Principles and Clinical Applications of Feature-Tracking Cardiac Magnetic Resonance Imaging: A Literature Review

Cardiac magnetic resonance (CMR) has emerged as a reference tool for the evaluation of ventricular function. The calculation of left ventricular (LV) and right ventricular (RV) ejection fraction (EF) is the most widely used method to assess ventricular function [1]. However, changes in EF occur during the late stages, and it may be normal in regional myocardial involvement. Additionaly, the techniques used to calculate ventricular volume differ considerably between CMR (8–12 discs drawn from short-axis slices) and echocardiography (20 discs obtained from apical four- and two-chamber views) resulting in discrepancies and limited intermodal comparability [2, 3, 4]. In an attempt to better quantify global and regional ventricular functions, a number of techniques have been developed. Myocardial strain imaging is a promising technique to quantify ventricular functions. The current reference method to measure myocardial strain is two-dimensional (2D) speckle-tracking echocardiography (STE), which can assess global and regional functions and has superior reproducibility. However, STE has a lower signal-to-noise ratio (SNR) and is difficult to apply in poor echogenic windows, ultrasound dropouts, and reverberations [5].

CMR-based strain imaging techniques include tissue tagging, displacement encoding with stimulated echoes (DENSE), strain encoding (SENC) and feature tracking (FT). CMR tagging consists of a preparation phase in which magnetic labels (black lines) are superimposed orthogonally on the myocardium at the beginning of a cine sequence. Visual assessment of the deformation of these lines provide immediate information on regional wall-motion abnormalities. While it is the most validated CMR technique to assess myocardial strain, it suffers from low spatial resolution and tag fading, reducing its accuracy when applied to thin walls. DENSE encodes tissue displacement into the phase of an image. In this technique, three radiofrequency pulses generate an echo, and gradients encode displacement into the signal phase. DENSE also provides high-quality strain assessment but is limited by a low SNR and tag fading. SENC uses magnetization tags parallel to the image plane combined with out-of-plane phase-encoding gradients along the selected direction. Strain is directly related to pixel intensity in the resulting images. The main advantage of SENC include little need for post-processing and the main limitation is tag fading [6].

By comparison, CMR-FT is a new and faster method to determine regional and global myocardial deformation. It calculates the strain values from standard steady-state free precession (SSFP) cine images without requiring extra sequences [7]. CMR-FT-derived strain paraments have prognostic value in patients with ischemic or nonischemic dilated cardiomyopathies incremental to common clinical and CMR risk factors [8]. The present article provides a comprehensive review of the principles and clinical applications of CMR-FT. The strengths and limitations of CMR-FT are also discussed.

Myocardial strain or deformation refers to the systolic shortening of the myocardium. It is expressed as the percentage change in the myocardial length from resting (end-diastole) to contractile state (end-systole). For example, if the end-diastolic length is L0 and end-systolic length is L1, myocardial strain (S) is:

Two different strain-calculation methodologies have been developed: Lagrangian strain, in which the deforming myocardium itself acts as a reference and displacements are calculated at a fixed material point in the myocardium; and Eulerian strain, in which spatial coordinates are fixed and the tissue strain is calculated at a specific location in space. The imaging modalities calculate the Lagrangian strain [9, 10].

Due to the complex three-dimensional (3D) anatomy of LV musculature, myocardial deformation occurs along different directions. Three different types of strain can be calculated, namely longitudinal strain, radial strain and circumferential strain. Longitudinal strain refers to the shortening of the myocardium along the long axis, from the base to apex. Circumferential strain represents systolic shortening of the LV circumference. As both the longitudinal and circumferential strains describe systolic shortening, they are expressed with negative values. Radial strain refers to myocardial deformation toward the center of the LV cavity (Fig. 1). As it measures the thickening of the LV wall during systole, it is expressed with positive value. Several other parameters can be derived from strain analysis, such as strain rate and torsion, as shown in the schematic diagram. Strain rate refers to the velocity or rate at which deformation takes place [11]. Torsion is the twisting of the LV myocardium created by the clockwise rotation of the myocardial base and the counterclockwise rotation of the myocardial apex during systole. It is defined by the circumferential longitudinal shear angle generated by the rotation of the basal segments and the apex relative to a stationary reference point at a the midventricular level [12, 13].

Fig. 1
Schematic representing the left ventricle (LV) and myocardial deformation directions. The radial strain (R) refers to the myocardial deformation toward the center of the LV cavity. The circumferential strain (C) represents the systolic shortening of the LV circumference. The longitudinal strain (L) refers to shortening of the myocardium along the long axis, from base to apex. Torsion represents the twisting movement of the LV myocardium created by an overall clockwise rotation of the myocardial base and a counterclockwise rotation of the myocardial apex during systole.

• Click for larger image
• Download as PowerPoint slide

CMR-FT is an optical flow method of calculating myocardial strain. It identifies features in the image and tracks them in successive images of a sequence. The features are anatomic elements typically identified along the interface between the cavity and myocardium. The tracking involves defining a small square window around the feature and then searching for the most similar pattern in the following frame [14]. Several CMR-FT software packages can be directly applied to SSFP cine images.

The first step in post processing involves tracing the the endo- and epicardial borders in semiautomated manner. The software then automatically tracks the borders across the whole cardiac cycle (Fig. 2). Global radial strain (GRS) and global circumferential strain (GCS) are derived from short-axis SSFP cine images and global longitudinal strain (GLS) is derived from two or three long-axis cine image stacks (Fig. 3). Although FT was developed for 2D images, it can track 3D regions. However, there is limited experience with 3D applications [15].

Fig. 2
Semiautomated tracking of short-axis and long-axis orientation. The figure shows a representative example of tracking in short-axis and long-axis (horizontal four chambers and vertical two chambers) steady-state free precession (SSFP) images. The global radial strain and global circumferential strain are derived from short-axis SSFP cine images and global longitudinal strain is derived from two or three long-axis cine image stacks.

• Click for larger image
• Download as PowerPoint slide

Fig. 3
Strain analysis by feature tracking. A: Endocardial and epicardial borders traced manually at the end-diastolic phase. B: Global peak radial strain. C: Global peak circumferential strain. D: Torsion curves generated by feature tracking.

• Click for larger image
• Download as PowerPoint slide

The two major software packages are provided by Tomtec Imaging Systems and Circle Cardiovascular Imaging (CVI). Although the vast majority of studies on strain have used Tomtec software, CVI has recently introduced an alternative tool called tissue-tracking (TT). While FT follows distinctive characteristics at the endocardial- and epicardial-cavity borders to track myocardial deformation, TT uses a mid-surface curvilinear coordinate system to track LV deformation and follows the motion of software-generated myocardial nodes on SSFP cine sequences thorough the cardiac cycle. Therefore, it would be better to introduce both programs using different strategies to compute strain rather than simplify CMR-FT strategy as typically identified along with the interface of cavity and myocardium.

There are no clear cut-off values for myocardial strain in healthy subjects. A recent metanalysis analyzed the normal ranges of CMR-FT values for 659 healthy subjects reported in 18 papers. The pooled means of strain values were -20.1% (95% confidence interval [CI], -20.9% to -19.3%) for LV-GLS, -23% (95% CI, -24.3% to -21.7%) for LV-GCS, 34.1% (95% CI, 28.5% to 39.7%) for LV-GRS, and -21.8% (95% CI, -23.3% to -20.2%) for RV-GLS [16]. A normal reference range for various strain values is difficult to calculate because of variability among studies and different users. While GLS is the most robust deformation parameter for STE; GCS is more reproducible than GLS for CMR-FT. Large variations in radial range are seen between studies [17]. The proposed explanation for this variation states that changes in the voxel pattern during the cardiac cycle within the myocardium may affect the consistency of FT, particularly in the radial direction. For example, it is possible that the compaction and exclusion of blood from the interstices in a trabeculated myocardium at end-systole can alter voxel appearance enough to make accurate tracking difficult. The higher degree of trabeculations seen at the LV apical level compared with the basal level may account, in part, for the increased variability in measurements seen in this region [17].

Similarly, CMR-FT strain and torsion measurements are subject to considerable intervendor variability. Schuster et al. [18] conducted the first comparison of CVI and Tomtec software for inter-vendor agreement, and a high coefficient of variation of up to 45% for global torsion and 10% for GCS was seen for two repeated interobserver measurements.

Left atrial (LA) FT is also feasible and may enhance the diagnostic value of CMR in many diseases. LA function is divided into reservoir, conduit, and contraction phases. The normal values for LA-GLS can be derived during the reservoir (29%±5%), conduit (21%±6%), and atrial contraction phases (8%±3%) [19]. An increase in atrial contraction values is seen in elderly subjects, consistent with the physiology of aging [20].

Recent studies have demonstrated the value of RV strain in diseases such as pulmonary hypertension, pulmonary embolism, RV heart failure, and congenital heart disease [21]. RV strain has also been found to have an independent prognostic value when compared with LV strain alone [22]. Few studies defining age-specific normal ranges of RV strain are available. Application of the technique to the RV is still limited and further vendor-specific reference values are required before its routine clinical utilization [23, 24].

The systolic and diastolic performance of the heart depends upon the integrity and coupling of subendocardial and subepicardial fibers. Contraction of subendocardial fibers results in longitudinal shortening, while contraction of subepicardial fibers leads to circumferential shortening. Radial thickening is the result of contributions from both fibers [25].

Subepicardial fibers are also the principle contributor to myocardial rotation due to their large radius of rotation. Most myocardial disease affects the subendocardium first, resulting in early reduction of longitudinal strain. Circumferential strain and torsion remain relatively well preserved or may show compensatory increases to preserve stroke volume and EF. Transmural involvement results in simultaneous subendocardial and subepicardial dysfunction, which impairs all the parameters. A classic example is acute transmural infarction, which affects both longitudinal and circumferential strains as well as torsion, and ultimately results in reduced EF [14].

CMR-FT–derived parameters have also shown prognostic significance. In a study of large multicenter cohort, 1012 patients underwent CMR for evaluation of ischemic or nonischemic cardiomyopathy. The authors found that a 1% worsening in GLS was associated with an 89.1% increase in the risk of death, after adjusting for clinical and imaging risk factors, including EF and late gadolinium enhancement (LGE) (hazard ratio [HR], 1.891 per %; p<0.001). GLS was independently associated with death after adjusting for clinical and imaging risk factors (including EF and LGE) in both ischemic and nonischemic dilated cardiomyopathy subgroups [26]. The utility and applications of CMR-FT in specific clinical scenarios are discussed below.

Both global and regional strain parameters have been widely investigated in patients with ischemic heart disease (IHD). An infarcted myocardium shows impaired strain parameters, and myocardial strain values are inversely related to the area at risk, infarct size, and infarct transmurality [27, 28, 29, 30]. Segmental strain allows more precise discrimination between subendocardial and transmural infarction in comparison with visual analysis of changes in wall thickness and wall motion abnormalities [27]. Studies have validated CMR-FT against gold-standard myocardial tagging in evaluations of both global and segmental strain after acute myocardial infarction [29].

Dobutamine stress CMR (DS-CMR) is an established tool to assess myocardial ischemia and hibernating myocardium [31]. Image analysis is based on visual assessment and varies considerably with operator. CMR-FT can evaluate inducible ischemia and contractile reserve in patients with chronic IHD. Schuster studied 10 healthy subjects to determine the feasibility and reproducibility of CMR-FT for quantitative wall-motion assessment with intermediate dosing of dobutamine. CMR-FT reliably detected quantitative wall motion, and the strain derived from SSFP cine imaging corresponded to inotropic stimulation [32]. In an another study, Schuster et al. [33] investigated the feasibility of CMR-FT to assess myocardial viability in patients with ischemic cardiomyopathy. The authors studied 15 patients with ischemic cardiomyopathy referred for viability assessment at 3 T at rest and during periods of low-dose DS. Dysfunctional segments without scars showed improvements in all three strain parameters after administration of low-dose dobutamine. There was no response to dobutamine in dysfunctional segments with scar transmurality above 75%. Circumferential strain improved in all segments up to a transmurality of 75%. Radial strain improved in segments with <50% transmurality and remained unchanged in segment with transmurality >50%. The study concluded that CMR-FT holds promise for quantitative assessments of viability in patients with ischemic cardiomyopathy [33].

CMR-FT has also been studied in detecting myocardial ischemia. Schneeweis et al. [34] studied 25 patients with suspected or known coronary artery disease (CAD) who underwent a standardized high-dose DS-CMR protocol at 1.5 T. None of the patients had wall-motion abnormalities at rest or scar tissue. All patients underwent an X-ray coronary angiography, which served as the reference standard. The earliest sign of ischemia detectable by circumferential strain occurred during intermediate-dose DS. With an increase in the dose from intermediate to high levels, the strain did not increase further in normal segments but the differences in the strain value between stenotic and normal segments became even more pronounced. The study found that a change in the strain values at intermediate dose may improve early detection of ischemia and increase patient safety during pharmacologic stress testing [34].

LGE is a predictor of all-cause mortality, cardiovascular mortality, and major adverse cardiovascular events (MACEs), independent of LVEF [35]. Recent studies have shown that various CMR-FT–derived parameters can predict clinical outcomes after myocardial infarction. A study by Buss et al. [36] found that CMR-FT–derived values of myocardial strain can estimate functional recovery in patients with first-time ST-elevation myocardial infarction (STEMI). CMR imaging was performed in 74 consecutive patients two to four days after successfully reperfused STEMI, using a 1.5 T scanner. Peak systolic circumferential and longitudinal strains were measured using SSFP cine sequences and compared to infarct size as determined by LGE. Follow-up CMR at six months was performed. A cut-off value of -19.3% for GCS identified patients with a preserved EF ≥50% at follow-up, with a sensitivity of 76% and a specificity of 85%, which was superior to that provided by longitudinal strain and non-inferior to that provided by LGE. The study concluded that CMR-FT can objectively assess infarct size without contrast agent and can estimate functional recovery with accuracy comparable to that provided by LGE [36]. Gavara et al. [37] who evaluated the prognostic value of strain using CMR-FT after STEMI, investigated 323 patients who underwent CMR one week post-STEMI. Global and segmental longitudinal, circumferential, and radial strain were assessed. During a median follow-up of 36 months, 54 cases of first MACEs were documented. All strain parameters were correlated with the incidence of a composite endpoint, including cardiac death, readmission for heart failure, and reinfarction. After adjusting for baseline and CMR variables, GLS (HR, 1.21; 95% CI, 1.11–1.32; p<0.001) was the only independent predictor associated with MACE. In particular, the MACE rate was higher in patients with a GLS ≥-11% (22% versus 9%; p=0.001) [37].

Another study documenting the prognostic role of CMR-FT in patients with STEMI was done by Nucifora et al. [38], who investigated 180 patients admitted with first STEMI. CMR with LGE imaging was performed to assess LV function, infarct size, and microvascular obstruction. FT analysis was applied to calculate the LV GCS. Patients were followed for a median of 95 months and outcome event was a composite endpoint that included aborted sudden cardiac death, cardiovascular death, and hospitalization for heart failure. After adjusting for other clinical and CMR imaging characteristics, LV GCS remained significantly and independently associated with the outcome event [38].

Yoon et al. [39] investigated the role of CMR-FT–derived global LV strain to estimate infarct size and clinical outcomes in patients with acute myocardial infarction. They demonstrated that all three global strain parameters (GRS, GCS, and GLS) were significant predictors of adverse cardiac events. In particular, GLS was independently associated with increase in the risk of adverse events by a factor of more than five, even after adjusting for LVEF and infarct size [39].

In summary, CMR-FT can be used effectively in ischemic cardiomyopathy to identify ischemic territories, infarct size, prognostication, and prediction of MACE.

Left-ventricle dysfunction related to dilated cardiomyopathy (DCM) or CAD can be distinguished by CMR on the basis of gadolinium-enhancement patterns within the myocardium. The classical subepicardial or midwall enhancement of idiopathic DCM reflects the presence of fibrosis and is a negative prognostic predictor. LGE is unlikely to detect diffuse microscopic fibrosis as no enhancement is seen in a subset of patients with DCM. A study by McCrohon et al. [40] showed that LGE may be absent in up to 60% of DCM patients. This type of data necessitates the development of new prognostic predictors which can stratify DCM patients beyond LVEF and LGE. Recent studies have shown that CMR-FT may be able to play a useful role in the risk stratification of non-ischemic cardiomyopathies, including DCM (Fig. 4) [41, 42, 43].

Fig. 4
Cardiac magnetic resonance feature tracking in a 45-year-old female patient with idiopathic dilated cardiomyopathy: (A) short axis and (B) two-chamber, long-axis, phase-sensitive, inversion-recovery images show subepicardial (anterior, anterolateral, inferolateral and inferior) (yellow arrows) and transmural (red arrows) type of late gadolinium enhancement pattern; (C) a vertical long-axis steady-state free precession image with delineated endocardial and epicardial contours and a corresponding (D) global peak longitudinal curve. The peak global longitudinal strain is severely reduced (-4).

• Click for larger image
• Download as PowerPoint slide

Buss et al. [41] evaluated the prognostic impact of CMR-FT in patients with non-ischemic DCM compared with LGE quantification and LVEF. A total of 210 subjects with DCM were evaluated with CMR. The predefined primary endpoints were cardiac death, heart transplantation, and aborted sudden cardiac death. During the median follow-up period of 5.3 years, GLS and mean longitudinal strain were independent prognostic parameters that surpassed the value of global and mean LV radial and circumferential strain, as well as LVEF, LGE mass and N-terminal pro-brain natriuretic peptide. A GLS greater than -12.5% predicted outcomes even in patients with LGE (p<0.001) and LVEF <35% (p<0.01). The study concluded that LV longitudinal strain assessed by CMR-FT is an independent predictor of survival in DCM and offers incremental information for risk stratification beyond standard CMR, clinical parameters, and biomarkers [41]. However, a recent study failed to confirmed these results. This study aimed to investigate the prognostic value of LV myocardial strain and LGE on CMR imaging in patients with idiopathic DCM and reduced EF (<40%). CMR images were evaluated for ventricular function, presence and extent of LGE, and LV myocardial strain. The primary outcome was a composite of all-cause death and heart transplantation. During a median follow-up of 47 months, LV strain showed no significant association (HR, 1.048; 95% CI, 0.945–1.163) with primary outcome, while LGE (HR, 4.73; 95% CI, 1.11–20.12) and serum sodium (HR, 0.823; 95% CI, 0.762–0.887) were independently associated with the primary outcome [42].

Other studies have also investigated the long-term prognostic significance of CMR-FT–derived GLS in heart transplant recipients. GLS was found to be associated with the long-term risk of death or MACE after adjusting for clinical and CMR risk factors. Furthermore, each 1% worsening in GLS was independently associated with a 15% greater risk of cardiovascular events [43].

CMR is a standard tool for the diagnosis and evaluation of hypertrophic cardiomyopathy (HCM). Two CMR based diagnostic features of HCM include increased wall thickness and the presence of LGE. Recent studies have shown that the functional abnormalities extend beyond the areas of LGE. Similarly, abnormal systolic strain values are seen in hypertrophied segments, irrespective of the presence of LGE [44]. These findings have necessitated the discovery of new diagnostic parameters to improve the diagnostic accuracy. CMR-FT has emerged as a promising technique for diagnosing HCM and it provides incremental value over traditional CMR-based methods.

Vigneault et al. [45] evaluated myocardial strain and circumferential transmural strain differences (cTSDs; the differences between epicardial and endocardial circumferential strain) in a genotyped cohort with HCM and explored correlations between cTSD and other anatomic and functional markers of disease status. CMR-FT identified myocardial dysfunction, not only in subjects with overt HCM, but also in carriers of sarcomere mutations without LV hypertrophy (LVH), suggesting a potential role for myocardial strain measurement in identifying early disease expression in these individuals [45].

Evaluation of HCM in young patients is limited by a lack of age-specific norms for wall thickness in CMR images. LV strain analysis may have a role to play in identifying and risk-stratifying pediatric patients with HCM. Bogarapu et al. [46] evaluated the utility of a novel CMR-FT technique for measuring LV strain and strain rate on SSFP cine images in the detection of myocardial fibrosis in pediatric HCM patients. A total of 29 HCM patients were investigated. Global longitudinal, circumferential. and radial strain and strain rate were lower in LGE-positive patients compared with LGE-negative patients. A GLS of ≤12.8 had a 91% sensitivity and 89% specificity for detection of LGE. This study concluded that CMR-FT may be useful in the identification of myocardial fibrosis and risk stratification of pediatric HCM without the use of contrast agents [46].

The common differential for HCM is hypertensive heart disease (HHD). Both diseases have LVH, focal LGE, increased native T1, as well as diastolic dysfunction in common. Taking potential co-occurrences of hypertension and HCM into account, diagnoses based on CMR alone are often uncertain. CMR-FT is an emerging method to differentiate between these two diseases. Neisius et al. [47] investigated the role of CMR-FT in differentiating these two entities. GLS was significantly higher in HCM patients compared to HHD and contols (-14.7±3.8 [HCM] versus -16.5%±3.3% [HHD] versus -17.2%±2.0% [controls]). GLS was associated with LV mass index and LVEF. The study concluded that CMR-FT–derived GLS can diffrentiate HHD and HCM with a capacity similar to that of LVH and fibrosis imaging markers [47].

Acute myocarditis is a frequent cause of sudden cardiac death in young patients, and up to 30% of biopsy-proven cases develop a secondary dilated cardiomyopathy with a poor prognosis [48]. Diagnosis of acute myocarditis remains challenging due to variable clinical presentations. The clinical diagnostic criteria as well as the current gold standard, endomyocardial biopsy, are of limited value due to their low sensitivity and specificity [49]. CMR imaging of myocarditis is based on the Lake Louise Consensus Criteria. However, the Lake Louise criteria exhibit varying sensitivity and specificity, and many patients with low levels of inflammation demonstrate an inconspicuous CMR result with no positive criteria [50]. Additional diagnostic parameters are therefore needed to improve the noninvasive diagnosis of acute myocarditis. The introduction of CMR-FT offers potential advantages in diagnosing acute myocarditis based on routinely acquired cine magnetic resonance images.

Baeßler et al. [51] investigated the potential role of CMR-FT-based strain analysis of both ventricles in diagnosing acute myocarditis, particularly in patients with preserved LVEF. They assessed 31 clinically suspected and confirmed CMR cases as well as 14 patients with a clinical diagnosis but inconspicuous CMR. Twenty healthy volunteers were included as a control. Analysis of global longitudinal, circumferential, radial strain and the strain rate of both ventricles was performed. A combined cut-off of -0.53 s^–1 for the basal RV circumferential strain rate and -29.0% for the LV circumferential strain allowed for a classification of acute myocarditis patients with preserved EF with a sensitivity of 89% and a specificity of 80%. The study pointed to a discriminative power, particularly for RV strain analysis in the CMR-based diagnosis of acute myocarditis [51]. In an another study, Dick et al. [52] evaluated the feasibility and reproducibility of CMR-FT–derived strain and strain rate. In logistic regression and receiver operating characteristic analyses, the peak early negative LA strain rate proved to be the best independent predictor of acute myocarditis (area under the curve=0.80). The study provided evidence of the discriminative power of atrial strain analysis in diagnosing acute myocarditis [52].

The addition of CMR-FT–derived strain parameters has been shown to increase the sensitivity of diagnosing acute myocarditis on CMR. Baeßler et al. [53] found that a multiparametric imaging model, including the novel T2-mapping derived parameter, the FT-derived strain parameter, and LGE yielded superior diagnostic sensitivity in suspected acute myocarditis when compared with any single imaging parameter.

Arrhythmogenic right ventricular cardiomyopathy (ARVC) is characterized by progressive fibrofatty degeneration, which preferentially affects the RV myocardium [54]. The diagnosis of ARVC is based on task force criteria that was revised in 2010 to improve diagnostic accuracy [55]. The original criteria was prone to significant inter-observer variability due to qualitative severity grading for systolic dysfunction and ventricular dilatation. The revised criteria define specific cut-off values for end-diastolic volume and EF, but the inclusion of quantitative metrics relies on subjective image interpretation. Strain imaging has been proposed as a less observer-dependent quantitative estimate for global and regional myocardial contraction [56].

Vigneault et al. [57] used FT to analyze regional wall-motion abnormalities in patients with ARVC. They found that strain was significantly impaired in overt ARVC compared with control subjects, both globally (p<0.01) and regionally (all segments of horizontal long axis view, p<0.01). A subtricuspid RV strain of less than -31% and RV end-diastolic volume index were significant predictors of disease presence [57]. Some studies have revealed a role for CMR-FT in the early detection and objective quantification of contraction abnormalities in ARVC. Heermann et al. [58] examined whether CMR-FT can serve as a quantifiable measure to confirm global and regional ventricular dysfunction in ARVC patients in support of the early detection of ARVC. They enrolled 20 patients with ARVC, 30 with borderline ARVC and 22 with a positive family history but no clinical signs of ARVC. Ten healthy volunteers were selected as controls. RV GLS rates in ARVC (-0.68±0.36 s^–1) and borderline ARVC (-0.85±0.36 s^–1) were significantly reduced compared with healthy volunteers (-1.38±0.52 s^–1, p≤0.05). CMR-FT could differentiate between manifest or borderline ARVC and healthy volunteers, even if RVEF was normal [58].

Involvement of LV has been increasingly recognized across a broad spectrum of ARVC severity. The early detection of LV involvement is of clinical importance because biventricular dysfunction is a stronger predictor of adverse outcomes, including sudden cardiac death and heart failure, compared with isolated RV disease. CMR-FT can help to assess LV dysfunction in ARVC patients with preserved LVEF. A study by Chen et al. [59] reported that CMR-FT is a novel method to assess global and regional myocardial contraction that may detect early LV dysfunction more easily than LVEF in ARVC.

Echocardiography is the primary imaging modality for serial follow-up imaging of postoperative patients with congenital heart disease. However, it is limited by poor acoustic windows in adults and complex postoperative anatomy. CMR provides comprehensive anatomical and physiological information in various congenital heart diseases [60]. CMR-FT–derived strain parameters offer additive value in predicting outcomes in comparison with conventional measurements of ventricular function and volume. Three studies have explored the prognostic value of CMR-FT–derived strain parameters in predicting ventricular tachycardia (VT) and sudden cardiac death in repaired cases of tetralogy of Fallot (TOF). Moon et al. [61] examined a case-control cohort that included 16 cases of repaired TOF. Patients with TOF who died or had VT were compared with age-matched patients with TOF with no adverse outcome. The results showed lower values of both RV and LV longitudinal strain in patients with repaired TOF who experienced adverse outcomes. The impaired longitudinal strain of both ventricles was strongly associated with adverse clinical outcomes [61]. Orwat et al. [62] conducted a prospective study of 372 patients to examine the relation between CMR-FT–derived myocardial deformation parameters and clinical deterioration in patients with repaired TOF. The results showed a significantly decreased LV circumferential strain and RV longitudinal strain in the event group when compared with the no-event group [62]. Another study conducted by Hagdorn et al. [63] showed that LV systolic circumferential strain rate is an independent predictor for the development of VT in patients with repaired TOF. Contrary to the previously reported studies, ventricular strain parameters did not predict deterioration of ventricular function in the studied population. The change in the strain-rate variables, even in the absence of affected strain parameters, highlights the superiority of strain-rate parameters in repaired TOF, rather than absolute proportion of deformation (strain) or volume change (EF). Second, in contrast to the RV variable, LV systolic circumferential strain rate has emerged as a prognosticator of VT. This could be because of mechanical interventricular coupling, by which an RV pressure load leads to decreased LV function. The study supports paying more attention to the LV in right-sided heart diseases [63].

The role of CMR-FT has also been investigated in children and adults with Fontan circulation [64]. Meyer et al. [65] evaluated the relationship between ventricular strain and cardiac systolic function in 51 patients with Fontan circulation. CMR and concomitant clinical assessment was performed at the start of the study and after two years. It was seen that GLS and GCS decreased over two years’ time while EF, cardiac index, and clinical parameters were preserved. The study concluded that CMR-FT-derived strain values may be a sensitive early indicator of systolic dysfunction [65]. The technique has also been used to study altered LV mechanics after the repair of a coarctation of aorta. Reduced global longitudinal and radial strains have been seen late after the effective repair of aortic coarctation with normal LVEF, suggesting that GLS may be an indicator of early LV dysfunction [66].

Diastolic dysfunction is the main cause of heart failure in patients with preserved ejection fraction (HFpEF) [67]. CMR-FT analysis has shown diagnostic role in HFpEF as well [68]. A retrospective study evaluated the role of CMR-FT in predicting the all-cause mortality in amyloid light chain (AL) amyloidosis. In a cohort of 76 newly diagnosed AL amyloidosis patients who underwent CMR for morphologic and functional evaluation CMR-FT–derived strain paramets were associated with all-cause mortality and added incremental prognostic information over biomarkers [69].

A bicuspid aortic valve (BAV) is commonly seen in asymptomatic young adults with normal or mildly impaired valve function. The latency period between the time of diagnosis and the onset of symptoms is often long and is characterized by progressive valvular obstruction and myocardial pressure overload. Various imaging techniques have been proposed to improve the early detection of myocardial dysfunction, such as tissue doppler imaging, speckle tracking, myocardial tagging, and T1 mapping. CMR-FT analysis is an alternative method for the early detection and risk stratification of LV diastolic dysfunction in minimally symptomatic or asymptomatic patients with BAV [70].

CMR-FT has also show promising role in identifying chemotherapy-induced cardiotoxicity. In a prospective longitudinal study of breast cancer patients treated with trastuzumab, significant temporal changes were seen in peak systolic strain at 6 and 12 months, correlating with reduction in LVEF. The results suggest that CMR-FT–derived systolic strain may identify the earliest signs of myocardial injury during trastuzumab therapy [71].

LA function has been the subject of study for many years. Ultrasound doppler and CMR derived volumetric approaches were the initial methods used to assess LA function [72]. Recently, CMR-FT has emerged as a promising technique for evaluating LA function. CMR-derived strain parameters are more sensitive than indexed LA volume for evaluating LA function [73]. A study by Kowallick et al. [19] found that CMR-FT reliably quantifies the strain and strain rate with high reproducibility on intra- and inter-observer bases. A differential pattern of LA function was also found in different diseases. For example, patients with HCM showed increased contractile function compared with healthy controls, while reduced atrial contractility was seen in patients with diastolic dysfunction and preserved LVEF [19]. Furthermore, LA dysfunction may contribute to cerebrovascular events in patients with atrial fibrillation (AF). CMR-FT–derived LA function reportedly improves risk stratification of cerebrovascular events in AF patients [74].

The optimum placement of LV leads is essential when pursuing the primary therapeutic target of cardiac resynchronization therapy (CRT). Guidelines recommend the pacing of the LV at the latest activated wall. The presence of scarring at the LV pacing site is associated with ineffective pacing and potentially proarrhythmogenic effects [75]. Several studies have demonstrated the role of CMR-derived anatomy and scar identification in determining the optimal site for LV pacing [76, 77]. The utility of FT-derived strain parameters in determining areas of late mechanical activation before CRT placement has also been investigated. Studies have shown that combined CMR-FT and LGE imaging is associated with marked LV reverse remodeling and superior clinical outcomes compared with deploying the lead over scarred and/or earlier activated segments [78].

CMR-FT–derived strain parameters have also been used to determine the consequences of pre-term birth in adult age. A study by Lewandowski et al. [79] found that young adults born pre-term have a unique adverse LV structure and function in terms of increased mass, shorter length, reduced volume, and apical displacement. Even if the LVEF was similar to those of subjects born at term, both longitudinal diastolic (peak strain rate and velocity) and systolic (peak strain, strain rate, and velocity) function as well as rotational (apical and basal peak systolic rotation rate) parameters were significantly lower [79].

Acceptance of the importance of multiparametric cardiovascular imaging in the diagnosis of ischemic and non-ischemic cardiomyopathies is growing. Scatteia et al. [6] investigated the relationship between LV strain and extracellular volume (ECV) and their collective impact on a mixed cohort of patients with ischemic and non-ischemic cardiomyopathies. The results showed an intrinsic link between altered CMR tissue properties and impaired myocardial mechanical performance. Decreased GLS, GCS and GRS were associated with increased ECV and native T1 and decreased post-contrast T1 in a dose-dependent manner. Event-free survival was consistently lower among those with reduced strains and increased ECV. The combined impact of ECV and strain emphasizes the complexity of myocardial tissue and function in the pathophysiology of heart failure. The study highlights the importance of multiparametric assessment of both tissue properties and mechanical performance in risk stratification [6].

No additional sequence acquisition is required for CMR-FT, saving time and resources. Since, it uses standard SSFP cine images, it can be applied to all routine scans. Furthermore, CMR-FT is a rapid and semi-automated technique with a short postprocessing time [80]. There are few limitations. One basic assumption about tissue tracking in 2D images is that the apparent in-plane deformations of gray-scale distributions correspond to actual displacements. This is not necessarily the case. For example, through-plane displacement of tapering, obliquely angulated forms could be misinterpreted as in-plane deformations in a 2D image. A second assumption is made about the coherent deformable nature of the myocardium, whch is not true for all regions. Although the use of relatively large interrogation windows can help to address this issue, whether it is effective or only apparent has yet to be established. It is also assumed that the measured deformation is the actual motion of the myocardium only. However, blood motion can affect tracking closer to the endocardium, resulting in unrealistic results. CMR-FT is also technically limited by pixel size; if the displacement of a feature is less than the pixel size, it may not be detected [9]. Finally, each of these methods of strain analysis require appropriate temporal resolution. The expected temporal resolution is 30–40 ms (at a heart rate of 60 bpm) for CMR-FT, 35 ms for tagging, 25 ms for SENC, 17 ms for DENSE, and 12.5–2 ms for STE [81, 82, 83, 84, 85].

CMR-FT is a relatively new and convenient post-processing tool that analyses the myocardial deformation without requiring extra sequences or contrast admininstration. A growing body of evidence indicates that it can detect cardiac disorders at the subclinical level, providing a window of opportunity to alter patient management before the advent of pronounced dysfunction. Further technical improvements and standardization of post-processing and interpretation are required to achieve inter-vendor and inter-organizational compliance.