Regional Myocardial Tissue Characterization by Cardiac Magnetic Resonance Is Associated With Heart Failure Development in Asymptomatic Transthyretin Amyloid Cardiomyopathy

Abstract

Background

Because the number of asymptomatic individuals diagnosed with transthyretin amyloid cardiomyopathy (ATTR-CM) has been rising, it is important to identify myocardial tissue characteristics associated with heart failure (HF) development in asymptomatic patients with ATTR-CM.

Methods and Results

Among 113 patients with ATTR-CM, 14 were asymptomatic (mean age 75±6 years; 79% male) and underwent echocardiography and cardiac magnetic resonance imaging (CMR). Over a median follow-up of 770 days, 6 patients developed HF. Compared with those who remained asymptomatic, these patients had a higher left ventricular mass index (LVMI) on echocardiography (P=0.020). On CMR, native T1 was significantly higher at the basal inferior (P=0.0277) and inferolateral (P=0.0017) segments, whereas the extracellular volume (ECV) fraction was significantly higher at the basal inferolateral (P=0.0296) and anterolateral (P=0.0092) segments. LVMI correlated with native T1 (r=0.599, P=0.024) and ECV at the basal inferolateral (r=0.780, P<0.001) and anterolateral (r=0.624, P=0.017) segments. Kaplan–Meier analysis showed that patients with higher LVMI, native T1 at the basal inferolateral segment, and ECV at the basal anterolateral segment had a higher incidence of HF development.

Conclusions

In asymptomatic ATTR-CM patients, HF development was associated with increased LVMI and regional myocardial tissue abnormalities in the basal lateral wall detected by native T1 and ECV on CMR; however, these findings should be considered exploratory and require validation in larger studies.

Central Figure.

Transthyretin amyloid cardiomyopathy (ATTR-CM) is a progressive form of heart failure (HF) caused by extracellular deposition of misfolded transthyretin protein within the myocardium.¹ It is characterized by increased left ventricular (LV) wall thickness, reduced LV cavity size, atrial enlargement, elevated LV filling pressures, and preserved or mildly reduced LV ejection fraction (LVEF). Disease-modifying therapies provide the greatest benefit in patients with New York Heart Association (NYHA) functional class I–II, rather than for those with advanced symptoms (class III–IV),^2,3 highlighting the importance of early diagnosis to improve outcomes.^4–6 In clinical practice, however, treatment is generally initiated only in patients with established HF symptoms, typically those with prior hospitalization for HF or those requiring diuretics for congestion.^7,8 Meanwhile, the number of asymptomatic individuals diagnosed with ATTR-CM has been rising,⁹ and these patients are often ineligible for current therapies yet may still undergo gradual disease progression, creating a critical treatment gap and underscoring the need for strategies to detect cardiac involvement before overt HF develops.

Imaging modalities play a pivotal role in identifying early myocardial changes. Cardiac magnetic resonance imaging (CMR) provides high-resolution morphological and functional assessment alongside detailed tissue characterization.¹⁰ Advanced techniques such as native T1 mapping and extracellular volume (ECV) quantification enable sensitive detection of regional amyloid infiltration before and after tafamidis treatment.^11,12 Our prior work revealed a heterogeneous distribution pattern, with preferential involvement of the basal inferoseptal regions even in asymptomatic patients.¹³ Building on these observations, we hypothesized that regional myocardial tissue characteristics assessed by CMR may help identify myocardial involvement associated with subsequent HF development in patients with asymptomatic ATTR-CM. Therefore, in this study we aimed to evaluate the LV tissue characterization parameters by CMR in relation to subsequent HF development.

Methods

Study Population

This retrospective, cohort study enrolled 113 patients with ATTR-CM at the University of Miyazaki Hospital and Miyazaki Medical Association Hospital between September 2018 and April 2025. All participants were of Japanese ethnicity. Asymptomatic patients were defined as those without a history of HF symptoms, classified as NYHA functional class I, and with no prior use of diuretics for volume overload. They demonstrated histologically confirmed amyloid infiltration in either the endomyocardium or carpal synovium and exhibited grade 2–3 myocardial uptake on technetium-99 m pyrophosphate (^{99 m}Tc-PYP) scintigraphy.¹⁴ None met the current criteria for initiation of disease-modifying therapy. Exclusion criteria included prior myocardial infarction, the presence of monoclonal gammopathy, and CMR performed at other institutions with unavailable 18-segmental measurements for native T1 mapping. Baseline assessment included comorbidities, extracardiac manifestations of amyloidosis, and relevant laboratory data parameters (B-type natriuretic peptide [BNP], troponin-T, and estimated glomerular filtration rate [eGFR]), obtained through physical examination and review of medical records, as previously described.¹³ Blood pressure was measured while in the seated position during hospital visits, and medication use was documented at baseline.

The study was conducted in accordance with the Declaration of Helsinki, and all patients provided written informed consent for the anonymous use of their clinical data. The protocol was approved by the Institutional Review Board of the University of Miyazaki (protocol number: 0-0651).

12-Lead ECG

The ECG parameters were assessed according to standard definitions.^15,16 Low QRS voltage was defined as ≤5 mm (0.5 mV) in all peripheral leads or ≤10 mm (1.0 mV) in all precordial leads. A pseudoinfarction pattern was defined as pathological Q waves (≥25% of the R-wave amplitude) or QS waves in 2 contiguous leads in the absence of documented ischemic heart disease. Atrioventricular (AV) conduction delay was defined as any degree of AV block. Intraventricular conduction delay was categorized as complete (QRS ≥120 ms) or incomplete (QRS 100–119 ms) right or left bundle branch block.

Radionuclide Scintigraphy

Radionuclide scintigraphy was performed 3 h after intravenous administration of 555–740 MBq of ^{99 m}Tc-PYP, followed by single-photon emission computed tomography and a low-dose, non-contrast CT scan for anatomical reference. Cardiac uptake was visually graded using the Perugini scale (0–3).¹⁷ All images were independently evaluated by 2 experienced radiologists (T. Terada, M.K.) who were blinded to the clinical data.

Conventional and 2D Speckle-Tracking Echocardiography

Echocardiography using the EPIQ CVx system (Philips, Amsterdam, Netherlands) equipped with a 1–5 MHz transducer was performed at rest with patients in the left lateral decubitus position.¹³ Image acquisition followed the guidelines of the American Society of Echocardiography and the European Association of Cardiovascular Imaging.^18,19 Measurements included LV systolic and diastolic dimensions and wall thickness, obtained from 2D images. LVEF was calculated using the modified Simpson biplane method from apical 2- and 4-chamber views. LV mass index (LVMI) was calculated from LV dimensions and wall thickness using the linear method of Devereux et al.²⁰ For strain analysis, apical 4-, 3-, and 2-chamber views were acquired and stored as cine loops over 2 cardiac cycles, with frame rates of 40–80 frames/s. Longitudinal strain was quantified using the AutoStrain LV/RV/LA software on the Philips EPIQ CVx system.¹³ Strain values were expressed as absolute percentages and analyzed at the basal, midventricular, and apical levels according to the 18-segment model displayed in a bull’s-eye plot.^21,22

CMR Acquisition

CMR examinations were performed at 2 institutions using the same field strength, acquisition sequence, and similar scanner platforms (3.0 Tesla whole-body scanner (Ingenia 3.0T CX [n=11] or Ingenia Elition 3.0T [n=3], Philips, The Netherlands)).¹³ Patients were imaged while in the supine position with anterior and posterior phased-array surface coils for signal reception. Initial localization was obtained using breath-hold real-time and steady-state free precession (SSFP) sequences to define the cardiac axes. LV function was assessed with cine imaging using a segmented SSFP pulse sequence, acquired in short-axis, long-axis, and 4-chamber views to encompass the entire LV. Imaging parameters included an 8-mm slice thickness, breath-hold acquisition, field-of-view 380 mm², matrix size 176×223, repetition time 3.0 ms, echo time 1.48 ms, and flip angle 50°.

Native T1 Mapping and ECV

Native T1 mapping was performed to quantify intrinsic myocardial tissue characteristics and assess extracellular space expansion. Analyses were conducted in a blinded fashion with respect to late gadolinium enhancement (LGE) findings.²³ LGE imaging was acquired 10–15 min after intravenous administration of gadobutrol (0.1 mmol/kg body weight). Native T1 mapping was performed during mid-to-end diastole using a single-breath-hold shortened modified Look-Locker inversion recovery sequence.²⁴ Imaging parameters included a 10-mm slice thickness, field-of-view 380 mm², matrix size 144×142, repetition time 2.0 ms, echo time 0.95 ms, and flip angle 20°. For ECV quantification, pre- and postcontrast T1 maps of the myocardium and blood pool were acquired at the same slice location. ECV values were calculated using the standard formula,²⁵ incorporating hematocrit measured on the same day as the CMR examination.

CMR Analysis

CMR analysis was performed using commercially available software (Vitrea, VMS-001SA V8.10, Canon Medical Systems). Blood pool values were obtained from regions of interest (ROIs) carefully placed in the LV cavity, avoiding artifacts, trabeculation, and partial volume effects. For regional analysis, native T1 mapping and ECV values were measured segmentally in 6 equiangular segments per slice at the basal, midventricular, and apical levels. Segmental analysis and bull’s-eye plots were generated using IntelliSpace Portal version 10.1 (Philips). Whole-slice native T1 and ECV values were defined as the mean across all 6 segments of the respective slice. Native T1 and ECV values for the midventricular septum were calculated as the average of the anteroseptal and inferoseptal segments on the midventricular short-axis slice. Institutional reference values for native T1 in healthy volunteers scanned at the midventricular septum were comparable between institutions; 1,262±38 ms on the Ingenia CX 3.0T system (n=23; male/female=14/9; mean age 42±7 years) and 1,262±27 ms on the Ingenia Elition X 3.0T system (n=17; male/female=11/6; mean age 40±12 years). In an additional cohort of normal individuals scanned using the Ingenia CX 3.0T system (n=20; male/female=15/5; mean age 61±16 years), and reference values for the midventricular septum were 1,266±30 ms for native T1 and 27±2% for ECV.¹³ Therefore, absolute native T1 values were used for analysis rather than z-score normalization.²³

Outcomes

The primary outcome was the occurrence of new-onset HF events, defined as progression to NYHA functional class II or higher, accompanied by ≥30% relative increase in BNP levels from baseline,^9,26,27 and the initiation of diuretic therapy (i.e., specifically loop diuretics, mineralocorticoid-receptor antagonists, or sodium–glucose cotransporter 2 inhibitors [SGLT2i]) for congestion management. HF symptoms were defined as the presence of symptoms and/or signs of congestion, such as dyspnea on exertion, orthopnea or peripheral edema, and/or symptoms suggestive of low cardiac output, including fatigue or reduced exercise tolerance.²⁶ The onset of HF was based on clinical symptoms and BNP elevation at either scheduled visits or unscheduled evaluations prompted by symptom worsening. In addition, patients without overt symptoms but with objective evidence of elevated intracardiac filling pressures and BNP levels, judged clinically as HF by their treating physicians, were also considered to have a HF event. Follow-up time was calculated from the date of the initial evaluation at our institution.

Statistical Analysis

Continuous variables are expressed as mean±standard deviation or as median with interquartile range (IQR) (25–75th percentiles), depending on distribution, which was tested using the Shapiro–Wilk method. Categorical variables are reported as frequencies (counts) and percentages. Between-group comparisons were conducted using Student’s t-test for normally distributed continuous variables and the Mann–Whitney U test for non-normally distributed data. Categorical variables were compared using Pearson’s chi-squared test. The association between echocardiographic LVMI and MRI-derived parameters (native T1 and ECV) across 18 myocardial segments was examined using Pearson’s correlation coefficient (r). Receiver operating characteristic (ROC) analysis was used to determine the predictive performance of LVMI, native T1 mapping, and ECV for new-onset HF. Time-to-event analysis for HF development was performed using the Kaplan–Meier method. Patients were stratified according to imaging parameters, and differences between groups were compared using the log-rank test. These analyses were considered exploratory. Interobserver agreement for native T1 measurements was assessed using the intraclass correlation coefficient (ICC) based on a 2-way model with absolute agreement. The single-measure ICC and 95% confidence interval were calculated. Interobserver agreement for native T1 measurements was further evaluated using Bland–Altman analysis. For each myocardial segment, the difference between 2 observers’ measurements was plotted against the mean of the 2 measurements. The mean difference (bias) and the 95% limits of agreement were calculated as the mean difference±1.96 standard deviations. A 2-tailed P value <0.05 was considered statistically significant. All analyses and visualizations were conducted using SPSS version 30 (IBM Corporation, Armonk, NY, USA) and GraphPad Prism version 8.0 (GraphPad Software, La Jolla, CA, USA).

Results

Baseline Characteristics

Of the 113 patients with ATTR-CM, 15 (13%) were asymptomatic at enrollment. Genetic testing for TTR identified 13 wild-type and 2 variant-type (Val122Ile) cases. Of them, 14 patients (12 wild-type and 2 variant-type) were included in the subsequent analyses (Figure 1). During a median follow-up of 770 days, new-onset HF occurred in 6 of 14 patients. The time to HF onset ranged from 84 to 896 days after the initial evaluation (Figure 2): 3 developed de novo HF, 1 experienced worsening HF due to atrial tachycardia, 1 due to sick sinus syndrome, and 1 was asymptomatic but had elevated intracardiac filling pressures. The median increase in BNP from baseline to HF symptom onset was 59.6% (IQR, 39.6–85.4%) among patients initiating diuretics and subsequently starting disease-modifying therapies. The baseline characteristics of the patients are shown in Table 1. BNP and troponin-T levels did not distinguish between patients who developed HF and those who did not. ECG findings, including low-voltage and pseudoinfarction patterns, showed minimal discriminatory value for predicting HF development. On conventional 2D echocardiography, the LV interventricular septal thickness and LVMI were higher in patients who developed HF than in those who remained asymptomatic. In contrast, global and regional longitudinal strain parameters derived from speckle-tracking echocardiography did not differ significantly between groups. CMR demonstrated a trend toward elevated native T1 mapping and ECV values in the mid-septal wall, as well as higher whole-myocardial native T1 mapping and ECV values in the basal level, among patients who developed HF; however, these differences did not reach statistical significance.

Figure 1.

Patient recruitment flowchart. ATTR-CM, transthyretin amyloid cardiomyopathy; CMR, cardiac magnetic resonance imaging; HF, heart failure.

Figure 2.

Time to heart failure (HF) development in asymptomatic ATTR-CM. Kaplan–Meier curve showing the cumulative incidence of HF during follow-up. HF events were based on clinical symptoms and BNP elevation requiring initiation of HF therapy. Tick marks indicate censored observations.

Table 1.

Patients’ Baseline Characteristics

	All patients (n=14)	Without HF development (n=8)	New-onset HF (n=6)	P value
Age (years)	75±6 (61–84)	76±5	74±7	0.480
Sex (male/female)	11/3	6/2	5/1	0.707
BMI (kg/m2)	23±4 (16–31)	23±2	23±5	0.993
SBP (mmHg)	133±14 (110–155)	128±16	139±10	0.167
DBP (mmHg)	80±9 (61–93)	76±9	85±6	0.064
PR (/min)	74±13 (57–103)	76±16	72±10	0.542
Comorbidities
Hypertension, n (%)	11 (79)	5 (63)	6 (100)	0.091
Dyslipidemia, n (%)	5 (36)	4 (50)	1 (17)	0.198
Diabetes mellitus, n (%)	3 (21)	3 (38)	0 (0)	0.091
Carpal tunnel syndrome, n (%)	10 (71)	6 (43)	4 (67)	0.733
Coronary intervention (yes/no/not tested)	1/10/3	0/5/3	1/5/0	0.072
99 mTc-PYP (grade 1/2/3), n	0/5/7	0/4/3	0/1/4	0.198
Biopsy site (heart/carpal tunnel), n	11/3	6/2	5/1	0.707
Wild-type/variant-type, n	12/2	7/1	5/1	0.825
NYHA functional class (I/II/III/IV)	14/0/0/0	8/0/0/0	6/0/0/0
NAC staging (Ia/Ib/II), n	9/4/1	5/2/1	4/2/0	0.656
BNP (pg/mL)	89±59 (12–205)	71±49	113±66	0.194
Troponin-T (ng/mL)	0.031±0.019 (0.009–0.081)	0.030±0.023	0.033±0.012	0.768
eGFR (ml/min/1.73 m2)	55 [50, 70]	50 [49, 77]	58 [53, 67]	0.491
ECG
AF/AFL, n (%)	0 (0)	0 (0)	0 (0)
Low voltage, n (%)	1 (7)	1 (13)	0 (0)	0.369
First atrioventricular block, n (%)	4 (29)	3 (38)	1 (17)	0.394
LBBB, n (%)	0 (0)	0 (0)	0 (0)
RBBB, n (%)	3 (21)	1 (13)	2 (33)	0.259
IRBBB, n (%)	1 (7)	1 (13)	0 (0)	0.352
Pseudoinfarct pattern, n (%)	4 (29)	2 (25)	2 (33)	0.417
Pacing rhythm, n (%)	1 (7)	0 (0)	1 (17)	0.231
Medications
Loop diuretic, n (%)	0 (0)	0 (0)	0 (0)
MRA, n (%)	0 (0)	0 (0)	0 (0)
β-blocker, n (%)	1 (7)	0 (0)	1 (17)	0.231
ACEi/ARB, n (%)	6 (43)	3 (38)	3 (50)	0.640
SGLT2i, n (%)	1 (7)	1 (13)	0 (0)	0.369
Echocardiography
LVEF (%)	55±7 (42–63)	57±8	53±9	0.318
IVSTd (mm)	14±3 (9–18)	12±3	16±2	0.021
LVWTd (mm)	13±3 (9–18)	12±3	15±2	0.128
LVDd (mm)	42±7 (29–61)	42±9	42±5	0.945
LVMI (g/m2)	131±32 (82–193)	115±19	153±33	0.020
GLS	14±3 (8–17)	14±2	13±3	0.335
Regional LS
Base	11±4 (7–18)	11±4	10±3	0.812
Mid-ventricle	12±4 (4–19)	13±4	10±4	0.232
Apex	19±5 (8–26)	19±4	18±7	0.886
CMR
LVEF (%)	63±11 (47–82)	62±8	63±14	0.902
CI (L/min/m2)	2.4±0.9 (0.9–4.0)	2.5±0.9	2.3±1.1	0.762
LVEDV (mL/m2)	59±18 (29–107)	61±21	57±15	0.671
LVESV (mL/m2)	21 [13, 25]	21 [14, 26]	23 [12, 26]	0.897
Native T1 (ms) at midventricular septum	1,376±67 (1,268–1,469)	1,370±65	1,384±74	0.721
ECV (%) at midventricular septum	41±8 (31–54)	39±8	44±9	0.224
Whole native T1 mapping (ms)
Base	1,375±45 (1,292–1,454)	1,367±44	1,386±48	0.460
Mid-ventricle	1,362 [1,307, 1,386]	1,373 [1,316, 1,400]	1,356 [1,307, 1,371]	0.755
Apex	1,372±67 (1,234–1,475)	1,360±69	1,390±65	0.425
Whole ECV (%)
Base	42±7 (30–52)	39±7	45±5	0.094
Mid-ventricle	37±6 (27–46)	35±6	39±7	0.284
Apex	37±6 (29–45)	36±6	39±6	0.329

Continuous variables are expressed as mean±standard deviation (minimum–maximum) or median [25th, 75th percentile], and categorical variables are expressed as number (percentage). Comparisons between groups were performed using Student’s t-test or Mann–Whitney U test for continuous variables and Pearson’s chi-squared test for categorical variables. ^{99 m}Tc-PYP, ^{99 m}Tc-phyrophosphate; ACEi, angiotensin-converting enzyme inhibitor; Af/AFL, atrial fibrillation/atrial flutter; ARB, angiotensin II receptor blocker; BNP, B-type natriuretic peptide; CI, cardiac index; DBP, diastolic blood pressure; ECV, extracellular volume; eGFR, estimated glomerular filtration rate; IBBB, incomplete bundle branch block; IVSTd, interventricular septal thickness at end diastole; LBBB, left bundle branch block; LS, longitudinal strain (absolute value); LVDd, left ventricular dimension at end diastole; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVMI, left ventricular mass index; LVPWTd, left ventricular posterior wall thickness at end diastole; MRA, mineralocorticoid-receptor antagonist; NAC, National Amyloidosis Center; NYHA, New York Heart Association; RBBB, right bundle branch block; SBP, systolic blood pressure; SGLT2i, sodium–glucose cotransporter 2 inhibitor.

Segmental Native T1 Mapping and ECV Parameters

Native T1 values were significantly higher in the basal inferior and inferolateral segments in patients who developed HF compared with those who remained asymptomatic (P=0.0277 and 0.0017, respectively; Figure 3A). No significant differences were observed in the other basal segments. Similarly, native T1 values in the midventricular and apical slices did not differ between groups (Figure 3B–C). ECV values were significantly higher in the basal inferolateral (P=0.0296) and anterolateral (P=0.0092) segments in patients who developed HF (Figure 3D). At the midventricular and apical levels, no significant between-group differences were observed (Figure 3E,F).

Figure 3.

Native T1 mapping and extracellular volume (ECV) fraction for 6 myocardial segments at the basal (A,D), midventricular (B,E), and apical (C,F) levels in patients who remained event-free (n=8, red) and those who developed heart failure (HF) (n=6, blue). Data are shown as dot plots (mean±standard deviation). Student’s t-test was used for statistical comparisons. A, anterior; AS, anteroseptal; IS, inferoseptal; I, inferior; IL, inferolateral; AL, anterolateral.

Correlation Between Echocardiographic LVMI and CMR Parameters

LVMI showed a significant positive correlation with native T1 values in the basal inferolateral segment (r=0.599, P=0.024; Figure 4A), but no significant associations were observed in the other myocardial segments (Supplementary Figure 1). Stronger correlations were identified between LVMI and ECV: LVMI significantly correlated with ECV in the basal inferolateral (r=0.780, P<0.001) and basal anterolateral (r=0.624, P=0.017) segments (Figure 4B,C). No significant relationships were detected in the remaining segments (Supplementary Figure 2).

Figure 4.

Correlations between left ventricular mass index (LVMI) measured by conventional echocardiography and native T1 or extracellular volume (ECV) values assessed by cardiac magnetic resonance imaging at corresponding sites: (A) native T1 at the basal inferolateral segment; (B) ECV at the basal inferolateral segment; (C) ECV at the basal anterolateral segment. Blue circles denote patients who developed new-onset heart failure (HF), and red circles denote patients who remained asymptomatic

Exploratory ROC Analysis of LVMI and Segmental CMR Parameters for New-Onset HF

ROC analyses were performed to evaluate the ability of LVMI, basal inferolateral and anterolateral native T1, and ECV to discriminate patients who developed new-onset HF (Figure 5A–C). The AUC values were 0.854 for LVMI, 0.969 for basal inferolateral native T1, and 0.885 for basal anterolateral ECV. The optimal cutoff values identified in this cohort were 142.1 g/m² for LVMI, 1,404 ms for basal inferolateral native T1, and 34.0% for basal anterolateral ECV, with corresponding sensitivity and specificity as shown in Figure 5. Kaplan–Meier analysis using cohort-derived cutoff values demonstrated that higher LVMI, native T1, and ECV values had a higher incidence of HF development (Figure 6A–C).

Figure 5.

Receiver operating characteristic (ROC) analyses for predicting new-onset heart failure: (A) left ventricular mass index (LVMI); (B) basal inferolateral (IL) native T1 mapping; (C) basal anterolateral extracellular volume (ECV). AUC, area under the curve; CI, confidence interval.

Figure 6.

Kaplan–Meier curves for event-free survival from new-onset heart failure (HF): (A) left ventricular mass index (LVMI); (B) native T1 mapping at the basal inferolateral segment; (C) extracellular volume (ECV) fraction at the basal anterolateral (AL) segment.

Reproducibility of CMR Native T1 Measurements

Interobserver agreement for segmental native T1 measurements was good to excellent across all myocardial segments, with ICC values ranging from 0.777 to 0.937 (Table 2). In addition to ICC analysis, Bland–Altman analysis demonstrated good agreement between observers, with a small mean difference and acceptable limits of agreement (Supplementary Figure 3).

Table 2.

Interobserver Agreement for Segmental Native T1 Measurements

Section	Site	ICC (2, 1)	95% confidence interval
Base	A	0.913	0.645–0.974
	AS	0.937	0.820–0.979
	IS	0.871	0.653–0.956
	I	0.932	0.803–0.978
	IL	0.854	0.618–0.950
	AL	0.841	0.587–0.945
Mid-ventricle	A	0.777	0.454–0.922
	AS	0.834	0.557–0.944
	IS	0.920	0.769–0.973
	I	0.895	0.706–0.965
	IL	0.835	0.572–0.943
	AL	0.870	0.653–0.956
Apex	A	0.911	0.745–0.971
	AS	0.874	0.652–0.958
	IS	0.890	0.697–0.963
	I	0.778	0.406–0.925
	IL	0.812	0.506–0.936
	AL	0.796	0.479–0.929

ICC, intraclass correlation coefficient; A, anterior; AS, anteroseptal; IS, inferoseptal; I, inferior; IL, inferolateral; AL, anterolateral.

Discussion

The mechanisms driving the transition from asymptomatic to symptomatic ATTR-CM remain incompletely understood. In our study, earlier onset of HF was associated with higher LVMI in asymptomatic patients, and segmental myocardial tissue characterization provided additional mechanistic insights into HF development in this population.

The number of asymptomatic individuals diagnosed with ATTR-CM has increased over time,⁹ although precise incidence data remain limited. In our cohort, 13% of patients were asymptomatic at enrollment. The LV interventricular septal thickness (16.0 mm) represents an early, quantifiable marker of adverse prognosis before symptom onset.⁹ Our findings aligned with this threshold, as LVMI >142.1 g/m² corresponded to approximately 16.5 mm in septal thickness. In conventional CMR analysis, a single ROI is often drawn in the midventricular septum to represent the entire LV.²⁸ However, our study demonstrated that amyloid deposition is not always homogeneous, and relying solely on mid-septal ROIs may underestimate myocardial involvement. Our previous studies have suggested that functional impairment initially appears in the basal inferoseptal segment, and myocardial involvement may extend to the inferior and lateral segments as the disease progresses.^13,29,30 Usuku et al. also reported that the basal inferior segment seemed particularly vulnerable to longitudinal strain impairment in ATTR-CM.³¹ An interesting finding of the present study was the slight difference in the segments associated with HF development between native T1 and ECV. Native T1 values were associated with HF development mainly in the inferior and inferolateral segments, whereas ECV was associated with HF development in the inferolateral and anterolateral segments. This difference may reflect the different pathological information provided by native T1 and ECV. Native T1 reflects both intracellular and extracellular changes and may detect earlier myocardial involvement, whereas ECV more directly reflects extracellular amyloid deposition.^11,32 Therefore, myocardial involvement may initially be detected by native T1 in the inferior and inferolateral segments, while increased ECV in the anterolateral segment may indicate more advanced amyloid deposition associated with progression to symptomatic HF. Amyloid infiltration in this region increases LV mass and associated compressive forces, which may impair the subendocardial coronary reserve,³³ and contribute directly to elevated LV filling pressures and clinical decompensation.

In our cohort, 42.9% of patients developed HF during a median follow-up of 770 days (2.1 years), comparable with prior studies reporting that 54.3% of asymptomatic ATTR-CM patients with grade 2–3 bone scintigraphy developed HF by 3 years,⁹ and 32.2% (38/118) of asymptomatic patients developed HF over a median follow-up of 3.7 years.³⁴ One patient in the non-HF group was receiving a SGLT2i for diabetes mellitus at enrolment. Given that SGLT2i therapy has demonstrated benefits in HF management for ATTR-CM,³⁵ it may have influenced the outcome. Early-stage cardiac amyloidosis remains challenging to diagnose due to the absence of precise criteria.³⁶ Initiation of disease-modifying therapy in asymptomatic patients has been associated with improved survival.³⁴ Clinically, segmental native T1 and ECV measurements in asymptomatic ATTR-CM may help identify patients who may require closer monitoring. However, we acknowledged that the segmental findings in this study should be interpreted in the context of disease biology rather than as definitive statistical localization, and further validation in larger studies is required.

Study Limitations

First, the sample size was small, with only 6 HF events during follow-up, precluding reliable multivariable analysis and limiting assessment of the independent or incremental value of segmental CMR parameters beyond conventional markers such as LVMI or BNP. Second, multiple segmental comparisons were performed across 18 myocardial segments for both native T1 and ECV, increasing the risk of type I error. Given the exploratory nature of this study, no formal correction for multiple testing was applied; therefore, segmental findings should be considered hypothesis-generating. Third, although BNP elevation was confirmed in all patients at the time of HF events, the diagnosis of HF and initiation of diuretic therapy were determined by treating physicians without central adjudication, which may have introduced some degree of subjectivity. Fourth, native T1 values are influenced by scanner type, field strength, and acquisition protocols.^23,37 Although native T1 values were comparable between institutions in this study, residual inter-scanner variability cannot be completely excluded, and normalization approaches such as z-scores may be useful in larger multicenter studies.²³ Finally, accurate quantification of native T1 values in the basal lateral wall is technically challenging because of motion and susceptibility artifacts related to adjacent lung tissue.³⁸ In this study, Bland–Altman analysis demonstrated slightly greater interobserver variability in the basal inferolateral and anterolateral segments, indicating that measurements in the lateral wall may be more susceptible to variability and should be interpreted with caution.

Conclusions

In asymptomatic ATTR-CM patients, HF development was associated with increased LVMI and regional myocardial tissue abnormalities detected by native T1 and ECV on CMR. In particular, involvement of the basal lateral wall may represent a transitional stage from asymptomatic disease to symptomatic HF. These findings should be considered exploratory and require validation in larger prospective studies.

Disclosure

T. Tsuruda received lecture fees from Pfizer Japan Inc. K.K. is a member of Circulation Reports’ Editorial Team.

IRB Information

The protocol was approved by the Institutional Review Board of the University of Miyazaki (protocol number: 0-0651).

Author Contributions

T. Tsuruda designed the study, collected and analyzed data, performed statistical analyses, and edited the manuscript. H.N., and Keisuke Y. collected and analyzed the CMR data. M.O., and M.T. performed conventional and speckle-tracking echocardiography. T. Terada and M.K. analyzed the ^{99 m}Tc-PYP scintigraphy images. T.O. collected synovial samples and contributed to data analysis. A.Y. performed histological assessments. Y. Suiko, Y.M., S.K., K.M., H.T., M.Y., Kinuko Y., T.I., K.N., Y. Shibata, and K.K. provided critical revisions and editorial input. T. Tsuruda was primarily responsible for the final content of the manuscript. All authors read and approved the final manuscript.

Supplementary Files

Supplementary File 1

Supplementary Figure 1. Supplementary Figure 2. Supplementary Figure 3.

Funding Statement

Funding: None.

Data Availability

All data generated or analyzed during this study are included in this article and its supplementary materials. Further inquiries can be directed to the corresponding author.