Novel CMR-Derived ECV and LA Reservoir Strain Interaction Provides Independent Risk Stratification in Non-Ischemic Cardiomyopathy

Thursday, February 5, 2026

4:13 PM - 4:20 PM

Location: Galapagos / Aruba / Ilhabela (2nd Lower Level)

Presenting Author(s)

Maria-Luiza Luchian, MD, PhD

Post-doctoral Research Fellow
Cleveland Clinic
Cleveland, Ohio, United States

Primary Author(s)

Maria-Luiza Luchian, MD, PhD

Post-doctoral Research Fellow
Cleveland Clinic
Cleveland, Ohio, United States

Co-Author(s)

TK

Theerawat Korkerdsup, MD

Research fellow
Cleveland Clinic, United States
AY

Ahmad Younis, MD

Post-doctoral Research Fellow
Cleveland Clinic, United States
CA

Carl Ammoury, MD

Research Fellow
Cleveland Clinic Foundation
Cleveland, Ohio, United States
MF

Mera Fahim, MD, PhD

Post-doctoral Research Fellow
Cleveland Clinic, United States
XS

Xiaotan Sun, MSc

Research Data Scientist I
Cleveland Clinic
Beachwood, Ohio, United States
MN

Makiya Nakashima, MSc

Research Data Scientist II
Cleveland Clinic
Cleveland, Ohio, United States
David Chen, PhD

Director of Artificial Intelligence
Cleveland Clinic
Cleveland, Ohio, United States
Christopher Nguyen, PhD, FSCMR

Director, Cardiovascular Innovation Research Center
Cleveland Clinic
Cleveland, Ohio, United States
WT

Wilson Tang, MD

Cardiologist
Cleveland Clinic, United States
Deborah Kwon, MD, FSCMR

Director of Cardiac MRI
Cleveland Clinic
Cleveland, Ohio, United States

Background: In NICM, LGE is a well-established prognostic marker; yet in some patients the LGE burden is low or absent,suggesting a lower-risk profile,but without eliminating residual risk.Extracellular volume(ECV) quantifies myocardial interstitial expansion that LGE may not detect, and impaired left-atrial(LA) mechanics has been shown to provide independent prognostic value.We therefore hypothesized that ECV and LA reservoir strain(LASr) offer complementary and potentially interactive risk stratification, particularly when LGE burden is low.¹²³

Methods: We retrospectively analyzed consecutive NICM patients(2013–2017) who underwent CMR on Philips Achieva 1.5T and Ingenia 3.0T scanners.LA strain was measured by feature-tracking (CVI-42 software, Canada) on 2/4-chamber cine;ECV by MOLLI T1 mapping pre/post contrast.Primary endpoint:time from index CMR to first HF hospitalization, LVAD, transplant or death. Given events per variable constraints, the 4-variable Cox models included one LA strain metric (LASr,LAScd,LASb per +1%),ECV(per+5%),MR fraction(per +1%),and a medical risk score(z-scored linear predictor of age,sex,hypertension,diabetes,BMI).Secondary analyses contrasted ECV≥30%vs< 30% with stabilized IPTW with doubly robust Cox model and tested prespecified ECV×LA interaction.

Results: Of 203 patients,160 with complete LA strain were analyzed (35 events;follow-up 3.7±1.7 years;age 54.9±15.1; 62(39%) female; 92(58%) LGE).Means:LVEF32.5±10.6%,LGE 3.4±6.7%,ECV 33.1±5.8%,LASr 24.9±11.6%,LASb 12.4±6.6%,LAScd 12.7±7.6%,MR fraction 13.4±11.3%(Table 1).The LASr model performed best (AIC 293.09;C-index 0.758) vs LAScd (AIC 296.08;C-index 0.751) and LASb (AIC 295.45;C-index 0.746).In the LASr model, higher LASr was protective (HR0.94,p=0.003),ECV adverse (HR1.36,p=0.033) and MR fraction showed an inverse association (HR0.94,p=0.008) limited to MR< 30% (piecewise HR 0.77 per +5%);no dose–response for MR≥30%(LR p=0.44).The medical risk score was associated with higher risk (HR1.57 per +1 SD;p=0.006, Fig1A). After adjustment for ECV,LASr and the risk score, LGE was not independently associated (p=0.145).Models substituting LAScd or LASb showed similar protective effects (each HR 0.92, p< 0.001) with ECV remaining adverse (p< 0.01).In IPTW analyses, ECV≥30% was associated with higher absolute event risk and LASr remained protective (HR0.96, p=0.001).Interaction testing showed ECV×LA effect modification for LASr (HR0.97, p=0.049) and LAScd (HR0.94, p=0.009),but not LASb (p=0.46,Fig1B).Cross-sectionally,higher ECV correlated with worse LA mechanics (all p< 0.011)(Fig1C).Median-defined ECV×LASr strata separated clearly(log-rank p=0.001) worst survival in high-ECV/low-LASr(Fig2).

Conclusion: In NICM with limited/absent LGE,residual risk is best captured by the ECV–LASr interplay:preserved LASr blunts ECV-related risk,whereas high ECV plus reduced LASr identifies a high-risk phenotype warranting closer surveillance.Combined LASr–ECV assessment improves risk stratification and may guide tailored management.

Group characteristics based on ECV and LA reservoir interaction

Variable	N = 160		Low ECV + High LASR N = 25	Low ECV + Low LAsR N = 17	High ECV + High LAsR N = 55	High ECV + Low LAsR N = 63	p-value
Age (years)		54.8 ± 15.1	57.9 ± 17.0	52.9 ± 13.8	54.0 ± 15.1	54.9 ± 14.8	0.386
Female		62 / 160 (38.8%)	14 / 25 (56%)	5 / 17 (29%)	20 / 55 (36%)	23 / 63 (37%)	0.277
Hypertension		100 / 160 (63%)	16 / 25 (64%)	10 / 17 (59%)	36 / 55 (65%)	38 / 63 (60%)	0.922
Diabetes		32 / 160 (20%)	4 / 25 (16%)	4 / 17 (24%)	12 / 55 (22%)	12 / 63 (19%)	0.916
Hyperlipidemia		68 / 160 (43%)	11 / 25 (44%)	9 / 17 (53%)	23 / 55 (42%)	25 / 63 (40%)	0.815
Atrial Fibrillation		29 / 160 (18%)	4 / 25 (16%)	3 / 17 (18%)	10 / 55 (18%)	12 / 63 (19%)	0.999
Beta blockers		138 / 160 (86%)	20 / 25 (80%)	16 / 17 (94%)	46 / 55 (84%)	56 / 63 (89%)	0.516
ACEI/ARB		123 / 160 (77%)	21 / 25 (84%)	14 / 17 (82%)	41 / 55 (75%)	47 / 63 (75%)	0.785
aldosterone		59 / 160 (37%)	11 / 25 (44%)	6 / 17 (35%)	21 / 55 (38%)	21 / 63 (33%)	0.809
statin		56 / 160 (35%)	7 / 25 (28%)	8 / 17 (47%)	23 / 55 (42%)	18 / 63 (29%)	0.280
BMI (kg/m²)		29.2 ± 6.7	29.5 ± 6.3	31.7 ± 6.6	28.3 ± 5.6	29.1 ± 7.7	0.348
LGE scar present		92 / 160 (58%)	15 / 25 (60%)	11 / 17 (65%)	35 / 55 (64%)	31 / 63 (49%)	0.395
mrilved_volume( ml/m²)		124.8 ± 36.8	122.9 ± 41.3	123.4 ± 31.1	129.6 ± 37.5	121.8 ± 36.1	0.658
mrilves_volume (ml/m²)		86.8 ± 36.9	89.7 ± 44.1	84.8 ± 32.4	91.5 ± 38.6	82.0 ± 33.4	0.708
mri_lvef%		32.4 ± 10.5	29.8 ± 12.1	32.9 ± 10.3	31.5 ± 10.8	34.2 ± 9.7	0.375
mrilvmass (g/m²)		77.7 ± 23.0	73.8 ± 24.3	87.1 ± 17.4	80.2 ± 23.7	74.5 ± 22.6	0.068
mr_volume (ml)		10.2 ± 8.9	9.0 ± 8.7	12.2 ± 6.8	9.6 ± 7.9	10.7 ± 10.2	0.394
mr_fraction%		13.4 ± 11.3	14.5 ± 14.1	15.0 ± 8.1	13.0 ± 11.2	13.0 ± 11.2	0.653
rved_volume (ml/m²)		84.1 ± 26.4	84.1 ± 31.5	72.9 ± 15.1	85.8 ± 27.7	85.7 ± 25.3	0.307
rves_volume ( ml/m²)		48.1 ± 24.1	53.5 ± 32.9	39.9 ± 13.4	48.3 ± 24.2	48.1 ± 22.3	0.562
rvef%		44.9 ± 12.6	40.0 ± 14.5	45.6 ± 12.1	46.1 ± 11.9	45.4 ± 12.4	0.316
LA_reservoir%		24.9 ± 11.6	37.0 ± 9.3	18.1 ± 4.8	33.0 ± 7.6	15.1 ± 5.4	< 0.001
LA_conduit%		12.6 ± 7.5	20.3 ± 8.1	8.3 ± 3.7	17.0 ± 6.1	7.1 ± 3.1	< 0.001
LA_booster%		12.4 ± 6.5	17.6 ± 5.3	9.8 ± 4.2	16.0 ± 5.5	8.0 ± 4.9	< 0.001
LAEF%		41.4 ± 17.7	54.1 ± 10.5	34.7 ± 13.7	51.1 ± 13.4	29.7 ± 16.1	< 0.001
LAVI (ml)		39.0 ± 15.8	31.6 ± 10.1	36.6 ± 17.3	33.4 ± 12.0	47.5 ± 16.6	< 0.001
ECV%		33.1 ± 5.8	26.6 ± 1.8	27.2 ± 1.7	34.8 ± 4.1	35.9 ± 5.7	< 0.001
LGE%		3.3 ± 5.7	2.9 ± 5.5	4.6 ± 6.0	4.2 ± 8.2	2.6 ± 5.7	0.172

Variable	N = 160		Low ECV + High LAres N = 25	Low ECV + Low LAres N = 17	High ECV + High LAres N = 55	High ECV + Low LAres N = 63	p-value
Age (years)		54.9 ± 15.1	57.9 ± 17.0	52.9 ± 13.8	54.0 ± 15.1	54.9 ± 14.8	0.386
Female		62 / 160 (39%)	14 / 25 (56%)	5 / 17 (29%)	20 / 55 (36%)	23 / 63 (37%)	0.277
Hypertension		100 / 160 (63%)	16 / 25 (64%)	10 / 17 (59%)	36 / 55 (65%)	38 / 63 (60%)	0.922
Diabetes		32 / 160 (20%)	4 / 25 (16%)	4 / 17 (24%)	12 / 55 (22%)	12 / 63 (19%)	0.916
Hyperlipidemia		68 / 160 (43%)	11 / 25 (44%)	9 / 17 (53%)	23 / 55 (42%)	25 / 63 (40%)	0.815
Atrial Fibrillation		29 / 160 (18%)	4 / 25 (16%)	3 / 17 (18%)	10 / 55 (18%)	12 / 63 (19%)	0.999
Beta blockers		138 / 160 (86%)	20 / 25 (80%)	16 / 17 (94%)	46 / 55 (84%)	56 / 63 (89%)	0.516
ACEI/ARB		123 / 160 (77%)	21 / 25 (84%)	14 / 17 (82%)	41 / 55 (75%)	47 / 63 (75%)	0.785
aldosterone		59 / 160 (37%)	11 / 25 (44%)	6 / 17 (35%)	21 / 55 (38%)	21 / 63 (33%)	0.809
statin		56 / 160 (35%)	7 / 25 (28%)	8 / 17 (47%)	23 / 55 (42%)	18 / 63 (29%)	0.280
BMI (kg/m²)		29.2 ± 6.7	29.5 ± 6.3	31.7 ± 6.6	28.3 ± 5.6	29.1 ± 7.7	0.348
LGE scar present		92 / 160 (58%)	15 / 25 (60%)	11 / 17 (65%)	35 / 55 (64%)	31 / 63 (49%)	0.395
lved_volume( ml/m²)		124.8 ± 36.8	122.9 ± 41.3	123.4 ± 31.1	129.6 ± 37.5	121.8 ± 36.1	0.658
lves_volume ( ml/m²)		86.8 ± 36.9	89.7 ± 44.1	84.8 ± 32.4	91.5 ± 38.6	82.0 ± 33.4	0.708
lvef%		32.5 ± 10.6	29.8 ± 12.1	32.9 ± 10.3	31.5 ± 10.8	34.2 ± 9.7	0.375
lV_massi (g/m²)		77.7 ± 23.0	73.8 ± 24.3	87.1 ± 17.4	80.2 ± 23.7	74.5 ± 22.6	0.068
mr_volume ml		10.2 ± 8.9	9.0 ± 8.7	12.2 ± 6.8	9.6 ± 7.9	10.7 ± 10.2	0.394
mr_fraction%		13.4 ± 11.3	14.5 ± 14.1	15.0 ± 8.1	13.0 ± 11.2	13.0 ± 11.2	0.653
rved_volume (ml/m²)		84.1 ± 26.4	84.1 ± 31.5	72.9 ± 15.1	85.8 ± 27.7	85.7 ± 25.3	0.307
rves_volume (ml/m²)		48.1 ± 24.1	53.5 ± 32.9	39.9 ± 13.4	48.3 ± 24.2	48.1 ± 22.3	0.562
rvef%		44.9 ± 12.6	40.0 ± 14.5	45.6 ± 12.1	46.1 ± 11.9	45.4 ± 12.4	0.316
LA_reservoir%		24.9 ± 11.6	37.0 ± 9.3	18.1 ± 4.8	33.0 ± 7.6	15.1 ± 5.4	< 0.001
LA_conduit%		12.7 ± 7.6	20.3 ± 8.1	8.3 ± 3.7	17.0 ± 6.1	7.1 ± 3.1	< 0.001
LA_booster%		12.4 ± 6.6	17.6 ± 5.3	9.8 ± 4.2	16.0 ± 5.5	8.0 ± 4.9	< 0.001
LAEF%		41.4 ± 17.7	54.1 ± 10.5	34.7 ± 13.7	51.1 ± 13.4	29.7 ± 16.1	< 0.001
LA_volume (ml/m²⁾		39.0 ± 15.8	31.6 ± 10.1	36.6 ± 17.3	33.4 ± 12.0	47.5 ± 16.6	< 0.001
ECV%		33.1 ± 5.8	26.6 ± 1.8	27.2 ± 1.7	34.8 ± 4.1	35.9 ± 5.7	< 0.001
LGE%		3.4 ± 6.7	2.9 ± 5.5	4.6 ± 6.0	4.2 ± 8.2	2.6 ± 5.7	0.172

1.A. Adjusted hazard ratios from the parsimonious Cox model (ECV per 5%, LA reservoir per 5 pp, MR fraction%, medical risk score). Points = HR; bars = 95% CI; dashed line = HR=1. 1.B. Effect-modification: HR for ECV per +5% at LA strain tertiles (reservoir, conduit, booster) from a-priori–adjusted Cox models. Preserved reservoir/conduit attenuates ECV-related risk(LASr; HR 0.97 [0.94–0.99]; p=0.049, LAScd (0.94 [0.90–0.98]; p=0.009) 1.C. Cross-sectional association between ECV and LA strain (reservoir, conduit, booster) with linear fit and 95% CI; higher ECV was associated with worse LA mechanics: LASr −2.51 pp, LAScd −1.48 pp, and LASb −1.15 pp per +5% ECV (all p≤0.011)

Fig. 2. Kaplan–Meier event-free survival by combined ECV and LA reservoir (LASr) strata & numbers at risk shown below. Cut points (medians, n=160): ECV 32.0%; LASr 23.8%; (Cohort means: ECV 33.1%; LASr 24.9%). Numbers at risk are shown beneath the plot. The higher risk occurs with high ECV+ low LA reservoir strain (red): 2-yr EFS: 65% , (53-81%), while preserved LA mechanics attenuate risk when ECV is high (green) 97% (92-100%)