Papillary muscle involvement in myocardial infarction: Initial results using multicontrast late-enhancement MRI

Authors


Abstract

We hypothesized that multicontrast late-enhancement (MCLE) MRI would improve the identification of papillary muscle involvement (PM-MI) in patients with myocardial infarction (MI), compared with conventional late gadolinium enhancement (LGE) MRI using the inversion recovery fast gradient echo (IR-FGRE) technique. Cardiac LGE-MRI studies using both MCLE and IR-FGRE pulse sequences were performed on a 1.5 Tesla (T) MRI system in 23 patients following MI. In all patients, PM-MI was confirmed by the diagnostic criteria as outlined below: (a) the increased signal intensity of PM was the same or similar to that of adjacent hyper-enhanced left ventricular (LV) infarct segments; and (b) the hyper-enhanced PM region was limited to the PM area defined by precontrast cine images of steady-state free precession (SSFP). Visual contrast score was rated according to the differentiation between LV blood pool and hyper-enhanced infarct myocardium. Quantitative contrast-noise ratios (CNR) of infarct relative to blood pool and viable myocardium were also measured on MCLE and IR-FGRE images. Of these 23 patients, 13 studies demonstrated primarily involvement of the territories of the right coronary (RCA, 8 patients) and/or left circumflex (LCX, 5 patients) arteries and 10 involved the territories of left anterior descending artery (LAD) with some LCX involvement. Although both IR-FGRE and MCLE determined the presence and extent of LV MI, better visual contrast scores were achieved in MCLE (2.9 ± 0.3) compared with IR-FGRE (1.6 ± 0.8, P < 0.001). The CNRs of infarct relative to LV blood pool showed a significant statistical difference (n = 23, P < 0.00001) between MCLE (16.2 ± 7.2) and IR-FGRE images (4.8 ± 4.1), which is consistent with the result of visual contrast scores between infarct and LV blood pool. The CNRs of infarct versus viable myocardium did not demonstrate a significant statistical difference (n = 23, P = 0.61) between MCLE (14.4 ± 7.0) and IR-FGRE images (13.6 ± 6.1). MCLE clearly demonstrated PM-MI in all cases (100%, 23/23) while only 39% (9/23) could be visualized on the corresponding IR-FGRE images. In conclusion, MCLE imaging provides better contrast between blood pool and infarct myocardium, thus improving the determination of PM-MI. J. Magn. Reson. Imaging 2011;33:211–216. © 2010 Wiley-Liss, Inc.

THE PAPILLARY MUSCLE (PM) is an integral component of the mitral valve apparatus, which also includes mitral valve leaflets, annulus, and the chordae tendineae. Acute or chronic myocardial infarction (MI) with PM ischemia is a primary factor leading to the occurrence of mitral regurgitation, with associated substantial morbidity and mortality (1, 2). Mitral valve replacement or repair is usually indicated for severe ischemic mitral regurgitation (3, 4). For patients with moderate ischemic mitral valve regurgitation surgical mitral valve repair might be beneficial. Moreover, PM-MI is also a potential source of ventricular arrhythmia (5). Surgical excision and radiofrequency ablation of PM have been successfully applied to eliminate the ventricular arrhythmias (6, 7).

Although echocardiography is the method of choice for the demonstration of ischemic mitral valve regurgitation, the visual confirmation of PM involvement is very challenging using this modality (8). Myocardial positron emission tomography (PET) has been reported to depict the patterns of PM ischemia (9), but the spatial resolution and ionizing radiation may limit its application. Currently, late gadolinium enhancement (LGE) MRI has been recognized as a gold standard for the determination of MI and can demonstrate subendocardial MI and small microvascular obstruction with a high sensitivity (10–13). However, the better spatial resolution of conventional LGE-MRI using the inversion recovery fast gradient echo (IR-FGRE) technique has not yielded substantially improved identification of PM involvement in patients with MI. Poor contrast between the LV blood pool and infarct myocardium and the fact that a large portion of PM extends into the LV cavity are important factors in the identification of PM-MI (14). Better blood pool suppression may help to demonstrate PM-MI. Recently a newly developed multicontrast late-enhancement (MCLE) MRI has been proposed for the simultaneous detection of scar tissue and wall motion abnormalities in patients with chronic MI (15). In this pulse sequence, complete blood pool suppression can be achieved at a specific cardiac phase, enabling improved identification of LGE in infarcted segments (16). We hypothesize that this capability of MCLE imaging will improve the identification of PM-MI in patients with acute and chronic MI, compared with conventional IR-FGRE imaging.

MATERIALS AND METHODS

The institutional research ethics board has approved the study protocol, and patient informed consent was obtained from the subjects. Cardiac LGE-MRI studies using both MCLE and IR-FGRE pulse sequences in patients with MI were performed on a 1.5T GE Signa HDx system (GE Healthcare, Milwaukee, WI). Twenty-three patients (21 males, 2 females, average age of 62 ± 10 years old; 5 acute MI within 7 days, 18 chronic MI > 4 weeks) met the diagnostic criteria of PM-MI, as outlined below. ECG gating and an eight-channel phased-array cardiac coil were used for the study.

Before contrast administration, a short-axis oblique (SAO) steady-state free precession (SSFP) study covering the whole LV was obtained for the evaluation of LV function. Two or four chamber SSFP studies were acquired for the demonstration of mitral valve regurgitation. The MR parameters of the cine SSFP sequence were as follows (16): bandwidth 125 kHz, flip angle 45°, views per segment (VPS) 16, TR/TE 3.7/1.6 ms, field of view (FOV) = 32 cm, image matrix = 256 × 192, and slice thickness 8 mm. 20 phase-resolved images over the heart cycle were acquired on average through a 12-s breathhold. Both IR-FGRE and MCLE covering the whole LV in SAO and/or two- or four-chamber views were performed 10–20 min after a double-dose intravenous bolus injection of Gd-DTPA (0.2 mmol/kg of Magnevist, Berlex Inc., Wayne, NJ). For IR-FGRE, the inversion time (TI) varied from 200 to 300 ms, depending on the null point of healthy myocardium. For MCLE, a segmented SSFP readout is used following an inversion pulse, providing approximately 20 cardiac-phase-resolved images at varying effective TIs (15). The in-plane resolution was 1.5 × 1.5 mm and the through-plane resolution is 8 mm for both IR-FGRE and MCLE. The detailed MR parameters for IR-FGRE were as follows (15): TR/TE 6.0/3.0 ms, receiver bandwidth (rBW) ± 31.5 kHz, flip angle 20°, VPS 20, number of excitations (NEX) 2. The delay time (TDEL) was chosen to yield images in middle to late diastolic phase. Approximately 20 heartbeats (18-s breathholds on average) were required to produce a single LGE-MRI image using IR-FGRE. Parameters for the MCLE sequence were: TR/TE 2.7/1.3 ms, rBW ± 125 kHz, flip angle 30°, VPS 16, TDEL 500 ms, and NEX 1. The MCLE sequence took approximately 13 heartbeats to acquire (one to establish the steady state, and 12 for data acquisition with an average of 11-s breathholds).

PM-MI was considered present if both of the following criteria were satisfied: (i) The increased signal intensity in PM was the same or similar to that of adjacent hyper-enhanced infarct segments on the IR-FGRE or MCLE images; (ii) The hyper-enhanced PM region was limited to the PM area defined by precontrast SSFP images.

Blinded analysis was performed by two independent trained observers on a GE Advantage workstation (GE Healthcare, Milwaukee, WI). The visual contrast score between blood pool and hyper-enhanced LV infarct was rated as excellent (3), good (2), or fair (1) based on their differentiation. Any score difference between two observers was resolved by mutual agreement after reviewing the images together. Quantitative analysis of the contrast difference in MCLE and IR-FGRE images was further performed by two observers by measuring contrast-noise-ratios (CNR) of infarct relative to the LV blood pool and viable myocardium. The observers were blinded to SSFP data including long-axis images during the review of contrast-enhanced MRI data. As the MCLE technique produces multiple images with varied contrast among infarct, blood pool, and viable myocardium, the optimal MCLE image with the maximal visual contrast between the infarct and blood pool or viable remote myocardium was selected for CNR measurements. CNR measurements were based on a region of interest (ROI) analysis of the image data (17). On the optimal MCLE image, circular or elliptical 60–80 mm2 ROIs were placed on regions of infarct, LV blood pool, and viable myocardium, and the mean and standard deviation of the signal intensity in each compartment were obtained. The noise value was measured from the standard deviation of a circular ROI of 200 mm2 placed in the image background. For CNR measurements on IR-FGRE, ROI contours from the MCLE image were copied and propagated on the IR-FGRE image with the same slice localization. Fine contour adjustments were made manually to allow for movement during the cardiac cycle. The CNRs of infarct versus LV blood pool and CNRs of infarct versus viable myocardium were calculated from the difference of mean signal intensity in each compartment divided by the standard deviation of the noise. All data are expressed as the mean value ± SD. Statistical significance in parameter changes was evaluated using the Student's paired t-test.

The severity of ischemic mitral valve regurgitation was classified as mild, moderate, or severe based on the qualitative diagnostic criteria similar to that used in echocardiography (3, 4). LV function measurements were calculated from the endocardial contours planimetered on SSFP SAO images by use of a modified Simpson's method through commercially available Mass plus software (MEDIS, Netherlands) installed on a GE Advantage workstation.

RESULTS

LV function MRI measurements in 23 patients are summarized in Table 1. The majority of patients (87%, 20/23) had significant impaired LV systolic function as demonstrated by a low LV ejection fraction (29.2 ± 14.1%, n = 23). Based on the standard American Heart Association (AHA) 17-segment model (18), all MI patients with PM involvement demonstrated wall motion abnormalities and late hyper-enhancement involving multiple coronary artery territories. Of these 23 patients, 13 studies demonstrated primary involvement in the territories of the right coronary (RCA, 8 patients) and/or left circumflex arteries (LCX, 5 patients) and 10 involved the territories of the left anterior descending artery (LAD) with some LCX involvement.

Table 1. LV Function Measurements (n = 23)
 Age (yr)LVEF (%)ESV (ml)EDV (ml)SV (ml)LVMED (g)
  1. LVEF = LV ejection fraction; ESV = end-systolic volume; EDV = end-diastolic volume; SV = stroke volume; LVMED = LV mass at end-diastolic phase.

Mean62.229.2167.5222.154.6110.1
SD10.114.193.9100.214.338.6

Although both IR-FGRE and MCLE determined the presence and extent of LV MI (Fig. 1), better visual contrast scores between infarct and LV blood pool were achieved using MCLE (2.9 ± 0.3) compared with IR-FGRE (1.6 ± 0.8, P < 0.001). The CNRs of infarct versus LV blood pool showed a significant statistical difference (Fig. 2A, n = 23, P < 0.00001) between MCLE (16.2 ± 7.2) and IR-FGRE images (4.8 ± 4.1), which is consistent with the visual contrast scores. The CNRs of infarct versus viable myocardium did not show a significant statistical difference (Fig. 2B, n = 23, P = 0.61) between MCLE (14.4 ± 7.0) and IR-FGRE images (13.6 ± 6.1). MCLE clearly demonstrated PM-MI in all cases (100%, 23/23). However, only 39% (9/23) could be visualized on the corresponding IR-FGRE images (Fig. 3), which displayed poor contrast between LV blood pool and infarcted myocardium. Mitral valve regurgitation of mild (43.5%, 10/23) or moderate (8.7%, 2/23) severity was noted on four- and/or two-chamber SSFP scans in 52% (12/23) of patients in this study (Fig. 4).

Figure 1.

a,b: IR-FGRE (A) and MCLE (B), both demonstrating the late-enhancement of anteroseptal MI. MCLE images showed better contrast between blood pool, viable myocardium, and infarction. c: Graph demonstrating better contrast scores achieved with MCLE images, compared with IR-FGRE images.

Figure 2.

Comparison of CNR measurements using MCLE and IR-FGRE techniques. a: Significant improvements in CNR of infarct versus LV blood pool were achieved using MCLE over IR-FGRE. b: There was no statistical difference between CNR measurements of infarct versus viable myocardium using MCLE and IR-FGRE.

Figure 3.

An inferior MI with posteromedial PM involvement. a: SSFP image demonstrating the location of PM within the LV blood pool. b: IR-FGRE postcontrast image only demonstrating the late-enhancement of inferior MI. c,d. MCLE images demonstrating the late-enhancement of inferior wall and posteromedial PM (arrows). The trigger delay was 770 ms for image C and 812 ms for image D.

Figure 4.

An anteroseptal MI with anterolateral PM involvement. a,b: MCLE images reveal the late-enhancement of anteroseptal wall and anterolateral PM. c: Four-chamber SSFP image showing the mild ischemic mitral regurgitation (arrow).

DISCUSSION

This study demonstrates a significant difference in the detection of papillary muscle involvement in patients with myocardial infarction using different LGE-MRI techniques, with improved identification of PM-MI using multicontrast late-enhancement MRI, compared with conventional IR-FGRE. Although both LGE-MRI techniques have shown a similar effectiveness in the demonstration of principal left ventricular myocardial infarction, the contrast between the blood pool and infarction is much better on LGE-MRI images using the MCLE pulse sequence.

The conventional LGE-MRI technique using IR-FGRE remains the cornerstone in the cardiac MRI examination of myocardial viability (19, 20). For PM involvement detection, the heterogeneous contrast uptake in posteromedial PM using 2D IR-FGRE methods might indicate the origin of ventricular arrhythmias (7) and three-dimensional IR-FGRE has been proposed for better identification of PM involvement in patients with mitral valve prolapse (21). Although the conventional LGE-MRI technique using T1 surfing applied before IR-FGRE has improved the optimal selection of TI and achieved good suppression of viable myocardium, thus improving the demonstration of infarct tissue (22), it is still very difficult to achieve complete blood pool suppression of the LV, thus making it hard to visualize the PM-MI. The newly developed MCLE pulse sequence applies SSFP readout immediately after an inversion pulse, which permits visualization of infarct tissue as an area of fast T1 recovery with the simultaneous nulling of blood pool and viable myocardium (15). At certain cardiac phases on MCLE images, complete blood pool suppression can be achieved. Although in anatomy the PM has a broad base with the left ventricular wall, it is loosely connected to the solid LV wall through trabeculae carneae and a significant portion extends into the LV cavity, as is indicated from a multidetector CT imaging study using a submillimeter high spatial resolution (14). The capability of complete blood pool suppression provided by MCLE improves the visualization of structures in the LV cavity, thus making it a more favored LGE-MRI technique to confirm the presence of PM-MI. The multicontrast capability of MCLE also facilitated the identification of subtle myocardial damage in experimental myocardial infarction as evidenced by whole-mount histology (23). Moreover, the MCLE sequence provided cine images that also facilitated the simultaneous appreciation of wall motion abnormalities in the region of myocardial infarction.

One advantage of using cardiac MRI to evaluate PM-MI is the ability to confirm the presence of the associated mitral regurgitation using a two- or four-chamber SSFP technique. Although the reported incidence of ischemic mitral regurgitation varies greatly according to the technique used, roughly 25% patients following MI will develop ischemic mitral regurgitation (2, 8). More importantly, even mild mitral regurgitation may yield a worse prognosis (24). The existence of severe ischemic mitral regurgitation may require surgical interventions such as annular ring reduction, asymmetrical annuloplasty, or even valve replacement to eliminate mitral regurgitation. PM damage can be determined using intra-cardiac ultrasound, but it is difficult to confirm the presence of PM-MI noninvasively by trans-thoracic or trans-esophageal echocardiography. Late-enhancement MRI using MCLE may provide an alternative to better demonstrate PM damage in patients with MI, and can explore the role of PM damage in the prediction or generation of ischemic mitral regurgitation following MI. In this study, mild and moderate ischemic mitral regurgitation was noticed in 52% of patients with myocardial infarction and PM involvement. Due to the small sample size of this study, the correlation between mitral regurgitation and PM involvement has not been explored. Further prospective studies are warranted to investigate the diagnostic and prognostic value of the LGE-MRI technique using the MCLE pulse sequence in the evaluation of PM involvement in ischemic mitral regurgitation.

Infarct tissue heterogeneity is closely related to the occurrence of ventricular arrhythmia due to the presence of re-entry routes for the slow conduction of electrical depolarization in myocardial infarction. Infarct heterogeneity identified on contrast-enhanced MRI is a strong predictor of spontaneous ventricular arrhythmia with subsequent implantable cardioverter-defibrillator (ICD) therapy (25). Several studies have also indicated the role of PM-MI in the production of ventricular arrhythmias (7). Although it is challenging to demonstrate PM-MI using conventional LGE-MRI technique, in this preliminary study we have shown that MCLE is a better method to identify its presence. Further electrophysiology studies would be required to examine the contribution of PM-MI as detected by MCLE as potential sources of ventricular arrhythmias.

In conclusion, multicontrast late-enhancement MR imaging provides better contrast between left ventricular blood pool and infarct myocardium, which improves the determination of papillary muscle involvement in myocardial infarction. This may help to identify patients in whom significant mitral regurgitation and/or ventricular arrhythmias may develop.

Acknowledgements

Dr. Alexander Dick was supported by a Heart and Stroke Foundation (HSF) Canada Phase 2 clinician scientist award and Dr. Kim Connelly was supported by a Heart and Stroke Foundation (HSF) Canada Phase 1 clinician scientist award.

Ancillary