What is the mid-wall linear high intensity “lesion” on cardiovascular magnetic resonance late gadolinium enhancement?

Cardiovascular magnetic resonance imaging (CMR) is a very useful tool that can noninvasively evaluate cardiac anatomy, function, and blood flow [1, 2]. Myocardial viability is commonly assessed using late gadolinium enhancement (LGE). CMR LGE is a valuable technique that makes it possible to diagnose myocardial fibrosis and disorders noninvasively [3‐5]. The most widely used pulse sequence type of LGE is inversion recovery (IR) with magnitude reconstruction, which usually obtains an image 10–15 min after administration of gadolinium contrast agent [6, 7]. About 10 min after injection of the contrast medium, this medium is evenly distributed in extracellular or intravascular space and is in equilibrium. With LGE, infarcted myocardium or fibrosis show relatively high signal intensity as compared with normal myocardium, due to increased stroma. Therefore, this technique is used as a powerful option for the evaluation of non-ischemic cardiomyopathy, because it allows the differential diagnosis of cardiomyopathy, with high probability, based on various patterns of LGE [8‐13].

However, CMR produces some artifacts due to the direct effect of myocardial wall motion, or the indirect effect of shortening acquisition time, to eliminate movement [14]. In LGE studies, choosing the correct inversion time (TI) to null normal myocardium can be challenging for inexperienced operators. If TI is not adequate, the normal myocardium will not be nulled, the contrast between infarcted and normal myocardium will be reduced, and additional misleading artifacts will be produced. Due to these effects, we often observe some high intensity signals on LGE imaging, which are believed to be artifacts.

In particular, we sometimes experience a linear, mid-wall high intensity in the basal septum in the short axis view, which often poses diagnostic difficulty in the clinical setting. Such linear high intensity (LHI) appears as round-shaped high intensity in the orthogonal long axis view and is thereby clearly distinguishable from right ventricular (RV) blood pools behind RV trabeculation [15] or artifacts (Fig. 1). The linear mid-wall LGE in the septum is also a pattern of LGE that is found in dilated cardiomyopathy (DCM) [16]. In addition, mid-wall fibrosis in DCM, as determined by LGE, is a predictor of the combined end point of all-cause mortality and cardiovascular hospitalization, which is independent of ventricular remodeling [17]. Therefore, it is clinically very important to distinguish the LHI found in the septum from mid-wall fibrosis in patients with DCM. However, no previous papers have discussed this issue.

In recent years, multidetector computed tomography (CT) with submillimeter collimation and high-speed gantry rotation has allowed the representation of thinner vessels, such as septal branches, with high temporal resolution and isotropic voxels [18].

We noticed that the LHI observed in the short axis view on CMR LGE was anatomically close to the course of the anterior septal perforator arteries on coronary CT angiography (CorCTA) (Fig. 2). The anterior septal perforator arteries originate from the proximal part of the left anterior descending coronary artery (LAD) and irrigate two-thirds of the upper part of the interventricular septum. They vary in number, with an average of eight branches; the first septal artery is usually the largest and the longest [19, 20].

A recent case report demonstrated that the basal septal perforator vessel can mimic late iodine enhancement in delayed-phase cardiovascular CT [21]. On the other hand, since LGE imaging is performed when the signal of normal myocardium becomes null, it is considered that a blood vessel lumen running in the myocardium is rendered with a relatively high signal. We therefore hypothesized that the mid-wall LHI on LGE may represent contrast enhancement of the anterior septal perforator arteries. To investigate this hypothesis, we compared the linear, mid-wall LHI lesions identified on LGE in the basal septum with the anterior septal perforator arteries identified on CorCTAto determine the relationship between them.

Among the 111 patients, there were 55 LHI+ and 56 LHI- patients. We observed that the LHI on LGE and anterior septal perforator arteries on CorCTA were very similar in shape and running direction (Fig. 3a-h). The measured length of the anterior septal perforator arteries on CorCTA was significantly longer than that in LHI on CMR LGE (16 ± 10 mm vs. 14 ± 10 mm; P < 0.05) (Fig. 3i, j). A linear regression analysis revealed that there was good agreement between LHI length on LGE and that of septal perforator arteries on CorCTA (R² = 0.53, P = 0.13; Fig. 4). Among the 55 LHI+ patients, anterior septal perforator arteries were seen in the same region on CorCTA in 53 patients (96%) (Table 2). On the other hand, in the 56 LHI- patients, anterior septal perforator arteries were identified in only 39 patients on CorCTA (70%). The measured length of the anterior septal perforator arteries was significantly shorter in LHI- patients (10 ± 8 mm vs. 21 ± 8 mm for LHI+; P < 0.05) (Fig. 5).

Compared with mid-wall fibrosis in DCM, LHI has the following characteristics: 1) It runs on the epicardial side of the septum (one-third on the adventitial side of the septum) in 98% (54/55) of the patients with LHI and in 10% (2/21) of the those with DCM-related mid-wall fibrosis; 2) It starts from the anterior interventricular sulcus in 96% (53/55) of patients with LHI and in 24% (5/21) of those with DCM-related mid-wall fibrosis; 3) It stays in the anterior septum (does not reach the inferior septum) in 91% (50/55) of the patients with LHI and in 19% (4/21) of those with mid-wall fibrosis of DCM (Fig. 6).

CMR LGE is a useful tool for scar detection, based on the even distribution of contrast medium in the extracellular or intravascular space. However, with the increasingly high resolution of CMR in recent years, a high intensity, which has been considered to be nonspecific, has also been frequently observed [14]. It has not previously been described what this septal LHI seen on LGE represents. Our report demonstrated that the LHI found in the basal septum is likely to represent a normal septal perforator artery.

In the present study, anterior septal perforator arteries were identified at a high rate by CorCTA among LHI+ patients, and the length of these features correlated well. Also, their morphology and running directions were very similar. On the other hand, for LHI- patients, the anterior septal branch length measured on CorCTA was significantly shorter. Based on these results, we considered that the LHI observed in the basal septum in short-axis LGE images reflects contrast enhancement of the anterior septal perforator arteries.

In this study, LHI by CMR was visually confirmed as anterior septal perforator arteries on CorCTA in LHI+ patients. Since all LGE images were obtained using a 3D sequence, the septal high intensity could be identified well due to the high spatial resolution, high signal-to-noise ratio, and gapless slices. On the other hand, the measured length of the anterior septal perforator arteries on CorCTA was significantly longer than the LHI on CMR LGE. The reason may be that CorCTA has a higher spatial resolution than CMR, as shown in Fig. 3 (i) and (j), and could measure vessels to the periphery. In 56 of 111 patients, anterior septal perforator arteries were found on CorCTA, but these patients were defined as LHI- by CMR. Given that the mean length of the anterior septal perforator arteries on CorCTA was 10 ± 8 mm in LHI- patients, these arteries may not have reached the set LHI criteria (observable for 1.5 cm or more) in this study. It is possible that the LHI described here may be visualized increasingly with improvement in the spatial resolution and signal-to-noise ratio of CMR in future. Although in a small number (2 of 55 patients), for some patients that were defined as positive LHI by CMR, no marked anterior septal perforator arteries could be identified by CT. The LV myocardium includes not only the septal artery, but also vein, capillaries, arterio- and venoluminal vessels, and sinusoids, etc. [22]. Although very rare, these structures might also be recognized as LHI by LGE. However, we believe that the septal LHI is likely to be an artery, and not a vein, for the following reasons: first, it is very similar to the running of the septal perforator artery on CorCTA images; second, arteries with fast blood flow are considered to be more visible due to the CMR-inflow effect; third, LGE data is collected during diastole phase with emphasis on arteries. In this study, we could visualize a septal perforator artery using a slice thickness of 10 mm. First, a septal perforator artery is anatomically located along the short axis of the myocardium. Furthermore, by nullifying the myocardial signal using the IR sequences, it was possible to visualize the septal perforator artery running inside the myocardium at high signal and obtain high contrast. For these reasons, we considered that the septal perforator artery could be clearly depicted on the CMR short-axis slice, which has a lower spatial resolution than CorCTA.

It is clinically important to differentiate LHI from mid-wall fibrosis in DCM. In general, linear delayed enhancement of DCM is found “mid” of the septum myocardium (endo- and epi-cardial sparing) and extending from the anterior to the inferior wall (noncoronary territory distribution) [17]. There was excellent agreement between the pathological location of the mid-wall fibrosis and the premortem location of the mid-wall LGE [17]. On the other hand, anterior septal perforator arteries anatomically originate from the proximal part of the LAD and irrigate two-thirds of the upper part of the interventricular septum [23]. Furthermore, it has been reported that, in the short axis of cardiovascular CT, an anterior septal perforator vein runs on the endo-cardial side in the myocardium, and an anterior septal perforator artery runs on the epi-cardial side in the myocardium [21]. Based on the anatomical information, we present the LGE patterns of LHI and DCM considering the three abovementioned points. The mid-wall fibrosis is relatively specific for DCM and regarded as an important indicator of poor prognosis. Therefore, it is very important to distinguish the LHI found in the septum and mid-wall fibrosis in DCM. In DCM patients with mid-septal fibrosis, no apparent linear structure could be delineated as the normal septal perforator artery. Myocardial thinning and a range of mid-wall fibrosis may be involved in this phenomenon, but this result alone cannot clarify this in detail. As the number of comparison DCM patient was very small, further study is needed. In addition, improving the spatial resolution of delayed contrast images may allow differentiation between the septal perforator artery and mid-wall fibrosis.

The present study has several limitations. First, LGE was acquired with only IR sequences, and not with a phase-sensitive inversion-recovery (PSIR) method. PSIR techniques have the benefits of a high contrast-to-noise ratio and accurate depiction of the enhanced region [24]. Therefore, we speculate that PSIR techniques can potentially improve the detectability of the septal LHI. Second, only 1 reader examined the cases; thereafter, a second reader confirmed the assessment without an independent reading.

Finally, awareness of septum high intensity areas in LGE images is important for interpretation of imaging findings. Radiologists familiar with this finding may be able to avoid unnecessary examinations and follow-ups after CMR.

The LHI signal observed in the basal anterior septum in short-axis CMR LGE images may reflect the contrast enhancement of the anterior septal perforator arteries. It is necessary to interpret this septum high intensity against knowledge of anatomic structure, to avoid misinterpretations of LGE and prevent misdiagnosis.