2022 Volume 11 Issue 4
Creative Commons License

Cardiac Magnetic Resonance Imaging Feature Tracking for Quantifying Left Ventricle Deformation in Type 2 Diabetic Patients

 

Qurain Turki Alshammari1*, Wed Almutairi1,2, Eman Alshammari1,2, Ohood Alrashidi1, Meshari T Alshammari1,3, Amjad R Alyahyawi1,4, Yasser Alzamil1, SD Shahanawaz5, Shashi Kumar CG5

1Department of Diagnostic Radiology, College of Applied Medical Sciences, University of Hail, Hail, Saudi Arabia.

2Department of Diagnostic Radiology, Hail Health Cluster, Ministry of Health, Saudi Arabia.

3Translational Medical Sciences and National Institute for Health Research (NIHR) Nottingham Biomedical Research Centre, Nottingham University Hospitals NHS Trust and the University of Nottingham, Nottingham NG7 2UH, UK.

4Centre for Nuclear and Radiation Physics, Department of Physics, University of Surrey, Guildford, Surrey, UK.

5Department of Physical Therapy, College of Applied Medical Sciences, University of Hail, Saudi Arabia.

 

*Email: [email protected]


ABSTRACT

An evaluation of myocardial function is necessary to assess the condition of a patient's heart. Although ejection fraction is used as a measure of myocardial function, many common cardiac illnesses affect the myocardium locally rather than globally. It may therefore be possible to determine the motion of the regional wall of the heart by evaluating myocardial strain. Studying the efficacy of CMR feature tracking in diabetics Type 2 To determine whether the use of MRI feature tracking is effective in detecting myocardial strain deformations in diabetics and healthy individuals.

This study involved 80 patients, 40 of whom had Type 2 diabetes and 40 of whom did not, who were subjected to cardiac MRI and echocardiography. Post-processing was done offline using CMR feature tracking and echocardiography speckle tracking, followed by SPSS analysis of the data. T2DM patients and controls did not demonstrate any significant differences in the study. There was a significant decrease in the LV GLS and LV GRS values among T2DM patients (p 0.0001) as compared to normal controls. Where the LV GLS was in controls (-20±5) and diabetes (-15±2). The CMR-FT method is well suited for accommodating STE, providing additional information in clinical trials, which can be used interchangeably. This study recommends that feature tracking (FT-CMR) and speckle tracking echocardiography (STE-Echo) be incorporated into the clinical practices of hail hospitals due to their positive results in diagnosing myocardial deformation.

Key words: Cardiac magnetic resonance imaging, Feature tracking, Speckle tracking echocardiography, Type 2 diabetes mellitus


INTRODUCTION

Diabetes mellitus (DM) is one of the non-communicable chronic diseases and a major public health concern accompanied by high healthcare costs, morbidity, and early mortality.

This metabolic disorder has rapidly increased due to a lack of physical activity, poor food habits, obesity, and obesity [1]. According to the International Diabetes Federation (IDF), 537 million adults aged 20 to 79 are estimated to be affected by diabetes by 2021 - one out of every ten adults. This number is expected to increase to 643 million by 2030, and 783 million by 2045. There are 73 million diabetics in the Middle East and North Africa, and the number is expected to rise to 135.7 million by 2045 [2].

Diabetes mellitus is characterized by elevated blood sugar levels. This leads to several macrovascular and microvascular complications such as neuropathy, retinopathy, renal failure, and cardiovascular disorder [1].

Diabetes-related mortality and morbidity are primarily caused by cardiovascular complications. When diabetes is untreated or inadequately treated, it can result in overt diabetic cardiomyopathy. Myocardial dysfunction is the most common complication of diabetes, causing clinically silent myocardial dysfunction [3].

Diabetic cardiomyopathy can be categorized as either having coronary artery disease or hypertension [4]. In addition to advanced glycation end products, atherosclerosis, microinfarctions in the subclinical stage, mitochondrial dysfunction, and lipotoxicity, several metabolic impairments contribute to diabetic cardiomyopathy [5] These impairments may affect both left ventilatory function (LV) and right ventilatory function (RV), due to their systematic nature [6].

It is believed that cardiovascular diseases represent a significant international public health burden due to the difficulty of evaluating heart contractile function in modern cardiology, which continues to be an issue [6]. In fact, the conventional parameter that describes left ventricular (LV) function, ejection fraction (EF), has considerable limitations due to its volumetric nature, suboptimal reproducibility, and a loss of ability to reflect local LV function. As a result, LV mechanics can be more comprehensively characterized by measuring myocardial deformity (strain) non-invasively [7].

A myocardial strain is a deformation caused by the application of force to the myocardium. The rate at which the myocardial length changes from relaxed to contracted is called myocardial strain [8]. As opposed to strain, ejection fraction (EF) offers the possibility of examining components of contractions in longitudinal (LS), circumferential (CS), and radial directions (RS). Non-invasive methods to analyze myocardial deformation, such as cardiac magnetic resonance feature tracking (CMR-FT) and speckle tracking echocardiography (STE), are available in addition to cardiac magnetic resonance feature tracking [9].

Patients with type 2 diabetes have been subjected to echocardiographic studies to determine how much strain they are under. With cardiac magnetic resonance feature tracking (CMR-FT), a novel method has been developed for quantifying left and right ventricular myocardial deformation using high-resolution images [10-12]. Further, studies have shown that speckle-tracking echocardiography (STE) is a reliable, reproducible, and accurate method of measuring LV deformations in a wide range of clinical applications. A prospective clinical trial evaluating this technique has not yet been conducted in large populations. The lack of information about CMR-FT and STE for LV deformation assessment is largely due to this fact. Thus, this study aimed to determine if longitudinal, radial and circumferential strains in diabetic patients were correlated with features tracked through Cine SSFP Imaging and speckle tracking echo.

 

MATERIALS AND METHODS

 

The Institutional Ethics Review Board of the University of Hail and King Salman Hospital approved this study (No. H-2022-205) and in compliance with proper guidelines for medical research, we adhere to the Declaration of Helsinki (2000 edition). There were a total of 40 persons diagnosed with diabetes mellitus and 40 individuals who were healthy controls, who visited the radiology department at King Salman Specialist Hospital (KSSH) and underwent routine CMR-FT and STE examinations between December 2021 and March 2022 in the Hail region of Saudi Arabia.

A diagnosis of Type 2 Diabetes (T2DM) at age 18 was required, without any symptoms, signs, or history of heart disease (known coronary artery disease, cardiomyopathy, or valvular heart disease); sinus rhythm; normal ejection fraction over 50%; and absence of contraindications to MR imaging.

Participants with severe renal impairment (estimated glomerular filtration rate = 30 mL/min/1.73 mm2), uncontrolled blood pressure at rest (systolic blood pressure > 180mmHg and/or diastolic blood pressure > 100mmHg), or contraindicated for MRI were excluded.

 

Procedure

The research design is to analyze retrospectively MRI FT offline data, using the software of circle CVI-42 that measures the strain in a 17-segment software-generated mode, in a different direction.

Global longitudinal strain (GLS) is measured on the long axis (2,3 4 chambers views) while circumferential (GCS) and radial strains (GRA) are measured on the short axis without the need to use additional sequences and the same for echocardiography using the STE.

At end-diastole, the epi- and endocardial borders were defined on all pictures. Following the initiation of an automated calculation, the used software program automatically demarcated the boundary all through the cardiac cycle. The efficiency of the tracking and contouring is visually assessed and manually corrected as appropriate. To compute the left ventricular ejection fraction, further cine-CMR studies were performed (LVEF).

 

a)

b)

Figure 1. Myocardial strain is measured with FT-CMR and STE. Feature-tracking cardiac magnetic resonance imaging, a) and 2-D speckle-tracking echocardiography, b) are used for the analysis of myocardial segmentations and global longitudinal strains. A four-chamber view is called a 4CH; a two-chamber view is called a 2CH and a three-chamber view is called a 3CH.

 

Speckle tracking echocardiography (STE)

In this study, an experienced cardiologist utilized the Vivid E9 equipped with the M5S M5 3.5–5MHz transducer (GE Vingmed Ultrasound, Horten, Norway) for conventional two-dimensional Doppler ultrasonography on the patients. Each patient with attached ECG leads had an electrocardiogram performed. There were three consecutive image cycles of the apical 4-chamber, 2-chamber, and long-axis in apical structures for the data to be collected, and all three cycles were collected at a high frame rate (>45 s 1).

EchoPAC software (version 203, GE Vingmed Ultrasound, Norway) has been used as a measurement tool to measure global longitudinal strain and mean vertical wave (MW). Based on the timing of the events in the aortic valve spectrum, the time of aortic valve closure was determined. Therefore, the apical long-axis perspective, four-chamber perspective, and two-chamber perspective were analyzed using Ten, APLAX, A4C, and A2C per the study design. Upon dividing the LV myocardium into 18 segments, the program uses this information to automatically calculate the global longitudinal strain (GLS) (Figure 1).

 

Feature tracking CMR

An MR scanner with 1.5 Teslas was used to perform CMR imaging. We recorded 2- and 4-chamber CVs with ECG-gated b-SSFP cine sequences as well as a short-axis stack. As a result, we used the following parameters for scanning the cardiac cycle: 40 frames per cardiac cycle, 0.8mm x 0.8mm pixel spacing, 8mm slice thickness, 1.5mTR, and 3ms TE. In order to evaluate the LVEF, the SA stack was used.

For the measurement of strain in CMR-FT, commercial software available from "CVI" was used (CVI42, Version 5.6.5, Circle Cardiovascular Imaging Inc., Calgary, Canada) (Figure 2). An FT was performed at the diastole's end. An endocardial and epicardium boundary scan was conducted on the left ventricle. The RV boundaries could also be tracked using CV in the same way. Using tracking algorithms, tissue characteristics have been tracked during the cardiac cycle. The tracking accuracy was verified visually, and any modifications were made only to the initial outlines if necessary. Afterward, the process was repeated three times and the average was taken. A global longitudinal strain measurement (GLS) was performed on the LV and RV, along with a global circumferential strain measurement (GLS/GRS) on the LV. The GLS was determined using two, three, and four long-axis CVs and averaging peak strains. In contrast, the RV strain was exclusively derived from the 2-CV and 4-CV strains. Global short axis (SA) strains (GCS and GRS) were calculated from an average of three slices: basal slices (last slices showing the entire circular myocardium with no outflow tract), midventricular slices, and apical slices (blood pool maintained throughout the cardiac cycle).

 

Data analysis

Using SPSS, the data will be analyzed (v23). If relevant, data are reported as percentages or standard deviations with a confidence level of 95%. (CI). p-values less than 0.05 were deemed statistically significant.

Using linear regression analysis and Bland-Altman plots, the concordance with FT-CMR and STE myocardial strain measurements was determined. To determine if FT-CMR and STE12 exhibited considerable variation, a paired Student's t-test has been used. Inter-observer (two readers) and intra-observer (two readings) repeatability was determined by calculating the standard deviation of an estimate, coefficient of variation (CV), and intraclass correlation coefficient (ICC).

 

RESULTS AND DISCUSSION

 

The inclusion criteria were met by a total of 54 patients. Fourteen participants were removed from the trial as a result of low-quality pictures (n=5), tachycardia (n=6), or an irregular pulse (n=3). In the study overall number of 80 patients was first separated into two groups: healthy individuals (n=40, mean age 52±13 years, 20 men) and T2DM patients (n=40, mean age 54±11 years, 21 men).

There was a significant difference between the levels of arterial blood pressure, fasting blood glucose, and HbA1c in patients with T2DM compared to healthy controls (p0.05), however, the level of low-density lipoprotein was not significantly different (P>0.05). Patients with T2DM had significantly higher e', a′, and E/e′ values than healthy controls (p0.05). Compared to healthy controls and T2DM patients, there is a significant difference in the E and A variables (p>0.05). Patients with T2DM showed levels that were considerably lower than those in healthy individuals when compared to the values of LV GLS and LV GRS (p-0.001). In terms of LV GCS, there were no significant differences between normal LV GCS and diabetic LV GCS. This is summarised in Table 1. A significant difference was observed between FT-CMR and STE in GCS as well as in GLS (<0.001). Co-relation coefficient analysis revealed that there was a positive co-relation between FT-CMR and STE in GCS, GRS AND GLS respectively (<0.001) This is summarised in Table 2.

Table 1. Characteristics of STE and CMR-FT in type 2 diabetes mellitus subjects and control subjects.

 

Normal subjects (40)

T2DM (40)

p-value

Age (yrs)

52±13

54±11

0.36

Male%

20 (50)

21(55)

0.84

BMI (kg/m2 )

22±1.4

23±1.7

0.21

Heart rate (bpm)

72±9

75±11

0.34

LVEF%

66±4

60±3

<0.001

DBP (mmHg)

74±10

77±9

0.19

SBP (mmHg)

121±12

125±11

0.38

Fasting Plasma glycose (mmol/L)

4.97±0.6

14±3.9

<0.001

HbA1c (%)

5.18±0.5

10.20±2.09

<0.001

TCH (mmol/L)

3.8±0.7

3.9±0.6

0.62

TG (mmol/L)

1.1±0.2

1.1±0.5

0.34

HDL (mmol/L)

1.24±0.2

1.20±0.4

0.42

LDL (mmol/L)

2.1±0.5

2.2±0.6

0.17

ECHOCARDIOGRAPHY

 

 

 

E (m/s)

0.83±0.15

0.76±0.12

<0.001

A (m/s)

0.65±0.11

0.69±0.20

<0.001

E/A

1.31±0.29

1.23±0.31

0.08

e’ (m/s)

0.13±0.04

0.08±0.02

<0.001

a’ (m/s)

0.09±0.03

0.10±0.02

<0.001

E/e’

7.1±1.5

10.56±2.3

<0.001

Feature tracking CMR

 

 

 

LV GCS% (SAX)

-24±6

-22±5

0.32

LV GRS% (SAX)

22±11

16±7

<0.001

LV GLS% (LAX)

-20±5

-15±2

<0.001

 

Table 2.  Myocardial strain evaluating FT-CMR vs STE

Difference

 

FT-CMR

STE

Absolute

95%CI

p

LV GCS% (SAX)

-28±6

-27±5

-1.0 ± 0.8

1.2 to 0.6

<0.0001

LV GRS% (SAX)

25±11

21±7

4.0 ± 3.2

4.2 ± 12.3

0.04

LV GLS% (LAX)

-27±5

-21±2

6.0 ± 3.9

4.1 ± 4.2

<0.0001

Correlation

 

FT-CMR

STE

Intercept

95%CI

p

LV GCS% (SAX)

1.2

2.8

2.5

0.64

<0.0001

LV GRS% (SAX)

0.6

9.5

12.4

0.48

<0.0001

LV GLS% (LAX)

0.7

3.5

3.9

0.72

<0.0001

 

FT-CMR, Feature tracking cardiac magnetic resonance; STE, speckle tracking echocardiography; CI, confidence interval; SEE, Standard error of estimate; GLS, Global longitudinal strain; GRS, Global radial strain; GCS, Global circumferential strain.

BMI  Body Mass Index, SBP Systolic Blood Pressure, DBP  Diastolic Blood Pressure, TCH Total Cholesterol, TG Triglyceride, HDL High-Density Lipoprotein, LDL Low-Density Lipoprotein, LAD Left Atrial Diameter, LVEF  Left Ventricular Ejection Fraction, E  Peak velocity during early diastole of the anterior mitral valve, A peak velocity during late diastole of the anterior mitral valve, e′ peak early diastolic annular velocities using TDI, a′ peak late diastolic annular velocities using TDI, GCS (SAX) Global Circumferential Starin (Short axis), GRS (SAX) Global Radial Strain (Short axis), GLS (LAX) Global Longitudinal Stain (Long axis).