Native T1 Mapping in Differentiation of Normal Myocardium From Diffuse Disease in Hypertrophic and Dilated Cardiomyopathy

Article (PDF Available)inJACC. Cardiovascular imaging 6(4) · March 2013with51 Reads
DOI: 10.1016/j.jcmg.2012.08.019 · Source: PubMed
Abstract
OBJECTIVES: The aim of this study was to examine the value of native and post-contrast T1 relaxation in the differentiation between healthy and diffusely diseased myocardium in 2 model conditions, hypertrophic cardiomyopathy and nonischemic dilated cardiomyopathy. BACKGROUND: T1 mapping has been proposed as potentially valuable in the quantitative assessment of diffuse myocardial fibrosis, but no studies to date have systematically evaluated its role in the differentiation of healthy myocardium from diffuse disease in a clinical setting. METHODS: Consecutive subjects undergoing routine clinical cardiac magnetic resonance at King's College London were invited to participate in this study. Groups were based on cardiac magnetic resonance findings and consisted of subjects with known hypertrophic cardiomyopathy (n = 25) and nonischemic dilated cardiomyopathy (n = 27). Thirty normotensive subjects with low pre-test likelihood of cardiomyopathy, not taking any regular medications and with normal cardiac magnetic resonance findings including normal left ventricular mass indexes, served as controls. Single equatorial short-axis slice T1 mapping was performed using a 3-T scanner before and at 10, 20, and 30 minutes after the administration of 0.2 mmol/kg of gadobutrol. T1 values were quantified within the septal myocardium (T1native), and extracellular volume fractions (ECV) were calculated. RESULTS: T1native was significantly longer in patients with cardiomyopathy compared with control subjects (p < 0.01). Conversely, post-contrast T1 values were significantly shorter in patients with cardiomyopathy at all time points (p < 0.01). ECV was significantly higher in patients with cardiomyopathy compared with controls at all time points (p < 0.01). Multivariate binary logistic regression revealed that T1native could differentiate between healthy and diseased myocardium with sensitivity of 100%, specificity of 96%, and diagnostic accuracy of 98% (area under the curve 0.99; 95% confidence interval: 0.96 to 1.00; p < 0.001), whereas post-contrast T1 values and ECV showed lower discriminatory performance. CONCLUSIONS: This study demonstrates that native and post-contrast T1 values provide indexes with high diagnostic accuracy for the discrimination of normal and diffusely diseased myocardium.
Native T1 Mapping in Differentiation of
Normal Myocardium From Diffuse Disease in
Hypertrophic and Dilated Cardiomyopathy
Valentina O. Puntmann, MD, PHD,* Tobias Voigt, PHD,† Zhong Chen, MD,*
Manuel Mayr, MD, PHD,‡ Rashed Karim, PHD,* Kawal Rhode, PHD,* Ana Pastor, MD,*
Gerald Carr-White, MBBS, PHD,* Reza Razavi, MD,* Tobias Schaeffter, PHD,*
Eike Nagel, MD, PHD*
London, United Kingdom
OBJECTIVES The aim of this study was to examine the value of native and post-contrast T1
relaxation in the differentiation between healthy and diffusely diseased myocardium in 2 model
conditions, hypertrophic cardiomyopathy and nonischemic dilated cardiomyopathy.
BACKGROUND T1 mapping has been proposed as potentially valuable in the quantitative
assessment of diffuse myocardial fibrosis, but no studies to date have systematically evaluated its role
in the differentiation of healthy myocardium from diffuse disease in a clinical setting.
METHODS Consecutive subjects undergoing routine clinical cardiac magnetic resonance at King’s
College London were invited to participate in this study. Groups were based on cardiac magnetic
resonance findings and consisted of subjects with known hypertrophic cardiomyopathy (n 25) and
nonischemic dilated cardiomyopathy (n 27). Thirty normotensive subjects with low pre-test likelihood
of cardiomyopathy, not taking any regular medications and with normal cardiac magnetic resonance
findings including normal left ventricular mass indexes, served as controls. Single equatorial short-axis
slice T1 mapping was performed using a 3-T scanner before and at 10, 20, and 30 minutes after the
administration of 0.2 mmol/kg of gadobutrol. T1 values were quantified within the septal myocardium
(T1
native
), and extracellular volume fractions (ECV) were calculated.
RESULTS T1
native
was significantly longer in patients with cardiomyopathy compared with control
subjects (p 0.01). Conversely, post-contrast T1 values were significantly shorter in patients with
cardiomyopathy at all time points (p 0.01). ECV was significantly higher in patients with cardiomyopathy
compared with controls at all time points (p 0.01). Multivariate binary logistic regression revealed that
T1
native
could differentiate between healthy and diseased myocardium with sensitivity of 100%, specificity of
96%, and diagnostic accuracy of 98% (area under the curve 0.99; 95% confidence interval: 0.96 to 1.00; p
0.001), whereas post-contrast T1 values and ECV showed lower discriminatory performance.
CONCLUSIONS This study demonstrates that native and post-contrast T1 values provide indexes with
high diagnostic accuracy for the discrimination of normal and diffusely diseased myocardium. (J Am Coll Cardiol
Img 2013;xx:xxx) © 2013 by the American College of Cardiology Foundation
From the *Department of Cardiovascular Imaging, King’s College London, London, United Kingdom; †Philips Innovative
Technologies, Clinical Research, London, United Kingdom; and the ‡Cardiovascular Division, King’s College London, London,
United Kingdom. This study was funded by the Department of Health via the National Institute for Health Research comprehensive
Biomedical Research Centre award to Guy’s & St. Thomas’ NHS Foundation Trust in partnership with King’s College London and
King’s College Hospital NHS Foundation Trust. Dr. Schaeffter has received research support from Philips Healthcare. All other
authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Manuscript received July 3, 2012; revised manuscript received August 6, 2012, accepted August 9, 2012.
JACC: CARDIOVASCULAR IMAGING VOL. xx, NO. x, 2013
© 2013 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION ISSN 1936-878X/$36.00
PUBLISHED BY ELSEVIER INC. http://dx.doi.org/10.1016/j.jcmg.2012.08.019
M
yocardial fibrosis is a fundamental pro-
cess in the development of myocardial
dysfunction in various cardiomyopa-
thies, leading to myocardial remodeling
and poor outcomes (1–5). Cardiac magnetic reso-
nance (CMR) is increasingly applied as the first-
line investigation into the causes of cardiomyopa-
thies (6). Visualization of fibrosis by CMR is based
on a greater distribution volume and slower wash-
out of gadolinium contrast agents within tissues
with greater extracellular space due to edema or
fibrosis (7). Whereas regional fibrosis after ischemic
injury is readily distinguished by well-delineated
areas of increased signal intensity on T1-weighted
images by late gadolinium enhancement (LGE) (8),
it may be impossible to define an area of clearly
unaffected myocardium as a “nulled” reference in
diffuse fibrotic processes (Figs. 1A and 1B) (9). As
a consequence, such images may null the signal in
areas of fibrosis, obscuring the finding or
result in images with various gray values,
not allowing a clear “yes or no” decision
(9,10). Recently, several studies have pro-
posed the measurement of T1 relaxation
as potentially valuable for the quantitative
assessment of myocardial fibrosis (11–16).
In these studies, native myocardium with
ischemic scar showed longer T1 values
compared with unaffected remote myocar-
dium. After contrast administration, re-
gional and also diffusely scarred myocar-
dium showed shorter T1 relaxation and
delayed normalization of T1 times with
gadolinium washout. Whereas these ob-
servations show the potential of T1 map-
ping for the evaluation of myocardial fibrosis, these
studies used a variety of imaging methodologies and
post-processing approaches. The ability of T1 map-
ping to differentiate between normal and abnormal
myocardium is yet to yield a clinically robust appli-
cation. In the present study, we aimed to examine
the value of native and post-contrast T1 relaxation
in the differentiation of healthy and diffusely dis-
eased myocardium in 2 model conditions, hypertro-
phic cardiomyopathy (HCM) and nonischemic di-
lated cardiomyopathy (NIDCM).
METHODS
Consecutive subjects undergoing routine clinical
CMR at King’s College London were invited to
participate in this study. Groups were based on CMR
findings and consisted of subjects with known HCM
(n 25) and NIDCM (n 27) and controls (n
30). Diagnosis of HCM was based on the demonstra-
tion of a hypertrophied left ventricle associated with a
nondilated left ventricle (LV) in the absence of in-
creased LV wall stress or another cardiac or systemic
disease that could result in a similar magnitude of
hypertrophy (17,18). All patients with HCM had an
expressed phenotype with typically asymmetric septal
hypertrophy of increased LV wall thickness, permit-
ting unequivocal clinical diagnoses. NIDCM was
defined as an increase in LV volumes, a reduction in
global systolic function, and absence of evidence of
ischemic-like LGE (18). Thirty normotensive sub-
jects with low pre-test likelihood for LV cardiomyop-
athy, not taking any regular medications and, conse-
quently, with normal CMR findings including normal
LV mass indexes, served as the control group (19).
Additional exclusion criteria for all subjects were the
generally accepted contraindications to CMR (im-
plantable devices, cerebral aneurysm clips, cochlear
implants, severe claustrophobia) or a history of renal
disease with a current estimated glomerular filtration
rate 30 ml/min/1.73 m
2
. The study protocol was
reviewed and approved by the institutional ethics
committee, and written informed consent was ob-
tained from all participants.
CMR protocol. We integrated native and post-contrast
myocardial T1 mapping into our routine imaging
protocol for the determination of the underlying
etiology of cardiomyopathy; an outline is provided in
Figure 2. The CMR studies were performed with the
patient supine, using a clinical 3-T scanner (Achieva
TX, Philips Healthcare, Best, the Netherlands) and a
32-channel coil. After standardized patient-specific
planning (20), volumetric cavity assessment was ob-
tained by whole-heart coverage of gapless short-axis
slices. Thereafter, cine images of 3 long-axis views
(4-chamber, 2-chamber, and 3-chamber views) and
transverse axial views were acquired. All cine-images
were acquired using a balanced steady-state free pre-
cession sequence in combination with parallel imaging
(SENSitivity Encoding, factor 2) and retrospective
gating during a gentle expiratory breath-hold (echo
time [TE]/repetition time [TR]/flip-angle: 1.7 ms/
3.4 ms/60°, spatial resolution 1.8 1.8 8 mm).
LGE imaging was performed in a gapless whole heart
coverage of short axis slices 20 min after administra-
tion of a cumulative dose of 0.2 mmol/kg body weight
gadobutrol using a mid-diastolic inversion prepared
2-dimensional gradient echo sequence (TE/TR/flip-
angle 2.0 ms/3.4 ms/25°, interpolated voxel size 0.7
0.7 8 mm, and a patient-adapted prepulse delay, as
appropriate. A steady-state free precession, single-
ABBREVIATIONS
AND ACRONYMS
CI confidence interval
CMR cardiac magnetic
resonance
HCM hypertrophic
cardiomyopathy
HR heart rate
LGE late gadolinium
enhancement
LV left ventricular
NIDCM nonischemic
dilated cardiomyopathy
ROI region of interest
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
2
Figure 1. Examples of LGE Images in Conditions With Diffuse Myocardial Involvement
Look-Locker images and late gadolinium enhancement (LGE) images in burnt-out hypertrophic cardiomyopathy (A) and dilated cardio-
myopathy after previous myocarditis (B). There was a reduced overall relative difference in signal between affected and unaffected myo-
cardium in Look-Locker images, leading to difficulty in nulling of the normal myocardium. 4CH 4-chamber; SAX short-axis; 3CH
3-chamber; 2CH 2-chamber.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
3
breath-hold modified Look-Locker inversion re-
covery sequence was used for T1 mapping, per-
formed in an equatorial short-axis slice before
and at 10, 20, and 30 min after contrast admin-
istration. Imaging parameters were as follows: field of
view 320 320, TR 3.3 ms, TE 1.57 ms, flip angle
50°, interpolated voxel size 0.9 0.9 8 mm, 166
phase-encoding steps, heart rate (HR)–adapted trig-
ger delay, 11 phases (3 3 5), and adiabatic
pre-pulse to achieve a complete inversion (14 –16).
Image analysis. All routine CMR analysis was per-
formed using commercially available software
(ViewForum, Extended Workspace, Philips
Healthcare). Endocardial LV borders were manu-
ally traced at end-diastole and end-systole. The pap-
illary muscles were included as part of the LV cavity
volume. LV end-diastolic and end-systolic volumes
were determined using Simpson’s rule. Ejection frac-
tion was computed as end-diastolic volume end-
systolic volume/end-diastolic volume. All volumetric
indexes were normalized to body surface area.
The LGE images were visually examined for the
presence of regional fibrosis. Global enhancement was
defined as percent of enhanced area per total short-
axis stack, where enhanced area was defined by 6 SDs
above the manually selected normal area, appearing as
maximally suppressed myocardium (10).
T1 relaxation maps were obtained using Relax-
Maps tools supported by the PRIDE environment
(Philips Healthcare). Selective acquisition at a fixed
point of the cardiac cycle in end-diastole largely
suppressed the influence of cardiac motion, but the
relatively long duration of the sequence (17 heartbeats
to obtain a single slice map) occasionally led to some
undesired breathing motion. We therefore performed
a motion correction image preparation step using a
custom-made tool developed in house on the basis of
a hierarchical adaptive local affine registration tech-
nique, as previously described (21), in which a refer-
ence phase (source) is registered to each of the selected
target phases (11 in total). A rectangular region of
interest (ROI) large enough to enclose the whole of
the LV is manually drawn onto the source reference
phase before registration. After the initial image affine
registration step of the ROI, the source image is
subdivided into 4 smaller ROIs using equal subdivi-
sions. Each subdivision underwent an affine transfor-
mation again (each with 6 degrees of freedom) to align
the features of the target image ROIs with the
corresponding ROIs in the source phase image.
Coregistered images were then used to derive T1
values. Because previous studies showed substantial
segmental variation in T1 values, which was greatest
in lateral and smallest in the septal segments (15,16),
an interobserver consensus was reached to place the
ROIs within the septal myocardium (Figs. 3 and 4).
Care was also taken to avoid “contamination” with
signal from the blood pool. Reported T1 values were
derived blinded to the LGE images to enable high
interobserver and intraobserver reproducibility of T1
measurements. We additionally examined the influ-
ence of visualized LGE on the T1 values, regionally
and within the septal ROIs. T1 was determined
by fitting a 3-parameter exponential model to the
Start
Start
Native T1 T1
10min
T1
20min
T1
30min
Planning Short-axis
Stack
Transverse
Stack
Scar
Imaging
EndContrast Injection
Figure 2. Integration of MOLLI Sequences in Clinical Routine
Modified Look-Locker inversion recovery (MOLLI) sequences were inte-
grated into the clinical protocol for the investigation of cardiomyopa-
thy, with set spacing between the time points. Time points denote the
interval from gadolinium contrast administration.
Figure 3. T1 Map With Color Scale and Region of Interest
With interobserver consensus, the region of interest was placed
conservatively within the septal myocardium. Care was also taken
to avoid “contamination” with signal from the blood pool.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
4
measured data and applying Look-Locker correc-
tion, as previously described (14). Noise was calcu-
lated in an ROI drawn manually inside the lungs
and taken into account in the T1 computation (22).
Because longitudinal relaxation is HR dependent,
we also applied HR correction of T1 values when
HR exceeded 80 beats/min, as previously described
(14,16). In addition to the T1 values of myocardium and
blood pool, we calculated volume fraction of extravascular
extracellular matrix (ECV) according to the formula
(22–25) ECV [lambda (1 hematocrit)], where
ECV is the myocardial extravascular extracellular volume
fraction, and lambda [R1myocardium]/
[R1bloodpool] before and after gadolinium contrast
(where R1 1/T1).
Statistical analysis. Statistical analysis was per-
formed using SPSS version 20 (SPSS, Inc., Chi-
cago, Illinois). Differences from the control group
and between the groups were examined using one-
way and repeated-measures analysis of variance
with Bonferroni post hoc tests, as appropriate.
Reproducibility and agreement analysis was per-
formed using paired Student t tests and bivariate
correlations. Associations with demographic and
hemodynamic variables were detected by bivariate
linear regression analyses. Multivariate binary logis-
tic regression was used to test the ability of T1-
derived measures in discrimination between healthy
and abnormal myocardium. The sensitivity, speci-
ficity, and diagnostic accuracy, areas under the
curve, and cutoff values were derived using receiver-
operating characteristic curve analysis using the
point that maximized the trade-off between speci-
ficity and sensitivity. Z-scores were used to compare
the areas under the curve. All tests were 2 tailed,
and p values 0.05 were considered significant.
RESULTS
Patients’ characteristics, hemodynamic variables,
and cardiac function are presented in Table 1. All
groups had similar sex representations, HRs, and
body surface areas. In comparison with controls,
patients with HCM had increased global systolic
function and diastolic LV wall thickness (p 0.01).
Compared with controls, subjects with NIDCM
had increased cavity volumes and reduced global
systolic function (p 0.01). Both patient groups
had increased LV mass index (p 0.01). Eighteen
patients with HCM and 9 with NIDCM had
evidence of LGE on scar imaging, as patchy LGE
or intramyocardial stria.
T1 of myocardium and blood pool and ECV. T1 values
of blood pool were similar between the groups. T1
of native myocardium (T1
native
) was significantly
longer in cardiomyopathies compared with control
subjects (p 0.01) (Table 2). Conversely, post-
contrast T1 values were significantly shorter in the
presence of cardiomyopathy at all time points (p
0.01). Similarly, lambda values were significantly
Figure 4. Representative Images of Modified Look-Locker Inversion Recovery Imaging
The region of interest was placed conservatively within the septal myocardium to avoid “contamination” with signal from the blood
pool. Blood T1 values were obtained by placing the region of interest in the center of the ventricular lumen. SI signal intensity.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
5
higher in patients with cardiomyopathy in compar-
ison with controls (p 0.01). ECV values were
similar at all time points for the total cohort (10 vs.
20 vs. 30 min: 28 11 vs. 29 10 [p 0.35] vs.
27 9[p 0.22]). ECV in patients was signifi-
cantly higher compared with controls at all time
points (p 0.01).
There were no significant differences in mean T1
values between subjects with visually detectable LGE
and those without it (T1
native
: HCM, 1,241 51 ms vs.
1,234 71 ms, p 0.88, NIDCM, 1,290 52 ms vs.
1,301 49 ms, p 0.58; T1
20min
: HCM, 367 42 ms
vs. 361 39 ms, p 0.91, NIDCM, 346 43 ms vs.
358 48 ms, p 0.60). Regions of visualized LGE
showed no significant difference from the septal
ROI T1 values for T1
native
and T1
20min
maps
(T1
native
septal ROI vs. regional ROI mean difference
2.7 9.4, p 0.91; T1
20min
mean difference 1.9
13, p 0.92).
Reproducibility and agreement analysis. In a subset
of subjects (n 47), interobserver mean differences
for T1 values were 1.3 ms (95% confidence interval
[CI]: 12.4 to 15.4 ms) for native scans and 0.7 ms
(95% CI: 7.9 to 12.3 ms) for overall post-contrast
scans, whereas intraobserver mean differences were
0.3 ms (95% CI: 6.3 to 5.3) and 0.1 ms (95%
CI: 3.4 to 4.2), respectively. We demonstrate
excellent overall (pre-contrast and post-contrast)
intraobserver and interobserver agreement in T1
Table 1. Patient Characteristics and Global Morphological and Functional Measures
Controls Patients With HCM Patients With NIDCM
(n 30) (n 25) (n 27)
Men 19 (63) 16 (64) 18 (67)
Age, yrs 43 944 11 45 14
Systolic BP, mm Hg 113 6 115 6 120 6
Diastolic BP, mm Hg 69 872 675 10
HR, beats/min 64 10 66 966 9
Body mass index, kg/m
2
24 324 323 4
LVEDV index, ml/m
2
76 970 10 110 15*
LV ejection fraction, % 63 672 8† 34 6*
LV mass index, g/m
2
55 899 10* 106 11*
Maximal LVWT, mm 9 218 2* 10 2
Global enhancement, % 6 512 11* 10 8†
eGFR, ml/min/1.73 m
2
82 10 83 974 12
Hematocrit, % 44 343 343 2
Values are n (%) or mean SD. *p 0.01 (Bonferroni post hoc tests for differences from the control group)
and †p 0.05.
BP blood pressure; eGFR estimated glomerular filtration rate; HCM hypertrophic cardiomyopathy;
HR heart rate; LV left ventricular; LVEDV left ventricular end-diastolic volume; LVWT left
ventricular wall thickness; NIDCM nonischemic dilated cardiomyopathy.
Table 2. Native and Post-Contrast T1 Relaxation Times
Controls Patients With HCM Patients With NIDCM
(n 30) (n 25) (n 27)
Native
T1 myocardium, ms 1,070 55 1,254 43* 1,239 57*
T1 blood, ms 1,871 89 1,869 92 1,901 92
R1 myocardium, msec
1
10
5
94 480 4* 76 4*
10 min
T1 myocardium, ms 402 58 307 47* 296 43*
T1 blood, ms 247 34 251 39 245 38
R1 myocardium, msec
1
10
5
254 40 339 44* 345 41*
Lambda 0.46 0.13 0.72 0.18* 0.70 0.18*
ECV 0.27 0.1 0.40 0.1* 0.41 0.1*
20 min
T1 myocardium, ms 440 58 363 63* 355 44*
T1 blood, ms 294 32 299 36 297 33
R1 myocardium, msec
1
10
5
230 31 284 50* 284 30*
Lambda 0.49 0.12 0.73 0.2* 0.73 0.09*
ECV 0.27 0.09 0.41 0.12* 0.40 0.09*
30 min
T1 myocardium, ms 504 38 437 49† 444 45†
T1 blood, ms 381 42 369 37 351 45
R1 myocardium, msec
1
10
5
199 14 230 20* 227 23*
Lambda 0.47 0.09 0.67 0.10* 0.67 0.12*
ECV 0.26 0.07 0.38 0.11* 0.38 0.1*
*p 0.01 (Bonferroni post hoc tests for differences from the control group) and †p 0.05.
ECV extracellular volume; other abbreviations as in Table 1.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
6
values (intraobserver: r 0.99, p 0.0001; inter-
observer: r 0.98, p 0.0001).
Comparisons of T1 measures in discrimination between
normal and abnormal myocardium.
In multivariate
binary logistic regression model using native and
post-contrast T1 and ECV values, we identified
T1
native
as the independent discriminator of cardio-
myopathic myocardium (p 0.001), with sensitiv-
ity of 100%, specificity of 96%, diagnostic accuracy
of 98%, a positive predictive value of 98%, and a
negative predictive value of 100%. Comparison of
areas under the curve showed that T1
native
provided
the best distinction between controls and patients
with cardiomyopathy (T1
native
vs. T
10min
: z 2.2,
p 0.05; T1
native
vs. ECV
20min
: z 2.4, p 0.02).
Results of receiver-operating characteristic curve
analysis with corresponding cutoff values for the
performance of separate T1-derived measures in the
differentiation of normal from abnormal myocar-
dium are presented in Table 3 and Figure 5.
Analysis of relationships. Within the separate patient
groups, there was a positive association between
T1
native
and ECV and age (T1
native
:r 0.53, p
0.03; ECV: r 0.51, p 0.05). In controls and
patients with NIDCM, myocardial T1
native
showed
negative associations with indexed LV end-diastolic
volume (r 0.37 and 0.54, respectively, p 0.05).
Patients with NIDCM showed further associations
between myocardial T1
native
and ejection fraction
(r ⫽⫺0.61, p 0.01), whereas in patients with HCM,
there was a positive association between myocardial
T1
native
and indexed LV mass (r 0.51, p 0.01).
DISCUSSION
Our study reveals that diffusely diseased myocar-
dium can be reliably differentiated from healthy
myocardium by means of T1 mapping. We dem-
onstrate that in HCM and NIDCM, native and
post-contrast T1 values provide indexes with high
diagnostic accuracy, sensitivity, and specificity, with
T1
native
providing the greatest distinction between
healthy and diffusely diseased myocardium. We
further demonstrate that in NIDCM, T1
native
cor-
relates with measures of LV remodeling and global
systolic function, whereas in HCM, it shows an
association with indexed LV mass. Our findings
provide a novel and easy-to-use method for the
detection of diffusely diseased myocardial tissue by
CMR with an immediate potential for clinical
translation.
T1 mapping techniques provide quantifiable in-
formation on longitudinal relaxation through the
acquisition of images with different inversion times
T1 native
ECV10
ECV20
ECV30
Source of the Curve
Sensitivity
1 - Specificity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Figure 5. ROC Curves for T1-Derived Values in Differentiation Between
Healthy and Abnormal Myocardium
Receiver-operating characteristic (ROC) curve analysis for the performance of
separate T1-derived measures in the differentiation of normal from abnor-
mal myocardium. Myocardial T1
native
imaging provided the best distinction
between controls and patients with cardiomyopathy. ECV extracellular
volume.
Table 3. Results of Cutoff Values on Receiver-Operating Characteristic Curve Analysis
Cutoff
Value AUC (95% CI) Significance (p)
Specificity (%)
(95% CI)
Sensitivity (%)
(95% CI)
Diagnostic Accuracy
(%) (95% CI)
T1 value, ms
Native 1,184 0.99 (0.98–1.00) 0.000 97 (82–100) 100 (90–100) 99 (91–99)
10 min 330 0.90 (0.87–0.97) 0.000 71 (52–85) 86 (72–94) 80 (68–88)
20 min 407 0.86 (0.79–0.96) 0.000 71 (52–85) 82 (67–92) 77(65–86)
30 min 477 0.84 (0.74–0.96) 0.000 68 (49–83) 71 (59–87) 70 (57–79)
ECV
ECV
10min
23 0.86 (0.76–0.95) 0.000 70 (48–85) 80 (65–90) 78 (63–86)
ECV
20min
23 0.85 (0.77–0.95) 0.000 72 (50–87) 82 (66–92) 78 (65–87)
ECV
30min
23 0.88 (0.80–0.96) 0.000 68 (46–85) 82 (64–90) 77 (64–86)
AUC area under the curve; CI confidence interval; other abbreviations as in Table 2.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
7
and by multiparametric curve-fitting analysis. T1
maps are derived as parametric reconstructed im-
ages, in which the signal intensity of a pixel depends
on the absolute longitudinal relaxation properties of
this voxel (13–16,22). Several methodologies were
tested to acquire the myocardial T1 relaxation
values, including sets of saturation or inversion
recovery images (12) with varying inversion times
and, lately, the classical and modified Look-Locker
sequences (14 –16,23). The variant of the latter
sequence, which was also applied in the present
study, leads to a series of multiple images acquired
within the same phase of cardiac cycle through a
selective fixed delay time over successive heartbeats
(14). Some studies using this or related methodol-
ogies suggested that myocardial T1
native
values
could be used to discern post-infarct scar from
healthy myocardium (15,16). Most of these studies,
however, focused on the post-contrast T1 values
and reported significantly shorter T1 times com-
pared with controls. In a population of patients with
mixed causes of heart failure, Iles et al. (12) reported
shorter post-contrast T1 compared with controls,
even when excluding areas of regional fibrosis.
These investigators also observed an inverse rela-
tionship between post-contrast T1 values and the
amount of fibrosis on histology (12). The ability of
T1 imaging to quantify the amount of diffuse
fibrosis was also confirmed using a novel technique
of ECV imaging deriving ECV using continuous
infusion (24,25). Diastolic myocardial impairment,
an indirect marker of diffuse myocardial fibrosis,
was shown to correlate with abnormal post-contrast
T1 values in patients with heart failure, diabetic
cardiomyopathy, and amyloidosis (11,26–28).
Whereas Iles et al. (12) showed no significant
difference in myocardial T1
native
values in patients
with heart failure, our findings reveal for the first
time that native T1 values are significantly higher in
diffusely disease myocardium, which cannot be
accounted for by the presence of visualized LGE.
The disparity with the former findings may lie in
the differences in patient selection, as well as imag-
ing techniques and higher field strength used in the
present study, which explains higher measurements
of longitudinal relaxation in native myocardium
(15,16,29).
In our study, myocardial T1
native
provides the
greatest distinction between healthy myocardium
and diseased tissue with high negative predictive
value. As such, it bears potential for the develop-
ment of an easy-to-implement test in patients with
suspected diffuse fibrosis, which may be missed by
classic LGE imaging. Furthermore, in subjects with
low pre-test likelihood for the presence of cardio-
myopathy (descriptive of our control group), or
those in whom contrast administration is contrain-
dicated, it may serve as an effective screening test.
Future advances in sequence development that
would provide whole-heart coverage might derive a
useful approach to characterize regional differences
and potentially obviate the need for contrast admin-
istration.
The observed findings in myocardial T1
native
emphasize several important aspects with regard to
post-contrast T1 mapping. Because gadolinium
administration greatly shortens T1 values, the over-
all T1 tissue relaxation will depend on the dose and
relaxivity of the gadolinium contrast agent, the
intrinsic T1 values of the tissue (13,23,29), and the
timing of the acquisition and bioavailability after
gadolinium administration (7,9). Post-contrast T1
sampling can thus be affected by a variety of
independent variables, including renal function,
contrast type and dose of administration, variation
in sampling time points, and individual pharmaco-
kinetics (22,30). Post-contrast T1 imaging at the
rigid time points can prove cumbersome in clinical
routine; we improved the interstudy comparability
of the post-contrast T1 sampling by using consis-
tent time points in our routine cardiomyopathy
imaging protocol. Yet the aforementioned influ-
ences might explain the lower performance of
post-contrast T1 values and ECV.
In our study, T1 values were sampled in 2 model
conditions of diffuse myocardial fibrosis. Whereas
several investigators have looked at the role of T1
mapping in patients with heart failure, no previous
study has systematically assessed native and post-
contrast T1 values in patients with HCM. It is well
established that visualization of LGE in HCM has
important and independent prognostic implications
(5), but recent evidence suggests that a profibrotic
state through genetically driven collagen metabo-
lism precedes the overt phenotype with LV hyper-
trophy or fibrosis visible on LGE (4). We previ-
ously demonstrated that global enhancement
correlates with reduction of longitudinal ventricular
deformation in HCM, even when global systolic
function remains apparently unaffected (10).
Whether T1-derived measures are able to detect
subclinical change in collagen metabolism remains
to be determined. The identification of early phe-
notypes in which early fibrotic process could be
quantified and followed up before the effects on
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
8
cardiac geometry and function would add to the
management of this condition (4).
The observed changes in post-contrast T1 values
have been previously related to increase in extracel-
lular space and well described in models of acute
and chronic ischemic or inflammatory myocardial
injury. The physiological correlate with diffuse
myocardial pathology and observed increases in
native T1 is less well understood. Extracellular
matrix remodeling is orchestrated by fibroblasts
within the heart, but changes in extracellular
matrix composition affect cardiomyocyte survival
(1–3,31). The accumulation of extracellular ma-
trix in pathological hypertrophy or LV remodel-
ing with increasing age and LV mass may under-
lie the observed increase in myocardial T1
native
values (1). Further studies are needed to elucidate
the links between exact molecular mechanisms in
health and disease and the corresponding T1
native
readouts.
Study limitations. The small sample size of other-
wise similar groups may limit the associations with
sex and age and generalization of the present
findings, and a larger multicenter study is required
to reconfirm our imaging protocol beyond the
proof of concept for widespread clinical use.
Examination of the equivalence of multivendor
sequences may also be appropriate. Next, T1
sampling in the septum of a single short-axis slice
is based on the assumption that it is representa-
tive of the diffuse myocardial process, but future
studies with multiple slices are needed to deter-
mine and understand relevant regional variation.
The prolonged acquisition time required for lon-
gitudinal relaxation to occur in the native myo-
cardium, especially at high field strengths, its HR
dependence, and motion artifacts may all amount
to errors in the pixel-wise estimation of T1 values
(32). Because we used motion and HR correction
algorithms and a very conservative approach to
T1 sampling, we believe that we controlled for
the majority of these influences. Whether imag-
ing with shorter variants of modified Look-
Locker inversion recovery sequences would re-
solve some of the breathing motion artifacts and
still provide sufficient information on longitudi-
nal relaxation of the native myocardium remains
to be confirmed in future studies. Similarly, the
effects of T2 recovery when longitudinal relax-
ation is measured with modified Look-Locker
inversion recovery remain to be determined.
Whereas in the present study, we looked at 2
models of diffuse myocardial fibrosis, we assume
that the presence of edema and its influence on
T1 values is unlikely a relevant substrate. Lastly,
T1 values in other cardiomyopathies need to be
examined in future studies.
CONCLUSIONS
We demonstrate that native and post-contrast T1
values provide indexes with high diagnostic accu-
racy for the discrimination of normal and diffusely
diseased myocardium, as observed in patients with
HCM and NIDCM in the present study. Myocar-
dial T1
native
imaging provides the best distinction
between controls and patients with cardiomyopa-
thy. Further studies are needed to effectively trans-
late these findings into clinical use.
Acknowledgment
The authors acknowledge Lorna Smith, head of
research radiography in the Department of Cardio-
vascular Imaging Sciences, King’s College London,
for her assistance with high-quality data acquisition
and clinical application support.
Reprint requests and correspondence: Prof. Eike Nagel,
King’s College London, Department of Cardiovascular
Imaging, The Rayne Institute, 4th Floor Lambeth Wing,
St. Thomas’ Hospital Campus, London SE1 7EH,
United Kingdom. E-mail: eike.nagel@kcl.ac.uk.
REFERENCES
1. Hill JA, Olson EN. Cardiac plasticity.
N Engl J Med 2008;358:1370 80.
2. Van den Borne SW, Diez J, Blank-
esteijn WM, Verjans J, Hofstra L,
Narula J. Myocardial remodeling after
infarction: the role of myofibroblasts.
Nat Rev Cardiol 2010;7:30–7.
3. Barallobre-Barreiro J, Didangelos A,
Schoendube FA, et al. Proteomics
analysis of cardiac extracellular matrix
remodeling in a porcine model of
ischemia-reperfusion injury. Circula-
tion 2012;125:789 802.
4. Ho CY, Lopez B, Coelho-Filho OR,
et al. Myocardial fibrosis as an early
manifestation of hypertrophic cardio-
myopathy. N Engl J Med 2010;363:
552–63.
5. Bruder O, Wagner A, Jensen CJ, et
al. Myocardial scar visualized by cardio-
vascular magnetic resonance imaging pre-
dicts major adverse events in patients with
hypertrophic cardiomyopathy. J Am Coll
Cardiol 2010;56:875– 87.
6. Bruder O, Schneider S, Nothnagel
D, et al. EuroCMR (European Car-
diovascular Magnetic Resonance)
registry: results of the German pilot
phase. J Am Coll Cardiol 2009;54:
1457–66.
7. Rehwald WG, Fieno DS, Chen EL,
Kim RJ, Judd RM. Myocardial mag-
netic resonance imaging contrast
agent concentrations after reversible
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
9
and irreversible ischemic injury. Cir-
culation 2002;105:224 –9.
8. Kim RJ, Wu E, Rafael A, et al. The
use of contrast-enhanced magnetic
resonance imaging to identify revers-
ible myocardial dysfunction. N Engl
J Med 2000;343:1445–53.
9. Grebe O, Paetsch I , Kestler HA, et
al. Optimal acquisition parameters for
contrast enhanced magnetic resonance
imaging after chronic myocardial in-
farction. J Cardiovasc Magn Reson
2003;4:575–87.
10. Puntmann VO, Jahnke C, Gebker R,
et al. Usefulness of magnetic resonance
imaging to distinguish hypertensive and
hypertrophic cardiomyopathy. Am J Car-
diol 2010;106:1016 –22.
11. Mewton N, Liu CY, Croisille P,
Bluemke D, Lima JA. Assessment of
myocardial fibrosis with cardiovascular
magnetic resonance. J Am Coll Car-
diol 2011;57:891–903.
12. Iles L, Pfluger H, Phrommintikul A,
et al. Evaluation of diffuse myocardial
fibrosis in heart failure with cardiac
magnetic resonance contrast-enhanced
T1 mapping. J Am Coll Cardiol 2008;
52:157480.
13. Blume U, Lockie T, Stehning C, et al.
Interleaved T(1) and T(2) relaxation time
mapping for cardiac applications. J Magn
Reson Imaging 2009;29:480 –7.
14. Messroghli DR, Radjenovic A, Kozerke
S, Higgins DM, Sivananthan MU, Ridg-
way JP. Modified Look-Locker inversion
recovery (MOLLI) for high-resolution
T1 mapping of the heart. Magn Reson
Med 2004;52:141– 6.
15. Messroghli DR, Walters K, Plein S, et
al. Myocardial T1 mapping: applica-
tion to patients with acute and chronic
myocardial infarction. Magn Reson
Med 2007;58:34 40.
16. Piechnik SK, Ferreira VM, Dall’
Armellina E, et al. Shortened modi-
fied Look-Locker inversion recovery
(ShMOLLI) for clinical myocardial
T1-mapping at 1.5 and 3 T within a 9
heartbeat breathhold. J Cardiovasc
Magn Reson 2010;12:69.
17. Maron BJ, McKenna WJ, Danielson
GK, et al., for the American College of
Cardiology Foundation Task Force on
Clinical Expert Consensus Documents;
European Society of Cardiology Com-
mittee for Practice Guidelines. American
College of Cardiology/European Society
of Cardiology clinical expert consensus
document on hypertrophic cardiomyopa-
thy. A report of the American College of
Cardiology Foundation Task Force on
Clinical Expert Consensus Documents
and the Classification of Cardiomyopa-
thies. European Society of Cardiology
Committee for Practice Guidelines. J Am
Coll Cardiol 2003;42:1687–713.
18. Elliott P, Andersson B, Arbustini E,
et al. Classification of the cardiomy-
opathies: a position statement from
the European Society of Cardiology
Working Group on Myocardial and
Pericardial Diseases. Eur Heart J
2008;29:2706.
19. Natori S, Lai S, Finn JP, et al. Cardio-
vascular function in multi-ethnic study
of atherosclerosis: normal values by age,
sex, and ethnicity. AJR Am J Roent-
genol 2006;186 Suppl:S357– 65.
20. Kramer CM, Barkhausen J, Flamm
SD, Kim RJ, Nagel E, for the Society
for Cardiovascular Magnetic Reso-
nance Board of Trustees Task Force
on Standardized Protocol. Standard-
ized cardiovascular magnetic reso-
nance imaging (CMR) protocols, So-
ciety for Cardiovascular Magnetic
Resonance: Board of Trustees Task
Force on Standardized Protocols.
J Cardiovasc Magn Reson 2008;10:35.
21. Buerger C, Schaeffter T, King AP.
Hierarchical adaptive local affine reg-
istration for fast and robust respiratory
motion estimation. Med Image Anal
2011;15:551–64.
22. Karlsen OT, Verhagen R, Bovée
WMM. Parameter estimation from
Rician-distributed data sets using a
maximum likelihood estimator: Ap-
plication to t1 and perfusion measure-
ments. Magn Reson Med 1999;41:
614–23.
23. Schelbert EB, Testa SM, Meier CG,
et al. Myocardial extravascular extra-
cellular volume fraction measurement
by gadolinium cardiovascular mag-
netic resonance in humans: slow infu-
sion versus bolus. J Cardiovasc Magn
Reson 2011;13:16 –30.
24. Flett AS, Hayward MP, Ashworth
MT, et al. Equilibrium contrast car-
diovascular magnetic resonance for
the measurement of diffuse myocardial
fibrosis: preliminary validation in hu-
mans. Circulation 2010;122:138 44.
25. Ugander M, Oki AJ, Hsu LY, et al.
Extracellular volume imaging by mag-
netic resonance imaging provides in-
sights into overt and sub-clinical myo-
cardial pathology. Eur Heart J 2012;
33:1268–78.
26. Maceira AM, Joshi J, Prasad SK, et al.
Cardiovascular magnetic resonance in
cardiac amyloidosis. Circulation 2005;
111:186–93.
27. Jellis C, Wright J, Kennedy D, et al.
Association of imaging markers of
myocardial fibrosis with metabolic and
functional disturbances in early dia-
betic cardiomyopathy. Circ Cardio-
vasc Imaging 2011;4:693–702.
28. Martos R, Baugh J, Ledwidge M, et
al. Diastolic heart failure: evidence of
increased myocardial collagen turn-
over linked to diastolic dysfunction.
Circulation 2007;115:888 –95.
29. Sharma P, Socolow J, Patel S, Pettigrew
RI, Oshinski JN. Effect of Gd-DTPA-
BMA on blood and myocardial T1 at
1.5T and 3T in humans. J Magn Reson
Imaging 2006;23:323–30.
30. Knowles BR, Batchelor PG, Parish V,
et al. Pharmacokinetic modeling of
delayed gadolinium enhancement in
the myocardium. Magn Reson Med
2008;60:1524–30.
31. Anversa P, Hiler B, Ricci R, Guideri
G, Olivetti G. Myocyte cell loss and
myocyte hypertrophy in the aging rat
heart. J Am Coll Cardiol 1986;8:
1441–8.
32. Sievers B, Rehwald WG, Albert TS,
et al. Respiratory motion and cardiac
arrhythmia effects on diagnostic accu-
racy of myocardial delayed-enhanced
MR imaging in canines. Radiology
2008;247:106–14.
Key Words: diffuse fibrosis y
hypertrophic cardiomyopathy y
late gadolinium enhancement y
T1 mapping.
JACC: CARDIOVASCULAR IMAGING, VOL. xx, NO. x, 2013
MONTH 2013:xxx
Puntmann et al.
Native T1 Mapping and Diffuse Myocardial Fibrosis
10
    • "Our septal native T 1 values were increased, though not statistically , likely reflecting pathological remodelling including a degree of interstitial and replacement fibrosis, evident by cases with positive late enhancement and reduced RV ejection fraction. Quantitative T 1 -mapping, using native T 1 values, has been used for non-invasive tissue characterization in a variety of disease conditions[14][15][16]; however, its use has generally been restricted to the left ventricle due in part to the limitations of spatial resolution. T 1 -mapping sequences typically have an in-plane spatial resolution of approximately 2 mm, thus making it difficult to apply in the right ventricle, where, in normal hearts the mean diastolic wall thickness of the RV is 4–5 mm[30, 31], compared to 6–10 mm for the interventricular septum [31] . "
    [Show abstract] [Hide abstract] ABSTRACT: Aims: Anderson-Fabry disease (AFD) is characterized by progressive multiorgan accumulation of intracellular sphingolipids due to α-galactosidase A enzyme deficiency, resulting in progressive ventricular hypertrophy, heart failure, arrhythmias, and death. Decreased native (non-contrast) left ventricular (LV) T1 (longitudinal relaxation time) with MRI discriminates AFD from healthy controls or other presentations of concentric hypertrophy, but the right ventricle (RV) has not been studied. The aims of the current study were to evaluate native RV T1 values in AFD, with a goal of better understanding the pathophysiology of RV involvement. Methods and results: Native T1 values were measured in the inferior RV wall (RVI), interventricular septum (IVS), and inferior LV (LVI) in patients with AFD, patients with pulmonary hypertension, who provided an alternative RV pathological process for comparison, and healthy controls. A minimum wall thickness of 4 mm was selected to minimize partial volume errors in tissue T1 analysis. T1 analysis was performed in 6 subjects with AFD, 6 subjects with PH, and 21 controls. Native T1 values were shorter (adjusted p<0.05 for all comparisons), independent of location, in subjects with AFD (RVI-T1 = 1096±49 ms, IVS-T1 = 1053±41 ms, LVI-T1 = 1072±44 ms) compared to both PH (RVI-T1 = 1239±41 ms, IVS-T1 = 1280±123 ms, LVI-T1 = 1274±57 ms) and HC (IVS-T1 = 1180±60 ms, LVI-T1 = 1183±45 ms). RVI measurements were not possible in controls due to insufficient wall thickness. Conclusion: Native T1 values appear similarly reduced in the left and right ventricles of individuals with AFD and RV wall thickening, suggesting a common pathology. In contrast, individuals with PH and thickened RVs showed increased native T1 values in both ventricles, suggestive of fibrosis.
    Full-text · Article · Jun 2016
    • "The higher mean native myocardial T1 value in the DCM and HCM group can be explained by an increased interstitial space in DCM and HCM due to diffuse myocardial fibro- ses [8]. Concordant to our results Dass et al. and Puntmann et al. reported significantly higher T1 values in HCM and DCM compared to healthy individuals, even though a direct comparison of their values to ours is not possible since Dass et al. and Puntmann et al. used 3 T MR sys- tems [8,21]. In accordance to previous reports, we found a slightly higher mean native myocardial T1 value in the group of patients suffering from acute myocarditis compared to the group of healthy individuals, which can be explained by a diffuse myocardial edema in acute myocarditis [12]. "
    [Show abstract] [Hide abstract] ABSTRACT: Objectives T1 mapping allows quantitative myocardial assessment, but its value in clinical routine remains unclear. We investigated, whether the average native myocardial T1 value can be used as a diagnostic classifier between healthy and diffuse diseased myocardium. Methods Native T1 mapping was performed in 54 persons with healthy hearts and in 150 patients with diffuse myocardial pathologies (coronary artery disease (CAD): n = 76, acute myocarditis: n = 19, convalescent myocarditis: n = 26, hypertrophic cardiomyopathy (HCM): n = 12, dilated cardiomyopathy (DCM): n = 17) at 1.5 Tesla in a mid-ventricular short axis slice using a modified Look-Locker inversion recovery (MOLLI) sequence. The average native myocardial T1 value was measured using dedicated software for each patient. The mean as well as the range of the observed average T1 values were calculated for each group, and compared using t-test. The ability of T1 mapping to differentiate between healthy and diffuse diseased myocardium was assessed using receiver operating characteristic analysis (ROC). Results The mean T1 value of the group “healthy hearts” (955±34ms) differed significantly from that of the groups DCM (992±37ms, p<0.001), HCM (980±44ms, p = 0.035), and acute myocarditis (974±36ms, p = 0.044). No significant difference was observed between the groups “healthy hearts” and CAD (951±37ms, p = 0.453) or convalescent myocarditis (965±40ms, p = 0.240). The average native T1 value varied considerably within all groups (range: healthy hearts, 838-1018ms; DCM, 882-1034ms; HCM, 897-1043ms; acute myocarditis, 925-1025ms; CAD, 867-1082ms; convalescent myocarditis, 890-1071ms) and overlapped broadly between all groups. ROC analysis showed, that the average native T1 value does not allow for differentiating between healthy and diffuse diseased myocardium, except for the subgroup of DCM. Conclusions The average native T1 value in cardiac MR imaging does not allow differentiating between healthy and diffusely diseased myocardium in individual cases.
    Full-text · Article · May 2016
    • "One likely factor contributing to this difference is that unlike most prior studies that focused on patient cohorts with a specific and profound phenotype , we sampled patients with a range of cardiomyopathies and varying degrees of edema or fibrosis. Also, artifacts introduced by motion correction [18] have led many prior studies to restrict data analysis to the interventricular septum [9, 11, 20] . We analyzed myocardium across an entire short-axis slice, defining regions of interest based on LGE patterns. "
    Article · Jan 2016
Show more