Prevalence and distribution of regional scar in dysfunctional myocardial segments in Duchenne muscular dystrophy

Abstract

Background: The segmental relationship between cardiovascular magnetic resonance (CMR) peak circumferential strain (Ecc) and myocardial scar has not been well characterized in Duchenne muscular dystrophy (DMD), and it is unknown whether echocardiography accurately measures Ecc in DMD. We assessed segmental Ecc and scar using CMR with myocardial tissue tagging and late gadolinium enhancement (LGE) in patients with DMD, then compared CMR with echocardiographic velocity vector imaging (VVI) for regional Ecc based on independent observer assessments.

Results: Participants enrolled (n = 16; age 8-23) had median left ventricular (LV) ejection fraction of 0.52 (range 0.28-0.69), and 156 basal and mid-cavity myocardial segments from the 13 patients completing the LGE protocol were analyzed for strain and scar. Segmental CMR Ecc in the most negative quartile (quartile 4) ruled out scar in that segment, but scar was present in 46% of segments in the least negative (most dysfunctional) Ecc quartile 1, 33% of Ecc quartile 2 segments, and 15% of Ecc quartile 3 segments. Overall scar prevalence in inferior, inferolateral, and anterolateral segments was eight times higher than in inferoseptal, anteroseptal, and anterior segments (p < 0.001). This increased proportion of scar in lateral versus septal segments was consistent across CMR Ecc quartiles (quartile 1: 76% versus 11%, p = 0.001; quartile 2: 65% versus 9%, p < 0.001; quartile 3: 38% versus 0%, p < 0.001). Echocardiographic analysis could be performed in 12 of 14 patients with CMR exams and had to be limited to mid-cavity slices. Echo segmental Ecc in the most negative quartile made scar by CMR in that segment highly unlikely, but the correlation in segmental Ecc between CMR and echo was limited (r = 0.27; p = 0.02).

Conclusions: The relationship between scar and Ecc in DMD is complex. Among myocardial segments with depressed Ecc, scar prevalence was much higher in inferior, inferolateral, and anterolateral segments, indicating a regionally dependent association between abnormal Ecc and scar, with free wall segments commonly developing dysfunction with scar and septal segments developing dysfunction without scar. Although normal echocardiographic Ecc predicted absence of scar, regional echocardiographic Ecc by VVI has only a limited association with CMR Ecc in DMD.

Figure 1

DMD late gadolinium enhancement. Basal (panel A) and mid-cavity (panel B) slices of subepicardial and midmyocardial scar involving inferolateral and anterolateral segments in a patient with DMD. The white (hyperenhanced) region (arrow) is scar, while the black represents normal myocardium.

Figure 2

Distribution of myocardial scar in basal and mid-cavity segments. The bar graph shows inferolateral, inferior, and anterolateral predominance of scar, as well as the distribution of scar between basal and mid-cavity segments. Ant-Lat = anterolateral; Inf-Lat = inferolateral; Inf-Sept = inferoseptal; Ant-Sept = anteroseptal.

Figure 3

Segmental scar and circumferential strain by CMR versus VVI echocardiography. The distribution of myocardial fibrosis by quartile of Ecc as assessed with (A) CMR (all segments) or (B) VVI echocardiography (mid-cavity segments only) is shown. A) No segments with Ecc by CMR in the fourth (most negative) quartile had myocardial scar. Myocardial scar was most common in the two lowest (most dysfunctional) Ecc quartiles. B) Scar was also unlikely in segments falling into the fourth quartile of Ecc as assessed by echo VVI and was distributed in the lower three echo Ecc quartiles. 1^st quartile = least negative Ecc (most dysfunctional); 4^th quartile = most negative Ecc.

Figure 4

Increased proportion of scar in lateral versus septal segments across CMR circumferential strain quartiles. The figure shows an increased proportion of scar in group 1 (lateral) segments versus group 2 (septal) segments based on the analysis of all 156 myocardial segments. This effect was consistent across all lower three CMR quartiles of Ecc, indicating that there is increased prevalence of segmental scar at any level of segmental Ecc in group 1 (lateral) segments versus group 2 (septal) segments. In other words, the association between depressed Ecc and scar has a significant regional dependence, with dysfunctional lateral segments much more likely to have scar than dysfunctional septal segments. 1^st quartile = least negative Ecc (most dysfunctional); 4^th quartile = most negative Ecc.

Figure 5

Relationship between segmental transmurality of scar and regional Ecc. The relationship between the segmental transmurality of scar (characterized as 0-25% transmural, 26-50% transmural, or >50% transmural) and regional Ecc is shown, with regional Ecc characterized by (A) CMR (all segments), and (B) VVI echocardiography (mid-cavity segments only). Note that all segments with scar of greater than 50% transmurality were in the most dysfunctional CMR Ecc quartile (Q1), and segments with intermediate scar transmurality were distributed among the lower two Ecc quartiles (Q1 and Q2).