The longitudinal magnetization relaxation time of the myocardium, so-called T1 time, is highly sensitive to processes that increase the interstitial space and can be quantified with various techniques [7]. One of the most commonly used in clinical practice is the modified Look-Locker pulse sequence where multiple single-shot images are acquired intermittently in diastole during 9–17 cardiac cycles and the inversion recovery curves are generated (Fig. 1, panels A and B). The T1 time can be obtained for any myocardial segment and T1 maps can be generated by determining the T1 time at each pixel location (Fig. 1, panel C). Three T1 mapping-derived metrics have been proposed as markers of increased myocardial fibrosis: the native T1 time, the post-contrast T1 time and the myocardial extracellular volume (ECV). With the increase of interstitial fibrosis, the native T1 values (without the use of gadolinium contrast) become longer whereas the post-contrast T1 values become shorter. By combining them, myocardial ECV fraction can be computed, which quantifies the extracellular matrix space. In the absence of amyloid deposition or edema, collagen I is the main component of the extracellular matrix space and therefore the myocardial ECV fraction is considered a robust marker of myocardial fibrosis [8‐10]. The added value of these metrics over LGE is the ability to quantify the degree of fibrosis and, particularly, to detect diffuse interstitial fibrosis, often associated with early stages of the disease.

However, it should be noted that the cut-off values of the T1 mapping-derived metrics to define fibrosis cannot be currently established since the values show considerable overlap in normal and diseased myocardium [11]. Moreover, neither of the techniques is entirely specific to myocardial fibrosis; abnormal myocardial ECV fraction can be observed in infiltrative diseases (i.e., amyloidosis) and edema, while native T1 values may also be altered in iron deposition and diffuse fat infiltration [12]. Furthermore, standardization of CMR T1 mapping techniques is necessary to obtain reproducible measurements across different vendors and institutions.

LGE CMR is considered the reference standard to quantify myocardial replacement fibrosis and scar. The increased extracellular space and decreased capillary density of the fibrous tissue result in increased volume of distribution and prolonged wash-out of gadolinium in comparison to the normal myocardium [13]. 10–20 min after intravenous administration of gadolinium, inversion recovery images are acquired in mid to late diastole. The inversion time is chosen to null the normal myocardium and provide the best tissue contrast between fibrous tissue, which appears bright, and normal myocardium, which appears black. Distinct patterns of LGE have been described in various cardiac diseases and associated with adverse prognosis [14‐19] (Fig. 2).

Molecular magnetic resonance imaging with the use of collagen-specific contrast agents is a new experimental method for the assessment of myocardial fibrosis. These novel contrast agents have shown to improve visualization of scar and perfusion defects in animal models of myocardial infarction [20, 21]. Furthermore, an elastin/tropoelastin-targeting contrast agent has provided interesting insights into the pathophysiology of remote myocardium extracellular matrix remodeling in a mice model of acute myocardial infarction [22]. Several other molecular probes have been synthesized to study individual processes involved in fibrosis formation, like necrosis, apoptosis, inflammation and scar maturation [23]. Further efficacy and safety studies are needed before clinical implementation. However, the current evidence is promising for future improvements in fibrosis detection and monitoring of molecular processes associated with myocardial remodeling.

This method is based on alteration of the myocardial tissue magnetization to create trackable markers within the myocardium which are visualized as dark lines in the form of a grid pattern. This allows immediate visual assessment of myocardial deformation, but for a more objective approach and quantification additional post-processing is employed. Recent developments in pulse sequences and image processing have resulted in a plethora of new tagging techniques [26]. The main advantage of CMR tagging over feature tracking CMR is that the imposed tags are more clearly defined and easier tracked than the natural features and are not subjected to through plane displacements, thereby providing more reproducible measurements [27]. The main shortcomings of this technique are the need for additional, elaborate scan sequences with limited accuracy when applied to thin myocardium (such as the remodeled, thinned-wall LV, the right ventricle and the atria) and the time-consuming post-processing.

Feature tracking CMR is based on post-processing of standard steady state free precession cine images, similar to echocardiographic speckle tracking. Feature tracking CMR algorithms focus on the endo- and epicardial borders and detect the in- and outward motion of the cavity-tissue interface [27, 28]. Global and segmental LV longitudinal, circumferential and radial strain, strain-rates, and LV rotational mechanics can be derived from standard long- and short-axis views (Fig. 3). Global rather than segmental strain values appear the most reproducible [29‐31]. Additional methodology standardization is an important prerequisite for wider dissemination of this technique in clinical practice.

The pressure overload caused by AS increases LV wall stress and as a consequence the myocardium responds with myocyte hypertrophy to maintain LV systolic function. This myocardial hypertrophy is characterized by an increased muscle fiber diameter with parallel addition of new myofibrils [37]. Furthermore, there is an increase of interstitial fibrosis and myocyte apoptosis, partially as a consequence of oxygen supply–demand mismatch and myocardial ischemia [37‐39]. At a late stage in the natural history of severe AS, the LV myocardium is characterized by large areas of myocyte loss and replacement fibrosis causing LV systolic dysfunction and associated with poor prognosis [38].

The early changes in the interstitial space with increased deposition of collagen I can be assessed with CMR T1 mapping (Table 2) [8, 34, 40‐46]. Several studies have validated LV native T1 values and myocardial ECV fraction against histology in patients with AS undergoing aortic valve replacement [8, 34, 40, 41]. In 109 patients with moderate and severe AS, Bull and colleagues showed that LV native T1 values were significantly higher among patients with symptomatic severe AS compared with moderate and asymptomatic severe AS (1014 ± 38 vs. 955 ± 30 and 972 ± 33 ms, respectively; p < 0.05) (Fig. 4) [40]. A significant correlation was observed between native T1 values and collagen volume fraction assessed on myocardial biopsies (R = 0.65, p = 0.002). Similarly, Flett and coworkers validated the measurement of myocardial ECV fraction in 18 patients with severe AS [8]. ECV strongly correlated with the histological collagen volume fraction (R² = 0.86; p < 0.001). Although still not implemented in routine clinical practice, the measurement of myocardial ECV in patients with AS has important clinical implications [34, 43‐46]. Increased ECV has been associated with symptoms, worse LV systolic and diastolic function, higher levels of cardiac troponin T and ECG strain [34, 43‐46]. Recently, Chin et al. reported the prognostic implications of myocardial ECV fraction corrected for LV end-diastolic myocardial volume normalized to the body surface area (iECV) in 166 patients with mild to severe AS [34]. Patients with increased myocardial iECV (≥22.5 ml/m²) but without LGE (replacement fibrosis) showed significantly higher all-cause mortality and AS-related mortality rates (36 per 1000 patients-year for both) as compared to the patients with normal myocardium (iECV < 22.5 ml/m², 8 and 0 deaths/1000 patient-years) (Fig. 5).

LGE, myocardial replacement fibrosis, is detected in 19–62% of patients with severe AS [19, 32, 47, 48]. Two forms of LGE can be observed: the ischemic and the non-ischemic pattern. The ischemic pattern is characterized by subendocardial LGE along specific coronary artery territories whereas in the non-ischemic pattern the distribution of LGE can be diffuse, (multi)focal or linear, confined or patchy, and is predominantly located in the midwall myocardial layer and does not correspond to a specific coronary artery territory (Fig. 2) [19, 32, 47, 48]. The presence and the extent of LGE have been associated with increased LV mass, worse LV ejection fraction, the presence of symptoms, markers of myocardial injury such NT-pro-brain natriuretic peptide and high-sensitivity cardiac troponins and ECG strain (Table 2) [19, 32, 45, 46, 48, 49]. However, LGE was not significantly associated with transaortic gradients or the aortic valve area, common indices of AS severity [19, 32, 48], suggesting that there is different individual susceptibility to develop LV hypertrophy and myocardial fibrosis, likely influenced by multiple factors such as advanced age, male sex, obesity and certain genetic variants [50].

In addition, LGE is an important prognostic marker in patients with AS [19, 32, 33]. In 143 patients with moderate and severe AS who were followed for 2.0 ± 1.4 years, the presence of LGE was associated with an increase in all-cause and cardiac mortality (every 1% increase in LGE mass was associated with 5% increased risk of all-cause mortality; p = 0.005) [19]. When dividing the population according to the pattern of LGE, patients with midwall fibrosis (N = 54) had higher mortality than patients with infarct-type LGE (N = 40) (HR 8.59; 95% CI 1.97–37.38; p = 0.004 and HR 6.46; 95% CI 1.39–30.00; p = 0.017, respectively). Furthermore, in 154 patients with severe AS undergoing surgical aortic valve replacement, the presence of LGE was an independent predictor of all-cause and cardiovascular mortality (HR for all-cause mortality: 2.8; 95% CI 1.3–6.9; p = 0.025) [32]. Importantly, after aortic valve replacement, LGE does not completely regress and has been associated with incomplete LV functional recovery, worse New York Heart Association functional class and worse survival (Fig. 6) [32, 33, 49]. However, detection of LV myocardial fibrosis in patients with asymptomatic moderate and severe AS seems insufficient to identify the patients who will present valve related complications. In the prognostic importance of microvascular dysfunction in aortic stenosis (PRIMID AS) study, including 174 patients with asymptomatic moderate to severe AS, the group of patients who presented with cardiovascular death, major adverse cardiovascular events and development of typical AS symptoms, necessitating referral for aortic valve replacement, showed comparable extent of LGE than patients who remained asymptomatic or free of valve related complications during follow-up [51].

Interstitial and replacement myocardial fibrosis lead to impaired LV myocardial deformation which can be detected with strain imaging. Myocardial tagging and feature tracking CMR demonstrated that global as well as regional LV strains were significantly correlated with LGE extent in patients with hypertrophic cardiomyopathy, who exhibit a similar pattern of midwall fibrosis to patients with AS [52, 53]: global and regional LV strain values impair as LGE increases. Head-to-head comparisons between tagged and feature tracking CMR in moderate to severe AS have shown reasonable agreement for LV strain measurement, albeit feature tracking provided systematically higher values than CMR tagging [54, 55]. The correlation between CMR LV circumferential and longitudinal strain and strain rate and symptomatic status of patients with severe AS and preserved LV ejection fraction was demonstrated by Al Musa et al. [56]. LV longitudinal strain rate was the most sensitive parameter to discriminate between asymptomatic versus symptomatic patients (−106.3 ± 43.3%/s in patients with “no/mild” symptoms vs. −83.4 ± 24.8%/s in moderate and severely symptomatic patients; P = 0.048). The association between LV myocardial strain and outcomes after surgical or transcatheter treatment was demonstrated in two studies [36, 57]. Mahmod and coworkers showed that global LV circumferential, but not longitudinal strain measured on CMR significantly improved at 8 months after aortic valve replacement [57]. Similarly, LV circumferential strain by CMR tagging was significantly associated with all-cause mortality in 98 severe AS patients undergoing surgical and transcatheter aortic valve replacement (HR per each 1% deterioration of circumferential strain: 1.03; 95% CI 1.01–1.05; p = 0.009) [36]. Furthermore, the effect of procedural access (transfemoral vs. transapical) on LV mechanics was studied with CMR feature tracking in 44 patients undergoing transcatheter aortic valve replacement [58]. The transapical approach was associated with impaired peak systolic longitudinal strain of the apical segments as compared to the transfemoral approach (−8.9 ± 5.3 vs. −16.9 ± 4.3%, respectively; p < 0.001), while there were no differences in LV ejection fraction and peak systolic longitudinal strain of the basal and midventricular segments between both approaches (Fig. 7).

In aortic regurgitation (AR), pressure and volume overload induce growth of cardiomyocytes with addition of new sarcomeres in series and interstitial fibrosis, characterized by increased fibronectin and non-collagen components [59]. Several clinical studies have histologically proven pronounced myocardial fibrosis in severe AR at the time of valve surgery [37, 60, 61]. A few studies have also evaluated myocardial fibrosis with CMR [33, 62, 63]. Sparrow et al. compared myocardial T1 values measured with a modified Look-Locker technique before and after gadolinium contrast in eight patients with severe AR and 15 normal controls [62]. Patients with AR had significantly prolonged post-contrast T1 values in abnormally contracting segments compared to the controls (532 vs. 501 ms, respectively; p = 0.002), suggesting increased interstitial fibrosis. Furthermore, in nine patients with severe AR who underwent surgical aortic valve replacement, ECV measured on 3T CMR was strongly correlated with the extent of interstitial fibrosis on histology (r = 0.79, p = 0.011) [63]. Replacement fibrosis has been also described in 26 patients with severe AR by Azevedo et al. [33]. The authors reported a 69% prevalence of LGE, mostly following a multifocal pattern. The correlation between myocardial replacement fibrosis assessed with LGE and histopathology was good (r = 0.70, p < 0.001). Moreover, in a combined cohort of 26 patients with severe AR and 28 patients with severe AS, the amount of myocardial fibrosis was inversely correlated with LV functional improvement (r = −0.47; p = 0.02) and was associated with worse long-term survival after aortic valve replacement surgery (Chi square = 5.85; p = 0.02) (Fig. 6) [33]. Furthermore, in 14 patients with chronic severe AR, myocardial CMR tagging showed an impairment in global longitudinal and circumferential strain at 2 years after aortic valve replacement (p < 0.03 for both), despite an improvement in LV ejection fraction and a decrease in LV size (Fig. 8) [64]. Similarly, Ungacta et al. showed a decrease in posterior wall circumferential strain in patients with AR 6 months after valve replacement [65]. These findings suggest that the presence of LV myocardial fibrosis in patients with AR is a marker of adverse remodeling that may lead to further deterioration in LV strain and poor prognosis after aortic valve surgery.

Mitral regurgitation (MR) is a heterogeneous disease, broadly classified as organic (primary) or functional (secondary) based on the underlying mechanism. Organic MR is due to intrinsic valvular disease whereas functional MR is caused by regional and/or global LV remodeling without structural abnormalities of the mitral valve [66]. Degenerative mitral valve disease (myxomatous disease and fibroelastic deficiency) is the most frequent etiology of primary MR in developed countries. The indication for mitral valve repair/replacement is determined by the presence of symptoms or LV function deterioration and LV remodeling [1, 3]. However, LV remodeling and myocardial fibrosis may occur before the development of symptoms. Chronic LV volume overload associated with MR leads to myocardial hypertrophy and increased interstitial fibrosis [67]. In 35 asymptomatic patients with moderate to severe primary MR, Edwards et al. demonstrated higher ECV on CMR as compared to controls (0.32 ± 0.07 vs. 0.25 ± 0.02, p < 0.01) (Fig. 9) [68]. Furthermore, 31% of patients with MR exhibited a non-infarct LGE pattern on CMR. Patients who had non-infarct type LGE presented with higher ECV values compared to MR patients without LGE (0.35 ± 0.02 vs. 0.27 ± 0.03, p < 0.01). The ECV values correlated with LV end-systolic volume, measures of systolic and diastolic LV dysfunction as well as with peak oxygen consumption on treadmill testing. The distribution of LGE in patients with MR varies significantly. Han et al. demonstrated the presence of LGE of the papillary muscles in 63% of patients with MV prolapse [69] whereas Chaikriangkrai et al. observed LV replacement fibrosis in 40% of patients with chronic severe MR [35]. The presence of LV LGE was associated with worse clinical outcomes in terms of intensive care unit readmission, incidence of permanent pacemaker implantation and rehospitalization (HR 4.775; 95% CI 1.100–20.729; p = 0.037) [35].

These structural changes of the LV myocardium may be associated with subtle functional abnormalities. In 15 patients with chronic moderate and severe MR and preserved LV ejection fraction who underwent CMR with tissue tagging, Maniar et al. demonstrated preserved global longitudinal and circumferential strain but abnormal regional strain values: the septal LV segments exhibited impaired strain whereas the lateral segments showed compensatory hyper-contractility [70]. Similarly, Mankad et al. showed with CMR tagging abnormal regional strain patterns in patients with severe MR and preserved LV ejection fraction: while radial strain was increased (19 ± 9 vs. 16 ± 6%, p = 0.003), circumferential strain was reduced (12 ± 6 vs. 21 ± 6%, p ≤ 0.001) as compared to healthy controls [71]. Several authors have demonstrated a decrease in global longitudinal and circumferential strain parameters on CMR tagging in patients with severe degenerative MR after mitral valve repair, which might imply an ongoing myocardial fibrosis after surgery [72‐74].