1–5.4%. The volume repeatability was typically 2–3 mL/m², but slightly higher when AM–PM “scan-to-scan” data were assessed (up to 4.6 mL/m²), or when an inter-observer was used to perform one set of analyses (up to 3.7 mL/m²).

Table 2.

Left atrial volume data – for healthy volunteers (HV – at AM and PM), patient volunteers with left ventricular hypertrophy and ischemia (LVHI), and patient volunteers with left ventricular hypertrophy and diabetes (LVHD). Data were acquired by observers 1 and 2.

Cohort	Observer	Scan number (time-point)	LAEF (%)	LAEDV (mL/m²)	LAESV (mL/m²)	LASV (mL/m²)
Left atrial volumetry/function parameters (normalized to body surface area)
HV	1	1 (AM)	54.4 ± 4.7	39.5 ± 7.3	18.1 ± 4.4	21.4 ± 3.8
	1	1 (AM)	54.5 ± 4.6	39.0 ± 7.0	17.8 ± 4.2	21.2 ± 3.7
	1	2 (PM)	54.7 ± 3.7	39.4 ± 6.8	17.9 ± 3.9	21.5 ± 3.6
	2	1 (AM)	54.9 ± 5.2	40.4 ± 7.4	18.3 ± 4.4	22.1 ± 4.1
LVHI	1	1 (AM)	45.4 ± 8.1	45.8 ± 11.3	25.3 ± 7.8	20.5 ± 5.7
	1	1 (AM)	44.9 ± 7.6	46.4 ± 11.6	25.9 ± 8.3	20.5 ± 5.7
	2	1 (AM)	44.1 ± 8.4	46.8 ± 11.2	26.5 ± 8.2	20.3 ± 5.7
LVHD	1	1 (AM)	51.2 ± 3.7	41.1 ± 12.9	20.0 ± 6.4	21.1 ± 7.0
	1	1 (AM)	50.7 ± 4.1	41.4 ± 13.1	20.2 ± 6.3	21.2 ± 7.4
	2	1 (AM)	49.6 ± 4.2	41.1 ± 13.4	20.6 ± 6.7	20.5 ± 7.2
Coefficient of repeatability (CoR) measurements
HV	Intra	1 (AM) vs. 1 (AM)	3.5	3.2	2.1	2.1
	Intra	1 (AM) vs. 2 (PM)	4.7	4.6	2.2	3.6
	Inter	1 (AM) vs. 1 (AM)	5.0	3.4	2.1	3.0
LVHI	Intra	1 (AM) vs. 1 (AM)	4.0	2.7	2.8	1.3
	Inter	1 (AM) vs. 1 (AM)	4.5	3.7	3.6	2.6
LVHD	Intra	1 (AM) vs. 1 (AM)	3.1	2.9	1.7	2.0
	Inter	1 (AM) vs. 1 (AM)	5.4	3.3	2.5	3.0

Significant differences were identified between the LVHI and HV groups for mean LA-EDV (46.4 mL/m² vs. 39.0 mL/m², P < 0.05), LA-ESV (25.9 mL/m² vs. 17.8 mL/m², P < 0.05), and LA-EF (44.9% vs. 54.5%, P < 0.05). Similar changes were also identified between the LVHI group and the LVHD group for mean LA-ESV (25.9 mL/m² vs. 20.2 mL/m², P < 0.05) and LA-EF (44.9% vs. 50.7%, P < 0.05). The mean LA-EF was also slightly lower in the LVHD group relative to the HVs (50.7% vs. 54.5%, P < 0.05) but otherwise there were no significant differences measured between these two groups. The mean LA-SV remained stable across all study groups (21.2 +/− 3.7 mL/m² for HV, 20.5 +/− 5.7 mL/m² for LVHI and 21.2 +/− 7.4 mL/m² for LVHD), suggesting that blood volume delivery into the LV remains preserved – despite the structural differences identified.

Corresponding LV parameters have been measured but are not described in detail in this report since the focus of the work was to examine the LA. However, the relationship between LA size (LA-EDV and LA-ESV) and LVM was assessed (Fig. 3). The mean LVM values were found to be significantly raised in the LVHI group (68.5 g/m²) relative to the HV group (59.5 g/m², P < 0.05), and approaching significance relative to the LVHD group (62.0 g/m², P < 0.08), i.e. a similar pattern of change to that observed for the mean LA-EDV and LA-ESV measurements. For individual measures, the best correlation was found between LA-ESV and LVM for the HV group: R²= 0.59; but equivalent correlations were reduced in both patient groups (R²= 0.02 for LVHI, and R²= 0.12 for LVHD). A similar pattern was evident between LA-EDV and LVM, with R²= 0.31 for HV, R²= 0.12 for LVHI, and R²= 0.17 for LVHD. Better correlations were achieved between LV and LA measures for stroke volume (R²= 0.40 for HV; 0.72 for LVHI; and 0.78 for LVHD), along with maximum chamber volumes (R²= 0.67 for HV; 0.47 for LVHI; and 0.74 for LVHD). However, correlations were weaker when minimum chamber volumes were compared (R²= 0.24 for HV; <0.01 for LVHI; and 0.46 for LVHD).

Fig. 3.

Relationship between mean LA-EDV, LA-ESV, and LVM values for the three groups studied. Significant changes (increases) were noted in the LVHI group for mean LA-EDV, mean LA-ESV, and mean LVM relative to the HV population.

Discussion

In this study, we used a 3T MRI scanner to derive repeatable measures of LA structure and function when: (i) analyzed by single or multiple observers; (ii) acquired at single or multiple time-points; or (iii) derived from different study populations. With relatively easy LA boundary detection resulting in typical CoRs of 2–3 mL/m², the measurements are stable, even when different time-points or observers are involved.

Differences in these LA parameters between the clinical and healthy volunteer groups have also been observed. Although of interest, clinical interpretation should be made with caution since the precise inclusion criteria for each group were different (e.g. age, gender, blood pressure matching). However, Maciera et al. (25) have described no clear association of LA volume with age and gender (when normalized to BSA) which is reassuring since it suggests that our LA volume differences could be reflective of cardiovascular change. A limitation of the work is that the size of each group (study “n”) is relatively small (although still larger than previously reported “sample size” calculations for CMR studies (36)). It would be interesting to examine further the effect of age and gender on much larger “population-based” cohorts as part of future work.

Previous investigations of the LA have almost exclusively utilized 1.5T machines, and we can only find two cases where 3T scanners have been described (20,26). Wen et al. (26) used 3T MRI for evaluating the LA in a cohort of patients with clinical indications for coronary CT. Their mean maximum LA volume by MRI was 46.2 +/− 9.7 mL/m², which is similar to our LVHI group (46.4 +/− 11.6 mL/m²). Similarly, Hudsmith et al. (20) reported maximum LA volumes of 78 +/− 25 mL for a healthy volunteer group (n = 10 men) using 3T MRI with an SSFP sequence, which if corrected for body surface area (approximately 2 m²) results in similar data to our HV group (39.0 +/− 7.0 mL/m²). The above consistencies also suggest that similar data can be acquired on different scanners at different sites and with different vendors. This is encouraging from the perspective of larger multi-center studies, where consistency across different machines may be beneficial. A limitation of our work is that a direct comparison with 1.5T data acquired on the same participants was not undertaken. However, a report by Hudsmith et al. (20) stated that “field strength had no effect on LA volume and function assessment using either FLASH or SSFP,” suggesting that 1.5T and 3.0T can be used interchangeably.

Previous measurements of the LA have been acquired in the 2ch, 4ch, SA, or axial plane, with the SA orientation (“extension” of an LV assessment) being the most common. However, it is difficult to identify the LA at the mitral valve in this plane, so we used the 2ch orientation (clearer view of the mitral valve plane) as recommended initially by Jarvinen et al. (12). Our choice of slice coverage every 5 mm was generally successful, providing clear LA boundaries and no significant partial volume difficulties. However, it does require the volunteer/patient to remain absolutely still to ensure that each image slice corresponds to the correct anatomy. A complete LA acquisition took typically 5 min to complete. The use of 3T was chosen because of the potential to utilize the “additional signal” to acquire data faster (i.e. more “slices per breath-hold”) using more parallel imaging and thinner image slices. However, some extra time was required for patient-specific targeted shimming in order to minimize resonance-offset and flow-related artifacts. For data analysis, the time taken for completion of a typical LA segmentation was about 15 min. We acknowledge that quicker methods are available, such as those that use the biplane area-length method for deriving LA volume. However, these methods rely upon geometric assumptions which may be less valid for cohorts with CVD where atrial remodeling has taken place.

In conclusion, this pilot study has demonstrated that 3T MRI-derived measurements of LA structure and function are reproducible and able to distinguish differences between the healthy individuals and patient groups. These LA endpoints are anticipated to provide a useful aid for future clinical interpretation of LA structure and function, and may also improve longitudinal investigations where consistency between “baseline” and “follow-up”’ examinations is paramount.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We would like to acknowledge the financial support of: (i) the Medical Research Council (MRC) for the LVHI study; and (ii) Diabetes UK for the LVHD study.

Acknowledgments

The authors would like to acknowledge: (i) the assistance of the MRI radiographers who have participated in the study; (ii) Mrs Norma Gourlay for her assistance with patient and volunteer bookings, and (iii) the patients and volunteers who kindly gave their time to be scanned.

References

Tsang

TSM

Barnes

Bailey

. Left atrial volume: important risk marker of incident atrial fibrillation in 1655 older men and women. Mayo Clin Proc 2001; 76: 467–475.

Benjamin

Agostino

Belanger

. Left atrial size and risk of stroke and death. The Framingham heart study. Circulation 1995; 92: 835–841.

Papavassiliu

Germans

Flücher

. CMR findings in patients with hypertrophic cardiomyopathy and atrial fibrillation. J. Cardiovasc Magn Reson 2009; 11: 34–34.

Gulati

Ismail

Jabbour

. Clinical utility and prognostic value of left atrial volume assessment by cardiovascular magnetic resonance in non-ischemic dilated cardiomyopathy. Eur J Heart Fail 2013; 15: 660–670.

Maron

Haas

Maron

. Left atrial remodelling in hypertrophic cardiomyopathy and susceptibility markers for atrial fibrillation identified by cardiovascular magnetic resonance. Am J Cardiol 2014; 113: 1394–1400.

Gupta

Matulevicius

Ayers

. Left atrial structure and function and clinical outcomes in the general population. Eur Heart J 2013; 34: 278–285.

Artang

Migrino

Harmann

. Left atrial volume measurement with automated border detection by 3-dimensional echocardiography: comparison with magnetic resonance imaging. Cardiovasc Ultrasound 2009; 7: 16–16.

Hudsmith

Petersen

Francis

. Normal human left and right ventricular and left atrial dimensions using steady state free precession magnetic resonance imaging. J Cardiovasc Magn Reson 2005; 7: 775–782.

Muellerleile

Groth

Saring

. Evaluation of different magnetic resonance imaging techniques for the assessment of left atrial emptying. Eur Radiol 2012; 22: 1904–1911.

10.

Bank

Wickstrom

Handran

. Modeling left atrial volume, shape and contraction patterns in normal subjects by cardiac magnetic resonance imaging. Int J Cardiol 2013; 168: 3348–3350.

11.

Tsang

Abhayaratna

Barnes

. Prediction of cardiovascular outcomes with left atrial size: is volume superior to area or diameter? J Am Coll Cardiol 2006; 47: 1018–1023.

12.

Jarvinen

Kupari

Poutanen

V-P

. Right and left atrial phasic volumetric function in midly symptomatic dilated and hypertrophic cardiomyopathy: CINE MR imaging assessment. Radiology 1996; 198: 487–495.

13.

Poutanen

Ikonen

Vainio

. Left atrial volume assessed by transthoracic three-dimensional echocardiography and magnetic resonance imaging: dynamic changes during the heart cycle in children. Heart 2000; 83: 537–542.

14.

Sievers

Kirchberg

Addo

. Assessment of left atrial volumes in sinus rhythm and atrial fibrillation using the biplane area-length method and cardiovascular magnetic resonance imaging with TrueFISP. J Cardiovasc Magn Reson 2004; 6: 855–863.

15.

Hauser

McClennen

Katsimaglis

. Assessment of left atrial volume by contrast enhanced magnetic resonance angiography. J Cardiovasc Magn Reson 2004; 6: 491–497.

16.

Raman

Neff

. Volume cine CMR to quantify atrial structure and function in patients with atrial dysrhythmias. J Cardiovasc Magn Reson 2005; 7: 539–543.

17.

Therleksen

Groenning

Svendsen

. Atrial and ventricular volume and function in persistent and permanent atrial fibrillation, a magnetic resonance imaging study. J Cardiovasc Magn Reson 2005; 7: 465–473.

18.

Sievers

Addo

Kirchberg

. How much are atrial volumes and ejection fractions assessed by cardiac magnetic resonance imaging influenced by the ECG gating method? J Cardiovasc Magn Reson 2005; 7: 587–593.

19.

Germans

Gotte

MJW

Nijveldt

. Effects on aging on Left atrioventricular coupling and left ventricular filling assessed using cardiac magnetic resonance imaging in healthy subjects. Am J Cardiol 2007; 100: 122–127.

20.

Hudsmith

Cheng

ASH

Tyler

. Assessment of left atrial volumes at 1.5 Tesla and 3 Tesla using FLASH and SSFP cine imaging. J Cardiovasc Magn Reson 2007; 9: 673–679.

21.

Marui

Saji

Nishina

. Impact of left atrial volume reduction concomitant with atrial fibrillation surgery on left atrial geometry and mechanical function. J Thorac Cardiovasc Surg 2008; 135: 1297–1305.

22.

Patel

Fatemi

Norton

. Cardiac cycle dependent left atrial dynamics: implications for catheter ablation of atrial fibrillation. Heart Rhythm 2008; 5: 787–793.

23.

Nori

Raff

Gupta

. Cardiac magnetic resonance imaging assessment of regional and global left atrium function before and after catheter ablation for atrial fibrillation. J Interv Card Electrophysiol 2009; 26: 109–117.

24.

Batistarrika

Zudaire

Ferreira

. Assessment of left atrial volumes and function in orthotopic heart transplant recipients by dual-source CT. Invest Radiol 2010; 45: 72–76.

25.

Maciera

Cosin-Sales

Roughton

. Reference atrial dimensions and volume by steady state free precession cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2010; 12: 65–75.

26.

Wen

Zhang

. Assessing the left atrial phasic volume and function with dual-source CT: comparison with 3T MRI. Int J Cardiovasc Imaging 2010; 26: 83–92.

27.

Sarikouch

Koerperich

Boethig

. Reference values for atrial size and function in children and young adults by cardiac MR: a study of the German competence network congenital heart defects. J Magn Reson Imaging 2011; 33: 1028–1039.

28.

von Bary

Dornia

Eissnert

. Predictive value of left atrial volume measured by non-invasive cardiac imaging in the treatment of paroxysmal atrial fibrillation. J Interv Card Electrophysiol 2012; 34: 181–188.

29.

Kuhl

Lonborg

Fuchs

. Assessment of left atrial volume and function: a comparative study between electrocardiography, magnetic resonance imaging and multi slice computed tomography. Int J Cardiovasc Imaging 2012; 28: 1061–1071.

30.

Ahtarovski

Iversen

Lonborg

. Left atrial and ventricular function during dobutamine and glycopyrrolate stress in healthy young and elderly as evaluated by cardiac magnetic resonance. Am J Physiol Heart Circ Physiol 2012; 303: H1469–1473.

31.

Nacif

Barranhas

Turkbey

. Left atrial volume quantification using cardiac MRI in atrial fibrillation: comparison of Simpson’s method with biplane area-length, ellipse and three-dimensional methods. Diagn Interv Radiol 2013; 19: 213–220.

32.

Fredersdorf

Ucer

Jungbauer

. Lone atrial fibrillation as a positive predictor of left atrial volume reduction following ablation of atrial fibrillation. Europace 2014; 16: 26–32.

33.

Mosen

Steding-Ehrenborg

. Atrial remodelling is less pronounced in female endurance-trained athletes compared with that in male athletes. Scand Cardiovasc J 2014; 48: 20–26.

34.

Robertson

Lee

Voeller

. Quantification of the functional consequences of atrial fibrillation and surgical ablation on the left atrium using cardiac magnetic resonance imaging. Eur J Cardiothorac Surg 2014; 46: 720–728.

35.

Dubois

. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 1916; 17: 863–871.

36.

Bellenger

Davies

Francis

. Reduction in sample size for studies of remodelling in heart failure by the use of cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2000; 2: 271–278.