use of myocardial t1 mapping at 3.0 t to differentiate ... · myocardial t1 mapping in...

This copy is for personal use only. To order printed copies, contact [email protected]

This copy is for personal use only. To order printed copies, contact [email protected] RESEARCH • CARDIAC IMAGING

Anderson-Fabry disease (AFD) is an X-linked disorder of lysosomal metabolism characterized by progressive

accumulation of intralysosomal sphingolipids in multiple organs (1). Cardiac involvement in AFD can result in left ventricular hypertrophy (LVH), myocardial fibrosis, heart failure, and arrhythmia (2,3). Patients with AFD can be treated with enzyme replacement therapy, which should be initiated before irreversible cardiac changes occur (4). There-fore, timely diagnosis and monitoring of AFD are impera-tive. However, differentiation of AFD from other causes of LVH such as hypertrophic cardiomyopathy (HCM) remains a clinical challenge, given overlap in imaging findings (5).

Cardiac magnetic resonance (MR) imaging T1 mapping is useful in characterization of myocardial tissue. Cardiac MR imaging native left ventricular (LV) T1 values at 1.5 T are lower in patients with AFD compared with those in healthy volunteers and in patients with LVH from other causes (6–9). However, to our knowledge, cardiac MR

imaging T1 mapping at 3.0 T has not been previously re-ported in patients with AFD. Furthermore, only one small study has reported reduced right ventricular (RV) native T1 values in AFD compared with pulmonary hypertension (9).

The purpose of our study was to compare LV and RV 3.0-T cardiac MR imaging T1 values in AFD and HCM and to evaluate the diagnostic value of native T1 values beyond age, sex, and conventional imaging features. We hypothesized that native T1 values would be significantly lower in AFD compared with HCM and would provide incremental diagnostic value beyond age, sex, and conven-tional imaging features.

Materials and Methods

Study PopulationOur prospective study was approved by our institutional research and ethics board. Data collection was planned

Use of Myocardial T1 Mapping at 3.0 T to Differentiate Anderson-Fabry Disease from Hypertrophic Cardiomyopathy

Gauri R. Karur, MBBS, MD • Sean Robison, MBBS • Robert M. Iwanochko, MD, FRCPC • Chantal F. Morel, MD, FRCPC, FCCMG • Andrew M. Crean, BM, MRCP • Paaladinesh Thavendiranathan, MD, MS, FRCPC • Elsie T. Nguyen, MD, FRCPC • Shobhit Mathur, MBBS, MD • Syed Wasim, BSc • Kate Hanneman, MD, FRCPC

From the Toronto Joint Department of Medical Imaging, Toronto General Hospital, University of Toronto, 585 University Ave, 1 PMB-298, Toronto, ON, Canada M5G 2N2 (G.R.K., S.R., P.T., E.T.N., S.M., K.H.); Division of Cardiology, Peter Munk Cardiac Centre, University Health Network, University of Toronto, Toronto, ON, Canada (R.M.I., A.M.C., P.T.); and Fred A. Litwin Centre in Genetic Medicine, University Health Network & Mount Sinai Hospital, University of Toronto, Toronto, ON, Canada (C.F.M., S.W.). Received November 8, 2017; revision requested December 12; final revision received January 10, 2018; accepted January 12. Address cor-respondence to K.H. (e-mail: [email protected]).

P.T. supported by Canadian Institutes of Health Research (FRN 147814). K.H. supported by RSNA Research and Education Foundation Research Scholar Grant (RSCH1608) and Academic Innovation Fund (AIF) Grant, Toronto Joint Department of Medical Imaging, University of Toronto.

Conflicts of interest are listed at the end of this article.

Radiology 2018; 288:398–406 • https://doi.org/10.1148/radiol.2018172613 • Content code:

Purpose: To compare left ventricular (LV) and right ventricular (RV) 3.0-T cardiac magnetic resonance (MR) imaging T1 values in Anderson-Fabry disease (AFD) and hypertrophic cardiomyopathy (HCM) and evaluate the diagnostic value of native T1 values beyond age, sex, and conventional imaging features.

Materials and Methods: For this prospective study, 30 patients with gene-positive AFD (37% male; mean age 6 standard deviation, 45.0 years 6 14.1) and 30 patients with HCM (57% male; mean age, 49.3 years 6 13.5) were prospectively recruited between June 2016 and September 2017 to undergo cardiac MR imaging T1 mapping with a modified Look-Locker inversion recovery (MOLLI) acquisition scheme at 3.0 T (repetition time msec/echo time msec, 280/1.12; section thickness, 8 mm). LV and RV T1 values were evaluated. Statistical analysis included independent samples t test, receiver operating characteristic curve analysis, multi-variable logistic regression, and likelihood ratio test.

Results: Septal LV, global LV, and RV native T1 values were significantly lower in AFD compared with those in HCM (1161 msec 6 47 vs 1296 msec 6 55, respectively [P , .001]; 1192 msec 6 52 vs 1268 msec 6 55 [P , .001]; and 1221 msec 6 54 vs 1271 msec 6 37 [P = .001], respectively). A septal LV native T1 cutoff point of 1220 msec or lower distinguished AFD from HCM with sensitivity of 97%, specificity of 93%, and accuracy of 95%. Septal LV native T1 values differentiated AFD from HCM after adjustment for age, sex, and conventional imaging features (odds ratio, 0.94; 95% confidence interval: 0.91, 0.98; P = , .001). In a nested logistic regression model with age, sex, and conventional imaging features, model fit was significantly improved by the ad-dition of septal LV native T1 values (x2 [df = 1] = 33.4; P , .001).

Conclusion: Cardiac MR imaging native T1 values at 3.0 T are significantly lower in patients with AFD compared with those with HCM and provide independent and incremental diagnostic value beyond age, sex, and conventional imaging features.

© RSNA, 2018

Karur et al

Radiology: Volume 288: Number 2—August 2018 n radiology.rsna.org 399

prior to initiation of the study. Between June 2016 and Sep-tember 2017, a convenience sample of patients with AFD and HCM referred for 3.0-T cardiac MR imaging as part of the clinical management protocol were prospectively recruited for T1 mapping (Fig 1). Written informed consent was obtained from all study participants. Exclusion criteria included prior myocardial infarction, known severe coronary artery disease, or contraindications to MR imaging and imaging at 1.5 T. The final cohort consisted of 30 patients with gene-positive AFD with mean age 6 standard deviation of 45.0 years 6 14.1 (11 of 30 [37%] men with mean age of 46.6 years 6 15.5; 19 of 30 [63%] women with mean age of 44.1 years 6 13.5) and 30 patients with HCM with mean age of 49.3 years 6 13.5 (17 of 30 [57%] men with mean age of 51.2 years 6 9.3; 13 of 30 [43%] women with mean age of 46.8 years 6 17.7). The diagnosis of HCM was established based on international guidelines (10). Clinical and demographic data were obtained from the electronic patient record.

Cardiac MR Imaging TechniqueCardiac MR imaging was performed with a 3.0-T imager (Magnetom Skyra Fit; Siemens Health-care, Erlangen, Germany). Retrospectively gated cine steady-state free precession images were ob-tained in multiple planes. Frequency scouting and adjustment were used for steady-state free preces-sion imaging. LV and RV volumes, function, and mass were assessed by using a stack of short-axis cine steady-state free precession sections with cov-erage from cardiac base to the apex (repetition time msec/echo time msec, 40/1.19; field of view, 360 mm; section thickness, 8 mm; intersection gap, 2 mm). T1 mapping was performed with a steady-state free precession readout modified

Look-Locker inversion recovery (MOLLI) acquisition scheme by using recommended MOLLI inversion groupings (before administration of contrast material, 5[3]3; after administration of contrast material, 4[1]3[1]2) (typical imaging parameters: 280/1.12; field of view, 360 mm; section thickness, 8 mm). Three short-axis sections were acquired at the base, midven-tricle, and apex before and 12–15 minutes after administra-tion of 0.15 mmol/kg body weight of gadobutrol (Gadovist; Bayer Healthcare, Berlin, Germany) (11). Multiplane late gadolinium enhancement (LGE) imaging was performed ap-proximately 15 minutes after administration of contrast agent, with a phase-sensitive inversion recovery gradient-recalled echo sequence (781/1.97; field of view, 360 mm; section thickness, 8 mm; intersection gap, 2 mm).

Cardiac MR Imaging PostprocessingAll cardiac MR imaging analyses were performed by a fellow-ship-trained radiologist (G.R.K., with 3 years of cardiovascular imaging experience) blinded to all identifying information in-cluding clinical diagnosis. Quantification of LV and RV vol-umes, function, and mass was performed with commercially available software (QMass, Medis Suite version 3.0.18.6; Me-dis Medical Imaging Systems, Leiden, the Netherlands). LVH was defined as LV mass indexed to body surface area above 85 g/m2 in men and above 81 g/m2 in women (12). RV hypertro-phy was defined as RV mass indexed to body surface area above 29 g/m2 in men and above 28 g/m2 in women (12). Maximum end-diastolic LV and RV wall thicknesses were measured on short-axis cine steady-state free precession images. The pres-ence of accessory apical-basal muscle bundles and myocardial crypts, which are considered characteristic imaging features in HCM, was evaluated on cine steady-state free precession im-ages. An apical-basal muscle bundle was defined as a muscle band extending from LV apex to basal septal or anterior seg-ments without chordal attachment to the mitral valve (13). A

AbbreviationsAFD = Anderson-Fabry disease, AUC = area under receiver operating characteristic curve, CI = confidence interval, ECV = extracellular vol-ume, HCM = hypertrophic cardiomyopathy, LGE = late gadolinium enhancement, LV = left ventricle, LVH = left ventricular hypertrophy, MOLLI = modified Look-Locker inversion recovery, RV = right ventricle

SummaryCardiac MR imaging native T1 values at 3.0 T are significantly lower in patients with Anderson-Fabry disease compared with hypertrophic cardiomyopathy and provide independent and incremental diagnostic value beyond age, sex, and conventional imaging features.

Implications for Patient Care n Cardiac MR imaging native T1 mapping at 3.0 T distinguishes

Anderson-Fabry disease (AFD) from hypertrophic cardiomyopa-thy (HCM) without the need for administration of contrast agent.

n Septal left ventricular native T1 values have independent and in-cremental value over age, sex, and conventional imaging features to differentiate AFD from HCM.

n A septal left ventricular native T1 cutoff point of 1220 msec or lower at 3.0 T distinguishes AFD from HCM with high sensitiv-ity, specificity, and accuracy.

Figure 1: Flowchart shows number of participants with Anderson-Fabry disease (AFD) and hypertrophic cardiomyopathy (HCM) included in study. CAD = coronary artery disease, MI = myocardial infarction.

Myocardial T1 Mapping in Anderson-Fabry Disease and Hypertrophic Cardiomyopathy

400 radiology.rsna.org n Radiology: Volume 288: Number 2—August 2018

was considered to indicate statistical significance. The sample size was calculated to detect a difference in native T1 values between AFD and HCM as low as 45 msec with a standard deviation of 50 msec by means of an independent samples t test (8). For this difference to be detected with a power of 90% and a error of .05, a total of 27 participants per group was needed. All continuous data were first tested for normal distri-bution by using the Shapiro-Wilk test. Continuous variables are described as mean and standard deviation, and categorical variables as numbers and percentages. Comparisons between groups were made with independent samples t test for continu-ous values with normal distribution, Wilcoxon rank sum test for continuous values with nonnormal distribution, and Fisher exact test for categorical values. Correlation between continu-ous variables was assessed with Pearson correlation. Univariable logistic regression was used to evaluate the ability of myocar-dial T1 and ECV values to differentiate AFD from HCM. The linearity assumption of continuous variables was assessed with a quadratic term. Multivariable logistic regression analysis was

myocardial crypt was defined as a fo-cal myocardial defect with a depth of 50% or greater than the myocardium adjacent to it (14). LGE images were assessed for the presence of any LGE, LGE excluding RV insertion points, and LGE at the basal inferior-lateral myocardium. Basal inferior-lateral LGE is considered a characteristic imaging feature in AFD (5).

Inline, nonrigid motion correc-tion of individual T1 mapping im-ages was performed. Offline T1 map-ping analysis was performed with commercially available software (cmr42, version 5.6.3; Circle Cardiovascular Im-aging, Calgary, Canada). For assessment of global LV T1 values, the LV epicar-dium and endocardium were contoured at basal, midventricular, and apical sec-tions, with care taken to avoid blood pool and epicardial fat; T1 values were averaged across the three sections (Figs 2, 3). Septal LV T1 values were evalu-ated by drawing a single region of inter-est on the midventricular section in the midinterventricular septum, avoiding the anterior and inferior RV insertion points, and ensuring a minimum region-of-interest size of 1.5 cm2 (15–17). For assessment of RV T1 values, a single re-gion of interest was drawn in the lateral or inferior RV wall, ensuring a minimum region-of-interest size of 0.15 cm2 (18). If a region of interest could not be reli-ably drawn within the RV myocardium meeting this criterion, then RV T1 val-ues were excluded in that subject (16 of 60 [27%]). Areas with LGE were not excluded from T1 analysis. LV (global and septal) and RV myocardial extracellular volume (ECV) values were calculated with input of native and postcon-trast myocardial and blood pool T1 values and hematocrit as described by Arheden et al (19).

Intraobserver and Interobserver AgreementTo assess intraobserver agreement, a random subset of 20 studies was reanalyzed by the same reader following a minimum 2-week interval after the first analysis; the reader was blinded to the re-sults of the initial assessment and all identifying data. To assess in-terobserver agreement, the same subset was analyzed by a second experienced fellowship-trained reader (K.H., with 5 years of car-diovascular imaging experience) who was blinded to all identify-ing information, including the results of the initial assessment.

Statistical AnalysisStatistical analysis was performed with software (Stata, ver-sion 14.1; Stata, College Station, Tex). Two-tailed P , .05

Figure 2: A, C, Short-axis native T1 maps and, B, D, corresponding late gadolinium enhance-ment (LGE) images in, A, B, 54-year-old man and, C, D, 59-year-old woman with gene-positive Anderson-Fabry disease (AFD). Septal left ventricular (LV) native T1 findings were assessed by drawing region of interest (black arrowhead) in interventricular septum. Global LV native T1 findings were assessed by drawing epicardial (black arrows) and endocardial (white arrows) contours in LV myocardium. Right ventricular (RV) native T1 findings were assessed by drawing region of interest in RV lateral or free wall (white arrowheads). Areas with LGE (yellow arrow) were not excluded from T1 analysis. Septal LV, global LV, and RV native T1 findings are 1113 msec, 1173 msec, and 1234 msec in male subject and 1167 msec, 1263 msec, and 1237 msec in female subject, respectively.

Karur et al


Native T1 and ECV FindingsSeptal LV native T1 values did not differ significantly between patients with AFD who were treated with enzyme replacement therapy and those who were not (1160 msec 6 38 vs 1161 msec 6 51, respectively; P = .92). Among patients with AFD, septal LV native T1 values did not differ significantly between men and women (1150 msec 6 54 vs 1167 msec 6 42, respectively; P = .36) or between patients with LVH and those without LVH (1156 msec 6 26 vs 1162 msec 6 49, respectively; P = .83).

RV native T1 values correlated positively with global (r = 0.480; P = .001) and septal (r = 0.392; P , .001) LV native T1 values, respectively. Among patients with AFD, RV native T1 values were significantly higher compared with global (P = .01) and septal (P , .001) LV native T1 values. Global LV native T1 values were significantly higher compared with septal LV native T1 values (P , .001). Global LV ECV values were significantly higher compared with septal LV ECV values (P = .009).

Global LV, septal LV, and RV native T1 values were all sig-nificantly lower in AFD compared with HCM (P , .001, P ,

performed to evaluate the independent value of septal LV native T1 mapping in differentiation of AFD from HCM in a model with age, sex, and conven-tional imaging features (presence of basal inferolateral LGE, apical-basal muscle bundle, and myocardial crypt). For assessment of incremental diagnos-tic value, the likelihood ratio test was used to compare nested models with only age, sex, and conventional imaging features versus one with septal LV na-tive T1 values added. Receiver operat-ing characteristic curves were created to evaluate the ability of native T1 values to discriminate between groups. Areas under the receiver operating character-istic curve (AUC) for native T1 values were tested for equality (20). The op-timal LV native T1 value threshold for distinguishing AFD from HCM was identified, maximizing accuracy. Intrao-bserver agreement and interobserver agreement were assessed with indi-vidual intraclass correlation coefficient with one-way random-effects models and two-way random-effects models, respectively.

Results

Study PopulationDemographic information is summa-rized in Table 1. Ten patients with AFD (10 of 30 [33%]) were treated with en-zyme replacement therapy at the time of cardiac MR imaging. AFD and HCM groups did not differ significantly with regard to age (P = .24) or sex (P = .20). There was no significant difference in age between men and women in either the AFD group (P = .64) or the HCM group (P = .38).

Cardiac MR Imaging FindingsCardiac MR imaging findings are summarized in Table 2. Maximum LV wall thickness was significantly higher in HCM compared with AFD (P , .001). However, there was no sig-nificant difference in maximum RV wall thickness (P = .16). There was no significant difference in the prevalence of LVH (four of 30 [13%] vs nine of 30 [30%]; P = .21) or right ven-tricular hypertrophy (two of 30 [7%] vs one of 30 [3%]; P . .99) between AFD and HCM, respectively.

There was no significant difference in the presence of any LGE (23 of 30 [77%] vs 28 of 30 [93%]; P = .15) or LGE ex-cluding RV insertion points (17 of 30 [57%] vs 21 of 30 [70%]; P = .42), respectively, between AFD and HCM. However, basal inferolateral LGE was significantly more common in AFD com-pared with HCM (18 of 30 [60%] vs one of 30 [3%]; P = .001).

Figure 3: A, C, Short-axis native T1 maps and, B,D, corresponding late gadolinium enhance-ment (LGE) images in, A, B, 50-year-old man and, C, D, 57-year-old woman with hypertrophic cardiomyopathy (HCM). Septal left ventricular (LV) native T1 findings were assessed by drawing region of interest (black arrowhead) in interventricular septum. Global LV native T1 findings were assessed by drawing epicardial (black arrows) and endocardial (white arrows) contours in LV myocardium. Right ventricular (RV) native T1 findings were assessed by drawing region of interest in RV lateral or free wall (white arrowhead). Areas with LGE (yellow arrows) were not excluded from T1 analysis. Septal LV, global LV, and RV native T1 values are 1318 msec, 1327 msec, and 1313 msec in male subject and 1329 msec, 1298 msec, and 1332 msec in female subject, respectively.



DiscussionDifferentiation of AFD from HCM remains a clinical chal-lenge, given overlap in imaging findings (5). Confirming results of 1.5-T imaging, our findings show that LV and RV native T1 values are significantly lower in patients with AFD compared with HCM at 3.0 T. In addition, we found that septal native LV T1 values have independent and incremental value over age, sex, and conventional imaging features (presence of basal inferolateral LGE, apical-basal muscle bundle, and myocardial crypt) in differentiating AFD from HCM. A septal native LV T1 cutoff point of 1220 msec or lower with use of a MOLLI sequence at 3.0 T distinguishes AFD from HCM with high sensitivity and specificity.

Only a few prior studies have reported LV T1 values in pa-tients with AFD imaged at 1.5- T (6–9). Two prior studies re-ported that native T1 values obtained with shortened MOLLI and saturation recovery single-shot acquisition schemes at 1.5 T were lower in patients with AFD than in healthy control par-ticipants and patients with other causes of LVH (6,7). Notably, previous publications have not evaluated native myocardial T1 values at 3.0 T in AFD. Consistent with the findings of other studies, our study demonstrates that native RV T1 values are higher compared with native LV T1 values (21). Native RV T1 values correlate positively with LV T1 values, suggesting com-mon biventricular pathologic features (9). However, RV T1 val-ues have lower intraobserver and interobserver reproducibility compared with LV T1 values. Global LV native T1 and ECV values are significantly higher compared with septal LV native

.001, and P = .001, respectively) (Fig 4a). Global LV (P = .95), septal LV (P = .12), and RV (P = .22) ECV values did not differ significantly between AFD and HCM (Fig 4b).

The AUC to discriminate AFD from HCM was signifi-cantly higher for septal LV native T1 values (AUC, 0.979; 95% confidence interval [CI]: 0.951, 0.999) compared with global LV native T1 values (AUC, 0.832; 95% CI: 0.731, 0.933) (P = .001; Fig 5a). The AUC to discriminate AFD from HCM was significantly higher for septal LV native T1 values (AUC, 0.971; 95% CI: 0.930, 0.999) compared with RV native T1 values (AUC, 0.788; 95% CI: 0.647, 0.930) (P = .01; Fig 5b). A septal LV native T1 cutoff point of 1220 msec or lower distinguishes AFD from HCM with high sen-sitivity (29 of 30 [97%]), specificity (28 of 30 [93%]), and accuracy (57 of 60 [95%]).

Univariable logistic regression results are presented in Table 3. In a multivariable logistic regression model, septal LV native T1 values differentiate AFD from HCM with age, sex, and conventional imaging features controlled for (odds ratio, 0.94; 95% CI: 0.91, 0.98; P , .001). In a nested logistic re-gression model with age, sex, and conventional imaging fea-tures, model fit is significantly improved by the addition of LV native T1 values (x2 [df = 1] = 33.4; P , .001).

Intraobserver and Interobserver AgreementIntraobserver agreement and interobserver agreement were good to excellent for septal and global LV T1 and ECV values, but moderate to poor for RV T1 and ECV values (Table 4).

Table 1: Baseline Characteristics

Characteristic Anderson-Fabry Disease (n = 30) Hypertrophic Cardiomyopathy (n = 30) P ValueAge (y) 45.0 6 14.1 49.3 6 13.5 .24Female* 19 (63) 13 (43) .20Male* 11 (37) 17 (57) .20Hematocrit (g/L) 0.41 6 0.04 0.43 6 0.06 .07Height (cm) 167.3 6 9.7 172.1 6 10.9 .08Weight (kg) 75.4 6 16.7 87.5 6 18.9 .01Body mass index (kg/m2) 27.0 6 6.0 29.3 6 4.7 .10Body surface area (m2) 1.85 6 0.23 2.05 6 0.28 .004Systolic blood pressure (mm Hg) 120.9 6 17.9 127.8 6 18.4 .07Diastolic blood pressure (mm Hg) 76.4 6 10.9 77.0 6 10.0 .74Hypertension* 8 (27) 11 (37) .58Coronary artery disease* 0 (0) 1 (3) ..99Diabetes mellitus* 2 (7) 4 (13) .67Chronic kidney disease* 0 (0) 1 (3) ..99Medications* b-blockers 2 (7) 17 (57) ,.001 Statins 10 (33) 8 (27) .78 Calcium channel blockers 1 (3) 3 (10) .61 ACEIs/ARBs 9 (30) 2 (7) .04

Note.—Unless otherwise indicated, data are means 6 standard deviation. ACEI = angiotensin converting enzyme inhibitor, ARB = angio-tensin II receptor blocker.

* Data are the number of patients, with percentage in parentheses.

Karur et al


LV septal T1 value of 940 msec or lower with use of a shortened MOLLI sequence (without exclusion of areas of LGE) could be used to distinguish AFD from non-AFD causes of LVH. Thomp-son et al (6) reported that native LV global T1 cutoff points of 1146 msec or lower in women and 1120 msec or lower in men with use of a saturation recovery single-shot acquisition sequence allowed differentiation of patients with AFD from healthy con-trol participants and patients with LVH from other causes. Na-tive T1 values are higher at 3.0 T compared with 1.5 T because of the difference in magnetic field strength, which explains the higher cutoff point for AFD identified at 3.0-T imaging in our study (22,23).

T1 and ECV values, respectively, likely because of the absence of replacement fibrosis in the septum in patients with AFD (7). Although global LV native T1 values had slightly better repro-ducibility compared with septal LV native T1 values, septal LV native T1 values are recommended in the setting of patients sus-pected of having AFD, given better discrimination and ease of measurement. Although RV native T1 values are also reduced in AFD, RV native T1 assessment is not recommended for clinical use, given poor reproducibility.

Two prior studies in which 1.5-T imaging was used reported slightly lower native T1 cutoff points for AFD compared with the results from our study. Sado et al (7) reported that a native

Table 2: Cardiac MR Imaging Findings

Finding AFD (n = 30) HCM (n = 30) P ValueLVEDV (mL) 149.4 6 29.9 165.6 6 39.1 .08LVEDVi (mL/m2) 81.0 6 16.7 80.1 6 12.3 .81LVESV (mL) 57.9 6 13.6 60.9 6 18.8 .48LVESVi (mL/m2) 31.4 6 7.8 30.0 6 7.8 .47LVSV (mL) 91.7 6 19.7 104.3 6 25.8 .04LVEF (%) 61.3 6 4.7 63.1 6 6.2 .19LVM (g) 130.0 6 53.5 145.9 6 50.8 .24LVMi (g/m2) 68.7 6 29.4 71.1 6 22.1 .72Maximum LV wall thickness (mm) 12.9 6 4.0 18.2 6 4.6 ,.001LVH* 4 (13) 9 (30) .21RVEDV (mL) 151.3 6 35.5 148.1 6 42.6 .75RVEDVi (mL/m2) 82.0 6 18.7 72.1 6 14.7 .03RVESV (mL) 65.5 6 20.7 57.1 6 18.1 .10RVESVi (mL/m2) 35.5 6 10.9 28.1 6 7.2 .003RVSV (mL) 85.7 6 18.6 91.2 6 28.8 .39RVEF (%) 56.2 6 6.3 60.3 6 8.4 .03RVM (g) 37.8 6 9.0 36.9 6 9.1 .72RVMi (g/m2) 20.4 6 4.4 18.6 6 4.4 .12Maximum RV wall thickness (mm) 3.7 6 0.9 4.1 6 1.0 .16RVH* 2 (7) 1 (3) ..99Apical basal muscle bundle* 1 (3) 12 (40) .001Myocardial crypt* 3 (10) 9 (30) .10Any LGE* 23 (77) 28 (93) .15LGE excluding RV insertion points* 17 (57) 21 (70) .42Basal inferior-lateral LGE* 18 (60) 1 (3) .001Septal LV native T1 (msec) 1161 6 47 1296 6 55 ,.001Global LV native T1 (msec) 1192 6 52 1268 6 55 ,.001RV native T1 (msec)† 1221 6 54 1271 6 37 .001Septal LV ECV (%) 24.6 6 2.8 26.2 6 3.5 .07Global LV ECV (%) 25.5 6 2.2 25.6 6 3.1 .95RV ECV (%)† 34.0 6 5.7 32.1 6 3.7 .22

Note.—Unless otherwise indicated, data are means 6 standard deviation. AFD = Anderson-Fabry disease, ECV = extracellular volume, HCM = hypertrophic cardiomyopathy, LGE = late gadolinium enhancement, LV = left ventricle, LVEDV = left ventricular end diastolic volume, LVEDVi = LVEDV indexed to body surface area, LVEF = left ventricular ejection fraction, LVESV = left ventricular end systolic volume, LVESVi = LVESV indexed to body surface area, LVH = left ventricular hypertrophy, LVM = left ventricular mass, LVMi = LVM indexed to body surface area, LVSV = left ventricular stroke volume, RV = right ventricle, RVEDV = right ventricular end diastolic volume, RVEDVi = RVEDV indexed to body surface area, RVEF = right ventricular ejection fraction, RVESV = right ventricular end systolic vol-ume, RVESVi = RVESV indexed to body surface area, RVH = right ventricular hypertrophy, RVM = right ventricular mass, RVMi = RVM indexed to body surface area, RVSV = right ventricular stroke volume.* Data are the number of patients, with percentage in parentheses.† For RV native T1 and RV ECV values, n = 25 for AFD and n = 19 for HCM. A total of 16 patients were excluded from RV T1 and ECV analysis.



There was no significant difference in ECV values between groups, consistent with the results of a prior study performed at 1.5 T (6). ECV values may be within the normal range in AFD because AFD is an intracellular (lysosomal) storage disor-der (24). Although patients with HCM may have slightly higher ECV values compared to healthy volunteers as a result of diffuse interstitial fibrosis, ECV values can be within the upper range of normal, particularly when calculated from segments without LGE (24,25).

Distinguishing among different potential causes of LVH is a clinical and imaging challenge. Native cardiac MR imag-ing T1 values at both 1.5 T and 3.0 T are reduced in AFD compared with other causes of LVH; therefore, this imag-ing biomarker could be instrumental in the early identifica-tion of cardiac involvement in patients with AFD. Mul-tiple reports have stressed the role of early treatment and disease monitoring with cardiac MR imaging in patients

Figure 5: Graphs show receiver operating characteristic curves for left ventricular (LV) and right ventricular (RV) native T1 values for diagnosis of Anderson-Fabry disease (AFD) versus hypertrophic cardiomyopathy (HCM). (a) Area under curve (AUC) to discriminate AFD from HCM is signifi-cantly higher for septal LV native T1 mapping (blue dotted line; AUC, 0.979; 95% confidence interval [CI]: 0.951, 0.999) compared with global LV native T1 mapping (red dotted line; AUC, 0.832; 95% CI: 0.731, 0.933); n = 60; P = .001. (b) AUC to discriminate AFD from HCM is also significantly higher for septal LV native T1 mapping (blue dotted line; AUC, 0.971; 95% CI: 0.930, 0.999) compared with RV native T1 mapping (green dotted line; AUC, 0.788; 95% CI: 0.647, 0.930); n = 44; P = .01.

Figure 4: Box plots compare (a) septal and global left ventricular (LV) and right ventricular (RV) native T1 and (b) extracellular volume (ECV) values in patients with Anderson-Fabry disease (AFD) and hypertrophic cardiomyopathy (HCM).

Table 3: Univariable Model Examining Association between Cardiac MR Imaging T1 and ECV Values at 3.0 T and Diagnosis of Anderson-Fabry Disease Compared with Hypertrophic Cardiomyopathy

Finding Odds Ratio P ValueSeptal LV native T1 (msec) 0.94 (0.91, 0.98) ,.001Global LV native T1 (msec) 0.97 (0.96, 0.99) ,.001RV native T1 (msec) 0.97 (0.95, 0.99) .007Septal LV ECV (%) 0.86 (0.72, 1.02) .08Global LV ECV (%) 0.99 (0.82, 1.21) .95RV ECV (%) 1.09 (0.95, 1.25) .22

Note.—Data in parentheses are 95% confidence intervals. Odds ratios are for one-unit increases in native T1 values (1 msec) and extracellular volume (ECV) (1%), respectively. LV = left ventricle, RV = right ventricle.

Karur et al


Table 4: Intraobserver and Interobserver Agreement of T1 and ECV Measurements

Interobserver Agreement

Measurement Absolute Agreement Absolute Agreement ConsistencySeptal LV native T1 0.975 (0.938, 0.990) 0.792 (0.554, 0.912) 0.798 (0.558, 0.915)Global LV native T1 0.948 (0.875, 0.979) 0.950 (0.879, 0.980) 0.948 (0.873, 0.979)RV native T1 0.463 (0.005, 0.764) 0.224 (-0.374, 0.671) 0.212 (-0.339, 0.655)Septal LV ECV 0.648 (0.296, 0.847) 0.804 (0.503, 0.924) 0.840 (0.632, 0.935)Global LV ECV 0.719 (0.413, 0.881) 0.760 (0.478, 0.900) 0.753 (0.465, 0.897)RV ECV 0.770 (0.468, 0.912) 0.108 (-0.268, 0.576) 0.150 (-0.466, 0.668)

Note.—Data are individual intraclass correlation coefficients, with 95% confidence intervals in parentheses. ECV = extracellular volume, LV = left ventricle, RV = right ventricle.

with AFD (26–28). Renal impairment affects a substantial proportion of patients with AFD; end-stage renal disease was reported in 17% of men and 1% of women in a large cohort (29,30). In this setting, administration of gadolinium- based contrast material can be avoided with use of native myo-cardial T1 mapping for characterization of myocardial tissue. However, this remains to be evaluated in a separate study in-cluding participants with renal impairment. It is important to note that native T1 values are also shortened in the setting of fat and iron overload, although the clinical presentation typi-cally differs (31).

There were several limitations of this study. AFD is a relatively rare condition, thus limiting the number of participants. Imag-ing was performed with a single MR imager and T1 mapping sequence, and therefore results may not be generalizable to all cen-ters. Some of the patients with AFD included in our study were undergoing treatment with enzyme replacement therapy at the time of cardiac MR imaging, which could have affected T1 values. The diagnostic performance values for the cutoff point identified in our study are likely optimistic because they are derived from the same data that generated the model. Although care was taken to draw regions of interest in the RV myocardium, partial volume averaging and contamination of T1 values from the inadvertent inclusion of blood pool or epicardial fat are potential pitfalls that could result in artificially high RV native T1 and ECV values.

In conclusion, cardiac MR imaging native T1 values at 3.0 T are significantly lower in patients with AFD compared with HCM and provide independent and incremental diagnostic value beyond age, sex, and conventional imaging features.

Author contributions: Guarantors of integrity of entire study, G.R.K., C.F.M., K.H.; study concepts/study design or data acquisition or data analysis/interpreta-tion, all authors; manuscript drafting or manuscript revision for important intel-lectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately re-solved, all authors; literature research, G.R.K., A.M.C., S.M., K.H.; clinical stud-ies, G.R.K., S.R., R.M.I., C.F.M., A.M.C., S.M., K.H.; statistical analysis, G.R.K., A.M.C., P.T., K.H.; and manuscript editing, all authors

Disclosures of Conflicts of Interest: G.R.K. disclosed no relevant rela-tionships. S.R. disclosed no relevant relationships. R.M.I. disclosed no relevant relationships. C.F.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution received grants from Sanofi Genzyme and Shire Human Genetic Therapies for clinical research protocols; author received payment for lectures including service on speakers bureaus from Sanofi-Genzyme and Shire Human Genetic Therapies.

Other relationships: disclosed no relevant relationships. A.M.C. disclosed no rel-evant relationships. P.T. disclosed no relevant relationships. E.T.N. disclosed no relevant relationships. S.W. disclosed no relevant relationships. Activities related to the present article: institution received grant from Sanofi Genzyme and Shire Human Genetic Therapies for clinical research activities and payment for travel/accommodations/meeting expenses unrelated to activities listed; author received payment for lectures including service on speakers bureaus from the Canadian Fabry Association, Sanofi-Genzyme, and Shire Human Genetic Therapies. Other relationships: disclosed no relevant relationships. K.H. disclosed no relevant relationships.

References 1. Clarke JT. Narrative review: Fabry disease. Ann Intern Med 2007;146(6):425–433. 2. O’Mahony C, Elliott P. Anderson-Fabry disease and the heart. Prog Cardiovasc Dis

2010;52(4):326–335. 3. Morrissey RP, Philip KJ, Schwarz ER. Cardiac abnormalities in Anderson-

Fabry disease and Fabry’s cardiomyopathy. Cardiovasc J Afr 2011;22(1): 38–44.

4. Hughes DA, Elliott PM, Shah J, et al. Effects of enzyme replacement therapy on the cardiomyopathy of Anderson-Fabry disease: a randomised, double-blind, placebo-controlled clinical trial of agalsidase alfa. Heart 2008;94(2): 153–158.

5. Deva DP, Hanneman K, Li Q, et al. Cardiovascular magnetic resonance dem-onstration of the spectrum of morphological phenotypes and patterns of myo-cardial scarring in Anderson-Fabry disease. J Cardiovasc Magn Reson 2016; 18(1):14.

6. Thompson RB, Chow K, Khan A, et al. T1 mapping with cardiovascular MRI is highly sensitive for Fabry disease independent of hypertrophy and sex. Circ Cardio-vasc Imaging 2013;6(5):637–645.

7. Sado DM, White SK, Piechnik SK, et al. Identification and assessment of Anderson-Fabry disease by cardiovascular magnetic resonance noncontrast myocardial T1 map-ping. Circ Cardiovasc Imaging 2013;6(3):392–398.

8. Pica S, Sado DM, Maestrini V, et al. Reproducibility of native myocardial T1 mapping in the assessment of Fabry disease and its role in early detection of car-diac involvement by cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2014;16(1):99.

9. Pagano JJ, Chow K, Khan A, et al. Reduced right ventricular native myocardial T1 in Anderson-Fabry disease: comparison to pulmonary hypertension and healthy controls. PLoS One 2016;11(6):e0157565.

10. Gersh BJ, Maron BJ, Bonow RO, et al. 2011 ACCF/AHA guideline for the diag-nosis and treatment of hypertrophic cardiomyopathy: a report of the American Col-lege of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Developed in collaboration with the American Association for Thoracic Surgery, American Society of Echocardiography, American Society of Nuclear Car-diology, Heart Failure Society of America, Heart Rhythm Society, Society for Car-diovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Am Coll Cardiol 2011;58(25):e212–e260.

11. Kellman P, Arai AE, Xue H. T1 and extracellular volume mapping in the heart: estimation of error maps and the influence of noise on precision. J Cardiovasc Magn Reson 2013;15(1):56.

12. Kawel-Boehm N, Maceira A, Valsangiacomo-Buechel ER, et al. Normal values for cardiovascular magnetic resonance in adults and children. J Cardiovasc Magn Reson 2015;17(1):29.

13. Gruner C, Chan RH, Crean A, et al. Significance of left ventricular apical-basal muscle bundle identified by cardiovascular magnetic resonance imaging in pa-tients with hypertrophic cardiomyopathy. Eur Heart J 2014;35(39):2706–2713.

14. Deva DP, Williams LK, Care M, et al. Deep basal inferoseptal crypts occur more commonly in patients with hypertrophic cardiomyopathy due to disease-causing myofilament mutations. Radiology 2013;269(1):68–76.

15. Sheppard MN, Cane P, Florio R, et al. A detailed pathologic examination of heart tissue from three older patients with Anderson-Fabry disease on enzyme replacement therapy. Cardiovasc Pathol 2010;19(5):293–301.



16. Sado DM, Flett AS, Banypersad SM, et al. Cardiovascular magnetic resonance measurement of myocardial extracellular volume in health and disease. Heart 2012;98(19):1436–1441.

17. Niemann M, Herrmann S, Hu K, et al. Differences in Fabry cardiomyopathy be-tween female and male patients: consequences for diagnostic assessment. JACC Car-diovasc Imaging 2011;4(6):592–601.

18. Hanneman K, Crean AM, Wintersperger BJ, et al. The relationship between cardio-vascular magnetic resonance imaging measurement of extracellular volume fraction and clinical outcomes in adults with repaired tetralogy of Fallot. Eur Heart J Cardio-vasc Imaging 2017 Oct 17 [Epub ahead of print].

19. Arheden H, Saeed M, Higgins CB, et al. Measurement of the distribution volume of gadopentetate dimeglumine at echo-planar MR imaging to quantify myocardial infarction: comparison with 99mTc-DTPA autoradiography in rats. Radiology 1999;211(3):698–708.

20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44(3):837–845.

21. Swoboda PP, McDiarmid AK, Page SP, Greenwood JP, Plein S. Role of T1 Mapping in Inherited Cardiomyopathies. Eur Cardiol 2016;11(2):96–101.

22. Weingärtner S, Meßner NM, Budjan J, et al. Myocardial T1-mapping at 3T using saturation-recovery: reference values, precision and comparison with MOLLI. J Car-diovasc Magn Reson 2016;18(1):84.

23. Dabir D, Child N, Kalra A, et al. Reference values for healthy human myocardium using a T1 mapping methodology: results from the International T1 Multicenter cardiovascular magnetic resonance study. J Cardiovasc Magn Reson 2014;16(1):69.

24. Haaf P, Garg P, Messroghli DR, Broadbent DA, Greenwood JP, Plein S. Cardiac T1 mapping and extracellular volume (ECV) in clinical practice: a comprehensive review. J Cardiovasc Magn Reson 2016;18(1):89.

25. Swoboda PP, McDiarmid AK, Erhayiem B, et al. Assessing myocardial extracellular volume by T1 mapping to distinguish hypertrophic cardiomyopathy from athlete’s heart. J Am Coll Cardiol 2016;67(18):2189–2190.

26. Wu JC, Ho CY, Skali H, et al. Cardiovascular manifestations of Fabry disease: rela-tionships between left ventricular hypertrophy, disease severity, and alpha-galactosi-dase A activity. Eur Heart J 2010;31(9):1088–1097.

27. Weidemann F, Niemann M, Breunig F, et al. Long-term effects of enzyme replace-ment therapy on fabry cardiomyopathy: evidence for a better outcome with early treatment. Circulation 2009;119(4):524–529.

28. Cabrera G, Politei J, Antongiovani N, Amartino H; GADYTEF (Grupo Argentino de Diagnóstico y Tratamiento de la Enfermedad de Fabry). Effectiveness of enzyme replacement therapy in Fabry disease: Long term experience in Argentina. Mol Gen-et Metab Rep 2017;11:65–68.

29. Sunder-Plassmann G. Renal manifestations of Fabry disease. In: Mehta A, Beck M, Sunder-Plassmann G, eds. Fabry Disease: Perspectives from 5 Years of FOS. Oxford, England: Oxford PharmaGenesis, 2006.

30. Schiffmann R, Warnock DG, Banikazemi M, et al. Fabry disease: progres-sion of nephropathy, and prevalence of cardiac and cerebrovascular events before enzyme replacement therapy. Nephrol Dial Transplant 2009;24(7): 2102–2111.

31. Hanneman K, Nguyen ET, Thavendiranathan P, et al. Quantification of myocardial extracellular volume fraction with cardiac MR imaging in thalassemia major. Radiol-ogy 2016;279(3):720–730.

use of myocardial t1 mapping at 3.0 t to differentiate ... · myocardial t1 mapping in...

Documents