Approximately one-third of patients with DCM exhibit replacement myocardial fibrosis [31], and numerous previous studies have consistently shown a correlation between replacement myocardial fibrosis and various adverse outcomes, highlighting its independent predictive value for adverse events, such as cardiac events, all-cause mortality, and arrhythmic events (e.g., SCD, ventricular fibrillation, ventricular tachycardia) [32, 33]. The LGE technique can effectively distinguish between normal and fibrotic myocardia, making it the gold standard for the noninvasive assessment of replacement myocardial fibrosis widely employed in clinical practice [34].

Several studies have shown that the presence of LGE is a robust prognostic indicator of adverse outcomes in patients with DCM, particularly in terms of SCD events [7, 35, 36]. Moreover, Di Marco et al. [36, 37] observed that LGE exhibited superior predictive efficacy for arrhythmia events, particularly in patients without significant LVEF decline (hazard ratio [HR]: 10.4; p 0.001). This is a particularly noteworthy finding because it challenges the traditional reliance on LVEF as the primary indicator of risk in patients with DCM. The superior predictive power of LGE in this context suggests that LGE imaging can identify high-risk patients who may not be captured by LVEF alone.

In patients with significant LVEF decline, the left ventricle undergoes substantial remodeling, resulting in tissue instability at a critical level. The cascade of acute electrical instability and mechanical failure can ultimately lead to SCD [38]. However, in cases where the left ventricle is not extensively remodeled, myocardial fibrosis may be a primary contributor to VA and SCD. Reduced normal cardiomyocytes within fibrotic areas lead to altered electrophysiological properties, typically characterized by low weighted unipolar voltages. These changes affect myocardial conduction and excitability, increasing the risk of VA and SCD. Additionally, the altered electrical activity may be incongruous with the original neural innervation, potentially further influencing the occurrence of VA and SCD [39].

Since the initial study in 2006 that evaluated the prognostic significance of LGE in DCM, several investigations on LGE in DCM have been conducted [40, 41]. These studies generally corroborated that LGE can serve as an indicator of adverse outcomes [42, 43, 44]. However, the predictive capability of LGE across different studies has been substantially inconsistent, particularly concerning various clinical endpoints [45, 46, 47]. In particular, this variability may manifest as a HR of 1.45 for composite endpoint events and reach as high as 14 for arrhythmic events [34, 48]. These observations suggest that relying on LGE alone is not sufficient to achieve the desired accuracy of risk stratification. Therefore, incorporating more detailed LGE-related indicators is crucial for enhancing the effectiveness of LGE in predicting adverse outcomes. These refined indicators will enable a more accurate assessment of patient condition and risk stratification, facilitating the implementation of precise risk stratification strategies. (The brief summary of the cited article is shown in Table 1 (Ref. [6, 7, 20, 25, 26, 31, 34, 35, 36, 37, 40, 41, 42, 43, 44, 45, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59])).

Considering the high prevalence of LGE, reliance on the presence of LGE only as an indication of adverse events may not be sufficient. The multiplane and multiparameter characteristics of CMR enable quantitative evaluate LGE extent. Several studies have explored the prognostic value of the extent of LGE, revealing that each 1% increase in LGE extent can be a predictor of adverse outcomes [25, 49, 50]. Furthermore, a study focusing on patients with LVEF 50% reported that each 1% increase in LGE extent was notably correlated with a heightened risk of mortality by approximately 3% [51]. Therefore, this correlation may stem from an enhanced susceptibility to re-entrant arrhythmias as the extent of LGE increases [25], highlighting the importance of the accurate quantification of LGE using CMR.

However, Halliday et al. [35] and Behera et al. [52] noted a nonlinear relationship between LGE extent and prognosis. They revealed that the risk of mortality increases with LGE extent; however, this increase was not monotonic. For instance, the HRs for SCD events were found to be 1.59 for the 0%–2.55% LGE extent group, 1.56 for the 2.55%–5.10% group, and 2.31 for the $>$5.10% group [35]. Despite the increase in risk with increasing LGE extent, the increase rate is not uniform, and the increase rate of HRs for all-cause mortality varies. Consequently, modeling LGE extent as a linear predictor may underestimate the risk in most patients and overestimate the risk in those with the most extensive LGE. Notably, in Halliday et al.’s study [35], the LGE extent that was most predictive of adverse outcomes was found to be 1.29%, with a corresponding C-statistic of 0.70.

Furthermore, a recent study revealed the interaction between LGE extent and LVEF for predicting outcomes. In particular, the study revealed that in patients with LVEF $\ge$35%, the predictive value of LGE for adverse events becomes particularly pronounced when LGE extent exceeds 7.1%. In these patients, the risk of SCD or aborted SCD was 4.4 times higher (HR = 4.4; 95% CI: 2.4–8.3; p 0.001) [6]. This finding suggests that the combination of LGE extent and LVEF provides a more nuanced approach to risk stratification, emphasizing the importance of identifying the specific threshold of LGE extent that correlates with increased risk.

However, the current method for measuring LGE extent using CMR primarily reflects the percentage of the LGE area that occupies the left ventricular myocardial mass without considering the heterogeneity of regional LGE. The limitations of this method may lead to inaccurate assessment of cardiac structure and function. Research suggests that elements influencing the nonlinear relationship between LGE extent and prognosis include the fact that the critical isthmus sites of VA are located in tissue heterogeneity zones, which lie between myocardial fibrosis and healthy tissue [60]. Additionally, as the spatial extent and degree of heterogeneity of LGE increase, the heterogeneity of electrical conduction within cardiac tissue also rises, making it more likely for reentrant circuits to form and thereby increasing the probability of inducing arrhythmias. It is also important to note that the timing of arrhythmia induction is primarily determined by the level of local maximum fibrosis. This is because the activation patterns of arrhythmias exhibit a high degree of periodicity and regularity, and the regions with the highest degree of fibrosis will dictate the interval between arrhythmia episodes [61]. Moreover, previous studies investigating the impact of LGE extent did not adequately consider the potential effects of different LGE patterns. Di Marco et al. [36] found that epicardial and transmural LGE were significantly associated with a higher proportion of adverse outcomes than mid-wall LGE. However, whether epicardial and transmural LGE still maintain their high-risk characteristics at the same extent of LGE lesions remains to be clarified through further research.

The aforementioned studies have indicated that placing greater emphasis on the composition of LGE indicators rather than its volume is essential. However, there is currently a lack of consensus on the measurement methods and parameters for quantifying LGE, necessitating urgent standardization of LGE quantification methods before their use as decision-making tools. A brief summary of the cited article is shown in Table 1 (Ref. [6, 25, 35, 36, 49, 50, 51, 52]).

The localization of LGE in patients with idiopathic DCM predominantly occurs in the septum caused by specific pathological mechanisms, whereas LGE in DCM caused by viral myocarditis is predominantly observed in the free wall [62]. This disparity may be because cardiotropic viruses causing viral myocarditis mainly enter through the bloodstream, leading to direct exposure of the free wall to inflammatory factors present in the blood due to its contact with the pericardium [63]. The microstructural characteristics of fibrosis may differ depending on DCM etiology, thereby influencing its prognostic significance.

Previous studies have primarily focused on septal LGE and identified associations between all-cause mortality, VA events, and composite endpoints [25, 35, 52, 53]. With the advancement and maturation of CMR technology, attention has gradually shifted toward LGE in other myocardial regions. Some investigations have focused on LGE within the free wall; however, there remains a debate regarding whether free wall LGE is a protective or detrimental factor, with conflicting findings from these studies (HR: 0.77 and HR: 5.42) [35, 53]. This ambiguity highlights the need for further research to delineate the clinical implications of free wall LGE. Furthermore, studies exploring LGE in the inferior wall of the free wall have uniformly revealed a notable link between inferior wall LGE and adverse outcomes [25, 52], reinforcing the importance of considering the specific LGE location in the context of cardiac disease management.

Although the prognostic assessment of isolated free wall LGE in patients with DCM remains questionable, the presence of free wall and septal LGE often indicates poor prognosis. Previous studies have investigated the impact of LGE at different locations on patient prognosis and found a significant correlation between the multiple locations (both septal and free wall) of LGE and adverse outcomes in patients with DCM [35, 36, 53]. Particularly in Halliday et al.’s research [35], the HR for SCD events reached 5.82 when LGE was present in both the septum and free wall. This finding supports the hypothesis that prognosis deteriorates when LGE occurs at multiple locations within the same patient due to the potential indication of greater extent and higher LGE heterogeneity.

In a recent multicenter cohort study involving 1165 consecutive patients with DCM [20], the presence of LGE at right ventricular insertion points was associated with comparable outcomes to LGE-negative patients, while significantly reducing the risk of VA and sudden death compared with cases with left ventricular LGE. When analyzing LGE, particularly when quantifying LGE extent, distinguishing between LGE at the right ventricular insertion point and LGE at other locations in the left ventricle is advisable. The brief summary of the cited article is shown in Table 1 (Ref. [20, 25, 35, 36, 52, 53]).

The mid-wall pattern, which is the most common manifestation of LGE, has gained widespread recognition. Concurrently, subepicardial, subendocardial, and transmural LGE patterns have also been research spotlights, drawing considerable attention (all LGE patterns are shown in Fig. 1). In particular, numerous studies have definitively highlighted subepicardial myocardial fibrosis as a key LGE pattern that indicates poor prognosis, making this a significant area of focus [35, 52, 53]. Furthermore, de Frutos et al. [64] demonstrated that the distribution of LGE patterns in DCM is gene-specific. In particular, LGE-negative patients exhibited a higher frequency of mutations in the TNNT2, RBM20, and MYH7 genes, whereas those with subepicardial LGE demonstrated a greater prevalence of DMD, DSP, and FLNC gene mutations. These findings enable effective screening of high-risk families and provision of prognostic guidance.

Fig. 1.

Schematic illustration of different LGE patterns and corresponding MRI images. (A) Normal heart. (B) The mid-wall pattern of LGE located in septum. (C) The focal pattern of LGE located in right ventricular insertion point. (D) The subepicardial pattern of LGE located in free wall. (E) The subendocardial pattern of LGE located in septum. (F) The transmural pattern of LGE located in both septum and free wall. LGE, late gadolinium enhancement; MRI, magnetic resonance imaging.

Previous studies have demonstrated that subendocardial and transmural LGE patterns indicate previous myocardial infarction [7, 65]. However, some studies have identified subendocardial and transmural LGE patterns that are inconsistent with the distribution of coronary arteries and have confirmed their predictive value for adverse outcomes [34, 36]. These findings imply that these two LGE patterns cannot be simply classified as ischemic lesions. Although the etiology remains unclear and most patients with these LGE types are associated with microcirculatory disturbances and abnormal myocardial perfusion [54], it is also plausible that viral infections and other nonischemic mechanisms contribute to these patterns [55]. Consequently, it is essential to direct more attention to these neglected LGE subtypes.

However, current research shows there is controversy regarding whether the focal LGE pattern serves as a protective or adverse factor because their HR values were 3.16 and 0.21, respectively [35, 53]. This discrepancy may be attributed to the limited number of patients exhibiting this specific pattern. Furthermore, due to the thin myocardial walls in patients with DCM, distinguishing focal LGE from other patterns, such as subepicardial and subendocardial LGE, is challenging.

Similar to LGE location, diverse LGE patterns can be observed within the same patient, each exhibiting variations in fibrosis structure. The literature shows that the presence of multiple LGE patterns (the combination of mid-wall, subepicardial, and focal) is correlated with worse outcomes [34, 35, 53]. Note that Alba et al. [34] demonstrated a gradient response in the coexistence of multiple LGE patterns, with an elevated risk that increases as the number of patterns present increases. This trend was manifested with a HR of 1.24 with 1 pattern, 2.07 with 2 patterns, and 4.76 with $>$2 patterns, indicating a progressive correlation between the complexity of LGE patterns and the risk of adverse outcomes. The brief summary of the cited article is shown in Table 1 (Ref. [34, 35, 36, 52, 53, 54, 55]).

Furthermore, recent studies have integrated LVEF strata with LGE status, which has significantly enhanced the accuracy of risk stratification in patients with DCM, providing incremental valuable insights for the improvement of risk stratification [6, 36]. Di Marco et al. [36] identified high-risk LGE in patients with DCM, including epicardial, transmural, and combined septal and free wall LGE. Considering the differences in annual event rates among patients with different LVEF categories, the study divided LVEF values into three strata ($\le$20%, 21%–35%, and $>$35%). Combining LGE characteristics with LVEF strata, the researchers developed a risk model for patients with DCM and classified these patients into seven risk categories. Based on this model, patients with negative LGE and an LVEF $>$35% exhibited the lowest risk of SCD and VA events, with an annual event rate of 0.07%. In contrast, patients presenting with LGE and an LVEF 20% encounter a notably increased risk, with an annual event rate of 8.1%.

Building upon these findings, the subsequent step involves subdividing each parameter (i.e., extent, location, and pattern) of LGE and integrating them with specific cardiac function information (i.e., LVEF and myocardial deformation) or clinical indicators to potentially refine risk stratification. The brief summary of the cited article is shown in Table 1 (Ref. [6, 36]).

In a normal myocardium, the extracellular volume comprises intravascular compartments containing blood, and the stroma of the myocardium contains signaling molecules, such as fibrillary collagen and proteoglycans. However, in patients with cardiomyopathy, the presence of extracellular or interstitial edema, replacement fibrosis, or invasive fibrosis can increase interstitial component levels. This subsequently increases extracellular volume, prolongs myocardial T1 relaxation time, and is associated with native T1 values [30].

Native T1 values, which do not require the administration of contrast agents during imaging examinations, can serve as a complementary or alternative method for assessing LGE, particularly for patients with severe renal insufficiency or those who cannot tolerate contrast-enhanced examinations. However, technical factors, such as magnetic field strength, variations in T1 mapping techniques, and non-resonant artifacts, can introduce errors when assessing native T1 values. Furthermore, most cardiomyopathies exhibit relatively subtle changes in native T1 values, which may pose challenges for detection. ECV, representing the proportion of extracellular stromal volume in the entire myocardium, is calculated based on pre- and post-enhancement myocardial T1 values. It reflects changes in extracellular stroma and is strongly associated with extracellular matrix, as confirmed by biopsy [26]. This technique is a noninvasive method for evaluating structural changes and provides a deeper understanding of the underlying pathology of DCM. The combination of native T1 values and ECV allows for a comprehensive evaluation of interstitial myocardial fibrosis in patients with DCM. Although current research is limited, the T1 and ECV mapping techniques have shown promising prospects for prognostic evaluation of DCM.

Kitagawa et al. [56] demonstrated native T1 values as a predictor of left ventricular reverse remodeling and cardiac events in DCM patients. This finding is consistent with the results of another study that found native T1 values to be as predictive of arrhythmia events [57]. Complementing this, research has shown the ability of the ECV to forecast outcomes, such as HF in DCM [58]. Furthermore, a separate study demonstrated that native T1 and ECV could predict cardiac mortality and heart transplantation in patients with DCM without LGE [50].

In Li’s group [50], native T1 $>$936 ms and ECV $>$25.9% were confirmed to be independent predictors of adverse outcomes. Two other studies have discovered that per 10-ms increase in native T1 and per 3% increase in ECV were correlated with SCD events, and the number of abnormal ECV locations was also found to be linearly related to the annual major adverse cardiac events rate [58, 59]. Therefore, although T1 and ECV mapping show great promise for predicting the prognosis of patients with DCM, the evaluation metrics in each study varied, and the standardized threshold values for the relevant parameters were not determined. Furthermore, the optimal critical values for identifying high-risk individuals may differ according to different patient populations and imaging platforms. These are all issues related to T1 and ECV mapping that require attention. To address these issues, further studies using larger datasets and different patient populations are required.

The latest research has revealed that compared with using LGE or LVEF alone, models that combine LGE with native T1 or ECV are more effective in stratifying the risk of SCD in patients with DCM, highlighting a superior predictive ability [59]. Therefore, exploring the predictive value of T1 and ECV mapping in combination with other imaging markers, such as LGE and LVEF, along with further sublevel indicators, including the location and pattern of LGE, could improve risk stratification and provide a more comprehensive assessment of the underlying cardiac pathology. This multifaceted approach can help clinicians identify high-risk patients more accurately and tailor personalized management strategies to improve outcomes in this challenging patient population. The brief summary of the cited article is shown in Table 1 (Ref. [26, 50, 56, 57, 58, 59]).