The potential of cine balanced steady-state free precession mid short-axis myocardial signal intensity variations as a diagnostic marker: a proof-of-concept in dilated cardiomyopathy

Background: It was recently shown that the myocardial contrast fluctuations seen in short-axis (SAX) balanced steady-state free-precession (bSSFP) cine images are caused by longitudinal (through-plane) motion [1]. Although cine images provide in various ways functional information, these bSSFP myocardial signal intensity (mcSI) fluctuations are unexplored as a diagnostic marker. As a proof-of-concept, we therefore assessed if the reduced contractility in patients with dilated cardiomyopathy (DCM) causes altered mcSI profiles vs. healthy subjects. In addition, we evaluated if functional recovery after appropriate heart-failure (HF) therapy would result in (partial) normalization of the mcSI profile.

Methods: The pre-contrast mid SAX mcSI variation during the heart cycle was determined in DCM patients with a severely (EF< 35%; n=17) or moderately (EF >=35%; n=11) reduced ejection fraction (EF) versus healthy controls (n=49). In an additional seventeen DCM patients, images at baseline and after HF therapy (‘follow-up’) were analysed. Myocardial contours, EF and global longitudinal strain (GLS) were derived using commercial software (Neosoft SuiteHeart). Next, mcSI profiles were calculated from the pre-contrast mid SAX myocardium (in house scripts, Matlab R2022) and per subject scaled to the mcSI at mid-diastole (“Relative mcSI%”). The bSSFP CMR images were acquired at 1.5T (Philips, Ingenia) with acquisition parameters: TR 3.3ms, TE 1.7ms, flip angle 60°, R=2 , matrix 160 × 256; field of view, 350(FE)x430(PE) mm; slice thickness 8 mm and number of phases, 30. In addition, simulations were performed [1] to examine whether mcSI changes could be caused by the heart rate dependency of the mcSI profile assuming normal contraction.

Results: The mcSI from severely or moderately reduced EF patients and controls all differ during the heart cycle (Fig. 1). Functional recovery after HF therapy yielded a mcSI profile nearer to the one observed in healthy subjects. The maximum increase in relative mcSI shows a good correlation with both EF (R² = 0.53) and GLS (R² = 0.51). Although heart rate was also significantly lower at recovery (FU) vs. baseline (Table 1), and simulations indicate that the mcSI profile is heart rate dependent (Fig. 2), this heart rate dependency seems insufficient to explain the difference in mcSI profiles.

Conclusion: This proof-of-concept study shows for the first time that mcSI fluctuations of standard SAX cine images might be considered as a potential diagnostic marker, as this study demonstrates the sensitivity of the mcSI profiles to functional differences in DCM subjects and changes with treatment.

Left ventricular ejection fraction (LVEF), peak global longitudinal strain (peak GLS), heart rate and incidence of dyssynchrony in the control subjects, low EF, high EF, baseline (BAS) or follow-up (FU) datasets. Median with 1st and 3rd quartile; Mann–Whitney U test, significance levels indicated as follows: (*) EF=35; ($)EF=35 vs. CTRL; (&) BAS vs. FU; (§) BAS vs. CTRL; (£) FU vs. CTRL; 2 symbols; p<0.01, 3 symbols: p<0.001.

	n	LVEF (%)	peak GLS (%)	HR (bpm)	Dyssyn-chrony (%  subjects)
CTRL	49	58 (56 – 62)	-17.9 (-19.6 – -16.9)	61 (54 – 68)	0
EF< 35	17	20 (17 – 26) ***,$$$	-7.3 (  -9.4 –   -5.7) ***,$$$	77 (66 – 87) $$$	50
EF>=35	11	39 (36 – 49) NULL#	-14.1 (-15.7 – -12.0) NULL#	68 (63 – 76) NULL#	55
BAS	17	25 (17 – 33) &&&,§§§	-7.4 (-10.2  –  -5.7) &&&,§§§	80 (73 – 91) &&,§§§	47
FU	17	43 (38 – 48) £££	-12.7 (-13.0 – -11.4) £££	67 (61 – 73)	19

Figure 1: Top row shows the variation of the myocardial cine signal intensity (mcSI) in mid short axis cine images during the heart cycle for subjects with severely (EF <35%) or moderately (EF >= 35%) reduced EF vs. controls (left) and for baseline and follow-up acquisitions vs. controls (right) (mean +/- SE). Bottom panels show the correlation between the mcSI and EF (left) or peak global longitudinal strain (GLS) (right).
Figure 2: Heart rate (HR) dependency of the simulated myocardial signal intensity (mcSI) when modelling the cine magnetisation for through-plane motion as typical in healthy subjects. The horizontal axis is one cardiac cycle, as on the Figure 1 graphs. Although the heart rate changes significantly from baseline to recovery, or, between severe and moderately reduced EF (table 1), the in vivo change in mcSI (Fig. 1, top) are larger than the simulations predict and are correlated with functional parameters (Fig. 1, bottom).