Non-contrast T1-mapping detects acute myocardial edema with high diagnostic accuracy: a comparison to T2-weighted cardiovascular magnetic resonance

Abstract

Background: T2w-CMR is used widely to assess myocardial edema. Quantitative T1-mapping is also sensitive to changes in free water content. We hypothesized that T1-mapping would have a higher diagnostic performance in detecting acute edema than dark-blood and bright-blood T2w-CMR.

Methods: We investigated 21 controls (55 ± 13 years) and 21 patients (61 ± 10 years) with Takotsubo cardiomyopathy or acute regional myocardial edema without infarction. CMR performed within 7 days included cine, T1-mapping using ShMOLLI, dark-blood T2-STIR, bright-blood ACUT2E and LGE imaging. We analyzed wall motion, myocardial T1 values and T2 signal intensity (SI) ratio relative to both skeletal muscle and remote myocardium.

Results: All patients had acute cardiac symptoms, increased Troponin I (0.15-36.80 ug/L) and acute wall motion abnormalities but no LGE. T1 was increased in patient segments with abnormal and normal wall motion compared to controls (1113 ± 94 ms, 1029 ± 59 ms and 944 ± 17 ms, respectively; p < 0.001). T2 SI ratio using STIR and ACUT2E was also increased in patient segments with abnormal and normal wall motion compared to controls (all p < 0.02). Receiver operator characteristics analysis showed that T1-mapping had a significantly larger area-under-the-curve (AUC = 0.94) compared to T2-weighted methods, whether the reference ROI was skeletal muscle or remote myocardium (AUC = 0.58-0.89; p < 0.03). A T1 value of greater than 990 ms most optimally differentiated segments affected by edema from normal segments at 1.5 T, with a sensitivity and specificity of 92 %.

Conclusions: Non-contrast T1-mapping using ShMOLLI is a novel method for objectively detecting myocardial edema with a high diagnostic performance. T1-mapping may serve as a complementary technique to T2-weighted imaging for assessing myocardial edema in ischemic and non-ischemic heart disease, such as quantifying area-at-risk and diagnosing myocarditis.

Figure 1

CMR imaging modalities used for comparison – representative images in short-axis slices from: (Left panel) Normal volunteer (Middle panel) Patient with Takotsubo cardiomyopathy. Note high T2 signal and T1 values in an apical slice. (Right panel) Patient with regional edema in the anterior wall territory. Note high T2 signal and T1 values in the anterior, anterolateral wall and anterior septum. Rows: (A) Dark-blood T2-weighted images (reference ROIs in skeletal muscle and remote myocardium not shown). (B) Bright-blood T2-weighted images. (C) Colour ShMOLLI T1-maps. Green denotes normal myocardium. Red denotes increased T1 values (D) R² maps of ShMOLLI inversion recovery fit used to verify the quality of T1-maps. (E) LGE imaging.

Figure 2

Relationship between: (A) myocardial T1 values and T2 signal intensity (SI) ratio* derived from dark-blood T2w-CMR; (B) myocardial T1 values and T2 SI ratio derived from bright-blood T2w-CMR; (C) myocardial T2 SI ratio dark-blood vs. bright-blood T2w-CMR. *T2 SI ratio = SI_myocardium/SI_{skeletal muscle.}

Figure 3

Receiver operator characteristic curves for the detection of acute myocardial edema by ShMOLLI T1-mapping, dark-blood T2w-CMR and bright-blood T2w-CMR. For T2-weighted methods, reference regions of interest for comparison of myocardial signal intensity are specified in brackets. AUC = area-under-the-curve.

Figure 4

Receiver operator characteristic curves for the detection of acute myocardial edema according to patient groups using ShMOLLI T1-mapping, dark-blood T2w-CMR and bright-blood T2w-CMR. (A) Patients with regional edema vs. controls. (B) Patients with Takotsubo cardiomyopathy vs. controls. AUC = area-under-the-curve.

Figure 5

ROC curves for the detection of acute myocardial edema by ShMOLLI T1-mapping, dark-blood T2w-CMR and bright-blood T2w-CMR using only segments of good quality for all methods (left panel), and using all available segments, including segments previously rejected for artifacts for all methods (right panel.). This shows that ShMOLLI T1-mapping retains its superiority despite the re-inclusion of segments affected by artifacts. All comparisons are statistically significant (p < 0.03). AUC = area-under-the-curve.