Impact statement

Advances in the understanding of the molecular mechanisms that cause dilated cardiomyopathy offer the opportunity to personalise care and improve the outcomes of patients with this heterogeneous family of disease. Comprehensive characterisation of the disease with genetic testing and advanced imaging will play a key role. Precision therapies that target the primary disease mechanism will offer new hope for disease prevention in genetically susceptible individuals at risk of developing highly penetrant, malignant forms of the condition as well as effective treatments of early asymptomatic disease.

The genetic architecture of DCM

Using currently available next-generation sequencing panels, a causative rare genetic variant in about 20–30% of cases of DCM (Tayal et al., Reference Tayal, Ware, Lakdawala, Heymans and Prasad2021). The yield may be higher in select populations, such as those referred for advanced heart failure therapies, younger patients or those with a high burden of ventricular arrhythmia or conduction disease (Herman et al., Reference Herman, Lam, Taylor, Wang, Teekakirikul, Christodoulou, Conner, SR, McDonough, Sparks, Teodorescu, Cirino, Banner, Pennell, Graw, Merlo, Di Lenarda, Sinagra, Bos, Ackerman, Mitchell, Murry, Lakdawala, Ho, Barton, Cook, Mestroni, Seidman and Seidman2012; Lukas Laws et al., Reference Lukas Laws, Lancaster, Ben Shoemaker, Stevenson, Hung, Wells, Marshall Brinkley, Hughes, Anderson, Roden and Stevenson2022). Several different genes encoding a range of proteins with diverse functions have been implicated in DCM (Jordan et al., Reference Jordan, Peterson, Ai, Asatryan, Bronicki, Brown, Celeghin, Edwards, Fan, Ingles, James, Jarinova, Johnson, Judge, Lahrouchi, Lekanne Deprez, Lumbers, Mazzarotto, Medeiros Domingo, Miller, Morales, Murray, Peters, Pilichou, Protonotarios, Semsarian, Shah, Syrris, Thaxton, van Tintelen, Walsh, Wang, Ware and Hershberger2021). These are most commonly autosomal with X-linked and mitochondrial variants also uncommonly identified. Autosomal dominant inheritance is the norm and the most common genes implicated are those coding for sarcomeric proteins, including titin (TTN) and beta-myosin (MYH7). Other notable genes include those coding for cytoskeletal (FLNC), nuclear envelope (LMNA) and desmosomal (DSP) proteins. In some cases, the identification of a causative variant may influence treatment decisions, particularly those associated with a more malignant prognosis. However, the greatest value of genetic testing comes in predicting the risk of asymptomatic relatives by enabling the identification of those at high risk of developing the disease in the future. In gene elusive disease, all first-degree family relatives typically remain under clinical surveillance until the age of 60 years. If a causative genetic variant is identified, cascade screening will be able to identify the 50% of relatives who are carriers and who have an elevated risk of developing disease. Those who do not carry the variant can be discharged from follow-up.

A landmark study found rare truncating variants in the titin (TTNtv) gene in 25% of patients with advanced or familial DCM (Herman et al., Reference Herman, Lam, Taylor, Wang, Teekakirikul, Christodoulou, Conner, SR, McDonough, Sparks, Teodorescu, Cirino, Banner, Pennell, Graw, Merlo, Di Lenarda, Sinagra, Bos, Ackerman, Mitchell, Murry, Lakdawala, Ho, Barton, Cook, Mestroni, Seidman and Seidman2012). More recent studies found TTNtv in 13% of nonfamilial cases of DCM and ~ 0.5% of the general population (McNally and Mestroni, Reference McNally and Mestroni2017; Tayal and Prasad, Reference Tayal and Prasad2018; Schultheiss et al., Reference Schultheiss, Fairweather, Caforio, Escher, Hershberger, Lipshultz, Liu, Matsumori, Mazzanti, McMurray and Priori2019; Verdonschot et al., Reference Verdonschot, Hazebroek, Ware, Prasad and Heymans2019). These variants are associated with incomplete penetrance and variable expression (Japp et al., Reference Japp, Gulati, Cook, Cowie and Prasad2016) that may be attributed to additional gene-modifying factors, environmental factors, or the penetrance of disease later in life (Japp et al., Reference Japp, Gulati, Cook, Cowie and Prasad2016). Penetrance is likely to vary significantly between asymptomatic carriers with a family history of DCM and those found to carry a TTNtv as a secondary finding on testing performed for another reason. Nevertheless, subtle markers of reduced cardiac function have been found in carriers in the general population (Schafer et al., Reference Schafer, de Marvao, Adami, Fiedler, Ng, Khin, Rackham, van Heesch, Pua, Kui, Walsh, Tayal, Prasad, Dawes, Ko, Sim, Chan, Chin, Mazzarotto, Barton, Kreuchwig, de Kleijn, Totman, Biffi, Tee, Rueckert, Schneider, Faber, Regitz-Zagrosek, Seidman, Seidman, Linke, Kovalik, O’Regan, Ware, Hubner and Cook2016), suggesting they may be more susceptible to extrinsic insults, such as alcohol and cardiotoxic chemotherapy.

Recent research has emphasised the importance of common genetic variation or polygenic risk in determining the risk of developing DCM (Pirruccello et al., Reference Pirruccello, Bick, Wang, Chaffin, Friedman, Yao, Guo, Venkatesh, Taylor, Post, Rich, Lima, Rotter, Philippakis, Lubitz, Ellinor, Khera, Kathiresan and Aragam2020; Tadros et al., Reference Tadros, Francis, Xu, Vermeer, Harper, Huurman, Kelu Bisabu, Walsh, Hoorntje, Te Rijdt, Buchan, van Velzen, van Slegtenhorst, Vermeulen, Offerhaus, Bai, de Marvao, Lahrouchi, Beekman, Karper, Veldink, Kayvanpour, Pantazis, Baksi, Whiffin, Mazzarotto, Sloane, Suzuki, Schneider-Luftman, Elliott, Richard, Ader, Villard, Lichtner, Meitinger, Tanck, van Tintelen, Thain, McCarty, Hegele, Roberts, Amyot, Dube, Cadrin-Tourigny, Giraldeau, L’Allier, Garceau, Tardif, Boekholdt, Lumbers, Asselbergs, Barton, Cook, Prasad, O’Regan, van der Velden, Verweij, Talajic, Lettre, Pinto, Meder, Charron, de Boer, Christiaans, Michels, Wilde, Watkins, Matthews, Ware and Bezzina2021). Many patients without a rare variant cause are likely to have high polygenic risk contributing to the development of contractile dysfunction along with extrinsic factors. Polygenic risk is also likely to influence the penetrance of rare genetic variants, helping to explain the variable expressivity and incomplete penetrance commonly seen across families with a pathogenic variant. A future precision approach to determining disease risk in families with DCM is likely to integrate data on phenotype, rare and common genetic variation and the interaction with extrinsic insults.

Autosomal recessive transmission has also been described. This is of particular relevance in younger individuals and childhood cardiomyopathies. For example, biallelic loss of function mutations in the nebulin-related anchoring protein gene (NRAP) have been identified in some individuals with severe sporadic DCM and have been proposed to cause low-penetrant recessive disease (Iuso et al., Reference Iuso, Wiersma, Schuller, Pode-Shakked, Marek-Yagel, Grigat, Schwarzmayr, Berutti, Alhaddad, Kanon, Grzeschik, Okun, Perles, Salem, Barel, Vardi, Rubinshtein, Tirosh, Dubnov-Raz, Messias, Terrile, Barshack, Volkov, Avivi, Eyal, Mastantuono, Kumbar, Abudi, Braunisch, Strom, Meitinger, Hoffmann, Prokisch, Haack, Brundel, Haas, Sibon and Anikster2018; Koskenvuo et al., Reference Koskenvuo, Saarinen, Ahonen, Tommiska, Weckstrom, Seppala, Tuupanen, Kangas-Kontio, Schleit, Helio, Hathaway, Gummesson, Dahlberg, Ojala, Vepsalainen, Kytola, Muona, Sistonen, Salmenpera, Gentile, Paananen, Myllykangas, Alastalo and Helio2021). Several syndromic causes of DCM have been identified. These include dystrophinopathies such as Duchenne and Becker’s muscular dystrophy, and other eponymous syndromes including Barth syndrome (Hershberger et al., Reference Hershberger, Cowan, Morales and Siegfried2009; Hershberger et al., Reference Hershberger, Hedges and Morales2013). A raised creatinine kinase level and characteristic sub-epicardial fibrosis in the lateral wall are typical in patients with dystrophinopathies and cardiac manifestations may predate neuromuscular symptoms in Becker’s muscular dystrophy (Del Rio-Pertuz et al., Reference Del Rio-Pertuz, Morataya, Parmar, Dubay and Argueta-Sosa2022). Rare metabolic disorders, particularly inborn errors of metabolism, have also been associated with DCM (Guertl et al., Reference Guertl, Noehammer and Hoefler2000; Cox, Reference Cox2007). Broadly, these can be grouped into disorders of amino acid/organic acid metabolism, disorders of fatty acid metabolism, glycogen and lysosomal storage disorder and mitochondrial disorders (Guertl et al., Reference Guertl, Noehammer and Hoefler2000).

Stratifying arrhythmic risk in genetic DCM

The traditional approach to stratifying the risk of major ventricular arrhythmia in DCM relies on a combination of symptoms and left ventricular ejection fraction (LVEF). However, this “cause-agnostic” approach does not fully encapsulate the heterogeneity of DCM and a growing body of data support an increased risk of SCD with specific genotypes. This has begun to influence international guidelines on the selection of patients for implantable cardioverter defibrillators (ICDs) that recommend lower thresholds for such devices in patients with LMNA, FLNC, PLN or RBM20 variants and other high-risk features beyond LVEF (Zeppenfeld et al., Reference Zeppenfeld, Tfelt-Hansen, de Riva, Winkel, Behr, Blom, Charron, Corrado, Dagres, de Chillou, Eckardt, Friede, Haugaa, Hocini, Lambiase, Marijon, Merino, Peichl, Priori, Reichlin, Schulz-Menger, Sticherling, Tzeis, Verstrael, Volterrani and Group2022). This represents a wider trend in recent guidelines that attempt to risk stratify patients according to genotype and phenotype to make personalised decisions about their care, attempting to break down the traditional grouping of “non-ischemic cardiomyopathy” (Table 1).

Table 1. Genes with definite/strong association with DCM and their functional and phenotypic implications (14, 17)

Is genotype-specific therapy the answer?

Discovering a monogenic cause for DCM provides direct insight into the molecular mechanisms that drive disease. This creates the possibility of using precision therapies directed at the primary molecular basis of disease. Such approaches are not only relevant to those with symptomatic heart failure where they may be used alongside GDMT, but perhaps more importantly for asymptomatic individuals with early markers of disease or those with genetic susceptibility to developing disease later in life. Evidence-based treatments for the latter groups are currently lacking. Targeted therapy for disease prevention must be a priority. Strategies targeting the primary disease mechanisms may take different main approaches.

The immediately downstream molecular consequences of the variant represent attractive targets. Most genes associated with cardiomyopathy serve important functions within the cardiomyocyte, with their respective proteins carrying out specific functions. Disruptions in the function of these proteins, either through loss or gain of function, result in intracellular changes in signal transduction, prompting the cardiomyocyte to undergo adaptive changes (Reichart et al., Reference Reichart, Magnussen, Zeller and Blankenberg2019). Given the heterogeneity of DCM, downstream targeting of these processes would require a wide variety of agents to target the products of different genes implicated in the pathogenesis. One such target that was recently investigated in phase II and III trials was the heightened cardiac activity of ERK1/2, JNK, and p38 MAP kinases downstream from variants in LMNA associated with DCM (Muchir et al., Reference Muchir, Wu, Choi, Iwata, Morrow, Homma and Worman2012). Much hope arose from animal studies that demonstrated a reduction in adverse remodelling following the administration of a p38 inhibitor (Wu et al., Reference Wu, Muchir, Shan, Bonne and Worman2011; Laurini et al., Reference Laurini, Martinelli, Lanzicher, Puzzi, Borin, Chen, Long, Lee, Mestroni, Taylor, Sbaizero and Pricl2018). Unfortunately, these results were not translated into the phase III trial that was recently stopped due to futility.

Another example comes from the use of myosin modulators in sarcomeric DCM. Sarcomeric dysfunction is the primary mechanism of DCM in patients with TTNtv or relevant variants in MYH7. This is the opposite functional consequence of sarcomeric variants causing hypertrophic cardiomyopathy (HCM) that are associated with sarcomeric over-action. In the same way that promise has arisen from the use of myosin inhibitors, there is excitement about the potential use of agents such as danicamtiv and omecamtiv mecarbil that increase actin-myosin cross-bridging in sarcomeric DCM (Voors et al., Reference Voors, Tamby, Cleland, Koren, Forgosh, Gupta, Lund, Camacho, Karra, Swart, Pellicori, Wagner, Hershberger, Prasad, Anderson, Anto, Bell, Edelberg, Fang, Henze, Kelly, Kurio, Li, Wells, Yang, Teichman, Del Rio and Solomon2020; Teerlink et al., Reference Teerlink, Diaz, Felker, McMurray, Metra, Solomon, Adams, Anand, Arias-Mendoza, Biering-Sorensen, Bohm, Bonderman, Cleland, Corbalan, Crespo-Leiro, Dahlstrom, Echeverria, Fang, Filippatos, Fonseca, Goncalvesova, Goudev, Howlett, Lanfear, Li, Lund, Macdonald, Mareev, Momomura, O’Meara, Parkhomenko, Ponikowski, Ramires, Serpytis, Sliwa, Spinar, Suter, Tomcsanyi, Vandekerckhove, Vinereanu, Voors, Yilmaz, Zannad, Sharpsten, Legg, Varin, Honarpour, Abbasi, Malik and Kurtz2021). Similarly, emerging data suggest that TTNtv are associated with modifications in cardiac metabolism and energy utilisation (Verdonschot et al., Reference Verdonschot, Hazebroek, Derks, Barandiaran Aizpurua, Merken, Wang, Bierau, van den Wijngaard, Schalla, Abdul Hamid, van Bilsen, van Empel, Knackstedt, Brunner-La Rocca, Brunner, Krapels and Heymans2018; Ware and Cook, Reference Ware and Cook2018; Zhou et al., Reference Zhou, Ng, Ko, Fiedler, Khin, Lim, Sahib, Wu, Chothani, Schafer, Bay, Sinha, Cook and Yen2019). In particular, an upregulation in the transcription of important mitochondrial machinery may represent a compensatory response to sarcomeric dysfunction (Ware and Cook, Reference Ware and Cook2018). Targeting early mitochondrial dysfunction may therefore be a promising target for future investigation.

An example of precision therapy from current clinical practice is the use of sodium channel blockers such as flecainide or quinidine for DCM associated with SCN5A variants that result in an increased sodium current (Peters et al., Reference Peters, Thompson, Perrin, James, Zentner, Kalman, Vandenberg and Fatkin2022). A recent systematic review has shown such cardiomyopathies, typically associated with a high burden of ventricular arrhythmias, to be responsive to sodium channel blockers (Peters et al., Reference Peters, Thompson, Perrin, James, Zentner, Kalman, Vandenberg and Fatkin2022). It may be argued that such phenotypes are a primary electrical disease rather than a true cardiomyopathy. Nevertheless, the reversibility with widely available therapies emphasises the importance of achieving a genetic diagnosis, avoiding other unnecessary invasive procedures (Figure 3).

Figure 3. Cellular locations of some of the proteins with their respective genes associated with dilated cardiomyopathy.

Arguably, the most definitive treatment approaches are those that directly target the genetic variant (Verdonschot et al., Reference Verdonschot, Hazebroek, Ware, Prasad and Heymans2019). Various methods are currently being investigated to accomplish this objective, including: (1) gene editing – the use of CRISPR/Cas9 to directly edit the genetic sequence and restore normal protein function, (2) gene replacement therapy for cardiomyopathies associated with loss of function variants where the wild type gene is expressed, primarily through gene transfer techniques with viral vectors, (3) gene silencing therapy, primarily using small interfering RNA molecules to reduce the expression of abnormal functioning protein as a result of missense variants, and (4) exon skipping, involving the use of anti-sense oligonucleotides to mask exons during transcription and restoring the reading frame (Carrier et al., Reference Carrier, Mearini, Stathopoulou and Cuello2015; Gramlich et al., Reference Gramlich, Pane, Zhou, Chen, Murgia, Schotterl, Goedel, Metzger, Brade, Parrotta, Schaller, Gerull, Thierfelder, Aartsma-Rus, Labeit, Atherton, McGaughran, Harvey, Sinnecker, Mann, Laugwitz, Gawaz and Moretti2015; Prondzynski et al., Reference Prondzynski, Kramer, Laufer, Shibamiya, Pless, Flenner, Muller, Munch, Redwood, Hansen, Patten, Eschenhagen, Mearini and Carrier2017; Ma et al., Reference Ma, Zhang, Itzhaki, Zhang, Chen, Haddad, Kitani, Wilson, Tian, Shrestha, Wu, Lam, Sayed and Wu2018).

Much of the early progress in this area has been in Duchenne muscular dystrophy (DMD) where both exon skipping and gene editing have been used to restore dystrophin production in experimental models (Amoasii et al., Reference Amoasii, Hildyard, Li, Sanchez-Ortiz, Mireault, Caballero, Harron, Stathopoulou, Massey, Shelton, Bassel-Duby, Piercy and Olson2018). Early work has also demonstrated the potential of similar approaches in DCM associated with TTN, another similarly large gene with areas of redundant sequence (Gramlich et al., Reference Gramlich, Pane, Zhou, Chen, Murgia, Schotterl, Goedel, Metzger, Brade, Parrotta, Schaller, Gerull, Thierfelder, Aartsma-Rus, Labeit, Atherton, McGaughran, Harvey, Sinnecker, Mann, Laugwitz, Gawaz and Moretti2015; Romano et al., Reference Romano, Ghahremani, Zimmerman, Legere, Thakar, Ladha, Pettinato and Hinson2022). Pre-clinical studies in mice with the well-described PLN R14 gene deletion, have also used anti-sense oligonucleotides to decrease phospholamban activity, prevent cardiac dysfunction, and improve survival (Grote Beverborg et al., Reference Grote Beverborg, Spater, Knoll, Hidalgo, Yeh, Elbeck, Sillje, Eijgenraam, Siga, Zurek, Palmer, Pehrsson, Albery, Bomer, Hoes, Boogerd, Frisk, van Rooij, Damle, Louch, Wang, Fritsche-Danielson, Chien, Hansson, Mullick, de Boer and van der Meer2021; Deiman et al., Reference Deiman, Bomer, van der Meer and Grote Beverborg2022). This particular founder variant is associated with a malignant form of DCM, commonly encountered in the Netherlands.

Although substantial progress has been made in demonstrating the feasibility and potential of genome editing in cellular and murine models, numerous unanswered questions remain prior to advancing to human trials involving currently available techniques. Key considerations include ensuring the safety of viral delivery and accurately targeting the vector to the intended site with appropriate dosage (Colella et al., Reference Colella, Ronzitti and Mingozzi2018).

Precision therapy in gene elusive disease

Whilst genetic therapies hold great promise for those with monogenic disease, the majority will have little relevance for the majority of patients with DCM without a rare variant genetic cause. This group of patients are likely to have a diverse range of disease mechanisms including activation of fibroinflammatory pathways and metabolic dysfunction, driven by extrinsic causes including toxic insults, inflammatory or metabolic disease as well as genetic susceptibility related to common genetic variation (Reichart et al., Reference Reichart, Lindberg, Maatz, Miranda, Viveiros, Shvetsov, Gartner, Nadelmann, Lee, Kanemaru, Ruiz-Orera, Strohmenger, DeLaughter, Patone, Zhang, Woehler, Lippert, Kim, Adami, Gorham, Barnett, Brown, Buchan, Chowdhury, Constantinou, Cranley, Felkin, Fox, Ghauri, Gummert, Kanda, Li, Mach, McDonough, Samari, Shahriaran, Yapp, Stanasiuk, Theotokis, Theis, van den Bogaerdt, Wakimoto, Ware, Worth, Barton, Lee, Teichmann, Milting, Noseda, Oudit, Heinig, Seidman, Hubner and Seidman2022). Characterising these mechanisms in individual patients using precision phenotyping may help guide targeted therapy. The integration of advanced cardiac imaging and biomarkers offer huge potential to individualise management.

Cardiac imaging

Whilst echocardiography (TTE) serves as the initial modality for diagnosing patients with heart failure with reduced ejection fraction, it is unable to reliably discriminate the cause of left ventricular dysfunction. Much data supports the use of cardiac magnetic resonance (CMR) imaging as a valuable tool for discriminating between ischaemic and non-ischaemic aetiologies and refining the cause and mechanism of non-ischaemic LV dysfunction (Japp et al., Reference Japp, Gulati, Cook, Cowie and Prasad2016; Halliday, Reference Halliday2022). It does so through detailed tissue characterisation using late gadolinium enhancement (LGE) imaging and parametric mapping (Japp et al., Reference Japp, Gulati, Cook, Cowie and Prasad2016; Halliday, Reference Halliday2022; Merlo et al., Reference Merlo, Gagno, Baritussio, Bauce, Biagini, Canepa, Cipriani, Castelletti, Dellegrottaglie, Guaricci, Imazio, Limongelli, Musumeci, Parisi, Pica, Pontone, Todiere, Torlasco, Basso, Sinagra, Filardi, Indolfi, Autore and Barison2023). This insight currently provides important information that guides selection of patients for ICDs and may also help individualise other treatment decisions in the future.

LGE represents replacement myocardial fibrosis and is present in around one-third to a half of cases (Kuruvilla et al., Reference Kuruvilla, Adenaw, Katwal, Lipinski, Kramer and Salerno2014; Di Marco et al., Reference Di Marco, Anguera, Schmitt, Klem, Neilan, White, Sramko, Masci, Barison, McKenna, Mordi, Haugaa, Leyva, Rodriguez Capitan, Satoh, Nabeta, Dallaglio, Campbell, Sabate and Cequier2017). LGE presence has been found to be a predictor of mortality, hospitalisation, and sudden cardiac death (SCD). Furthermore, the presence, extent, and patterns of LGE may provide additional valuable predictive information regarding malignant ventricular arrhythmias (VAs) or left ventricular (LV) reverse remodelling (Kuruvilla et al., Reference Kuruvilla, Adenaw, Katwal, Lipinski, Kramer and Salerno2014). The presence of LGE is now included in guidelines for primary prevention ICD implantation (McDonagh et al., Reference McDonagh, Metra, Adamo, Gardner, Baumbach, Bohm, Burri, Butler, Celutkiene, Chioncel, Cleland, Coats, Crespo-Leiro, Farmakis, Gilard, Heymans, Hoes, Jaarsma, Jankowska, Lainscak, Lam, Lyon, McMurray, Mebazaa, Mindham, Muneretto, Francesco Piepoli, Price, Rosano, Ruschitzka, Kathrine Skibelund and Group2021; Zeppenfeld et al., Reference Zeppenfeld, Tfelt-Hansen, de Riva, Winkel, Behr, Blom, Charron, Corrado, Dagres, de Chillou, Eckardt, Friede, Haugaa, Hocini, Lambiase, Marijon, Merino, Peichl, Priori, Reichlin, Schulz-Menger, Sticherling, Tzeis, Verstrael, Volterrani and Group2022). The pattern of myocardial fibrosis on CMR may also point towards particularly genetic aetiologies. Variants in desmoplakin (DSP) and filamin C (FLNC) have been shown to be associated with ring-like patterns of myocardial fibrosis which has been associated with worse outcomes (Augusto et al., Reference Augusto, Eiros, Nakou, Moura-Ferreira, Treibel, Captur, Akhtar, Protonotarios, Gossios, Savvatis, Syrris, Mohiddin, Moon, Elliott and Lopes2020). Parametric mapping with CMR also offers the ability to quantify interstitial changes, including fibrosis and oedema. Another exciting emerging fibrosis imaging technique is 68-gallium-labelled fibroblast activation protein inhibitor (FAPI) positron emission tomography (PET). This nuclear technique offers the potential to image fibrosis activity, anticipate fibrotic remodelling and prevent clinical disease before it occurs using targeted anti-fibrotic therapies.

³¹Phosphorus magnetic resonance offers the unique ability to study myocardial energetics in vivo. Studies have confirmed that DCM is characterised by a decrease in the ratio of phosphocreatine to adenosine triphosphate, a marker of impaired energetics (Stoll et al., Reference Stoll, Clarke, Levelt, Liu, Myerson, Robson, Neubauer and Rodgers2016). This has been shown to improve with reverse remodelling and predict outcome (Neubauer et al., Reference Neubauer, Horn, Cramer, Harre, Newell, Peters, Pabst, Ertl, Hahn, Ingwall and Kochsiek1997). This technique offers the ability to characterise the metabolic phenotype of individual patients and perhaps identify those who may gain most benefit from targeted metabolic therapies.

Diffusion tensor CMR enables comprehensive evaluation of cardiac microstructure revealing intricate details of myocardial wall mechanics, including the rotational torsion of myocardial sheetlets. This emerging technique may offer unique insight into the response to therapies targeting the sarcomere (Nielles-Vallespin et al., Reference Nielles-Vallespin, Khalique, Ferreira, de Silva, Scott, Kilner, McGill, Giannakidis, Gatehouse, Ennis, Aliotta, Al-Khalil, Kellman, Mazilu, Balaban, Firmin, Arai and Pennell2017).

Blood biomarkers

Circulating biomarkers provide the opportunity to characterise metabolic derangement, collagen turnover as well inflammatory and immune activation (Rubis et al., Reference Rubis, Dziewiecka, Wisniowska-Smialek, Banys, Urbanczyk-Zawadzka, Krupinski, Mielnik, Karabinowska and Garlitski2022). This has the potential to guide therapy decisions. One potential disadvantage is that many are not cardiac-specific. For example, circulating serum biomarkers of fibrosis reflect collagen turnover not only in the heart but also in various organs such as vessels, liver, and bone. Nevertheless, the carboxy-terminal propeptide of procollagen type I (PICP) and the amino-terminal propeptide of procollagen type III (PIIINP) have been correlated cardiac fibrosis observed on histology (Izawa et al., Reference Izawa, Murohara, Nagata, Isobe, Asano, Amano, Ichihara, Kato, Ohshima, Murase, Iino, Obata, Noda, Okumura and Yokota2005; Lopez et al., Reference Lopez, Gonzalez and Diez2010; Rubis et al., Reference Rubis, Dziewiecka, Wisniowska-Smialek, Banys, Urbanczyk-Zawadzka, Krupinski, Mielnik, Karabinowska and Garlitski2022) and elevated levels of these peptides predict an unfavourable outcome in patients with HF (Martos et al., Reference Martos, Baugh, Ledwidge, O’Loughlin, Murphy, Conlon, Patle, Donnelly and McDonald2009; Sweeney et al., Reference Sweeney, Corden and Cook2020; Cleland et al., Reference Cleland, Ferreira, Mariottoni, Pellicori, Cuthbert, Verdonschot, Petutschnigg, Ahmed, Cosmi, Brunner La Rocca, Mamas, Clark, Edelmann, Pieske, Khan, McDonald, Rouet, Staessen, Mujaj, Gonzalez, Diez, Hazebroek, Heymans, Latini, Grojean, Pizard, Girerd, Rossignol, Collier and Zannad2021). There has been interest in using markers to select patients who may benefit the most from anti-fibrotic therapy (Cleland et al., Reference Cleland, Ferreira, Mariottoni, Pellicori, Cuthbert, Verdonschot, Petutschnigg, Ahmed, Cosmi, Brunner La Rocca, Mamas, Clark, Edelmann, Pieske, Khan, McDonald, Rouet, Staessen, Mujaj, Gonzalez, Diez, Hazebroek, Heymans, Latini, Grojean, Pizard, Girerd, Rossignol, Collier and Zannad2021; Raafs et al., Reference Raafs, Verdonschot, Henkens, Adriaans, Wang, Derks, Abdul Hamid, Knackstedt, van Empel, Diez, Brunner-La Rocca, Brunner, Gonzalez, Bekkers, Heymans and Hazebroek2021). Galectin-3 is another marker of fibro-inflammatory activity and has been identified as a prognostic marker due to its association with worse outcomes in DCM (Sharma et al., Reference Sharma, Pokharel, van Brakel, van Berlo, Cleutjens, Schroen, Andre, Crijns, Gabius, Maessen and Pinto2004; Besler et al., Reference Besler, Lang, Urban, Rommel, Roeder, Fengler, Blazek, Kandolf, Klingel, Thiele, Linke, Schuler, Adams and Lurz2017). It appears likely that fibrosis plays an important role in driving early disease in particular phenotypes. The extent to which biomarkers will be able to guide therapy prior to the emergence of symptomatic DCM is unknown. One advantage of using them in susceptible individuals or those with early disease is that extra-cardiac causes of fibrosis are less likely to be relevant in this younger, less co-morbid group. Additionally, other markers such has high-sensitivity troponin T (hsTnT) and N-terminal prohormone brain natriuretic peptide (nt-proBNP) may have an important role in predicting disease progression (Chmielewski et al., Reference Chmielewski, Michalak, Kowalik, Franaszczyk, Sobieszczanska-Malek, Truszkowska, Stepien-Wojno, Biernacka, Foss-Nieradko, Lewandowski, Oreziak, Bilinska, Kusmierczyk, Tesson, Grzybowski, Zielinski, Ploski and Bilinska2020; Suresh et al., Reference Suresh, Martens and Tang2022). Both, for example, have been associated with the risk of malignant ventricular arrythmias in LMNA mutation carriers (Figure 4).

Figure 4. Precision medicine for dilated cardiomyopathy.

Precision phenotyping in DCM

Another key challenge is integrating these multidimensional data in a simple, accessible way to create a ground truth for the patient we see in clinic. Several studies have used unbiased clustering analysis known as phenomapping, in patients with various forms of heart failure including DCM, to help define subgroups of patients (Shah et al., Reference Shah, Katz, Selvaraj, Burke, Yancy, Gheorghiade, Bonow, Huang and Deo2015; Verdonschot et al., Reference Verdonschot, Merlo, Dominguez, Wang, Henkens, Adriaens, Hazebroek, Mase, Escobar, Cobas-Paz, Derks, van den Wijngaard, Krapels, Brunner, Sinagra, Garcia-Pavia and Heymans2020; Tayal et al., Reference Tayal, Gregson, Buchan, Whiffin, Halliday, Lota, Roberts, Baksi, Voges, Jarman, Baruah, Frenneaux, Cleland, Barton, Pennell, Ware, Cook and Prasad2022). The heterogenous aetiology of DCM makes it imminently suitable for this form of classification. Tayal et colleagues used a machine-learning based approach to cluster patients based on clinical, imaging, genetic and circulating characteristics and identified distinct subclasses of DCM with shared and distinct disease mechanisms (Tayal et al., Reference Tayal, Gregson, Buchan, Whiffin, Halliday, Lota, Roberts, Baksi, Voges, Jarman, Baruah, Frenneaux, Cleland, Barton, Pennell, Ware, Cook and Prasad2022). Verdonschot and colleagues used a similar approach incorporating transcriptomics to identify distinct transcriptomic profiles, including, pro-fibrotic, pro-inflammatory and metabolic subtypes (Verdonschot et al., Reference Verdonschot, Merlo, Dominguez, Wang, Henkens, Adriaens, Hazebroek, Mase, Escobar, Cobas-Paz, Derks, van den Wijngaard, Krapels, Brunner, Sinagra, Garcia-Pavia and Heymans2020). Both groups then used common clinical variables to discriminate between the groups so that this approach could be translated more easily into clinical practice.

By untangling the upstream causes and downstream active processes unique to each patient, such approaches may illuminate the targets for therapeutic intervention. The heterogeneity of DCM necessitates a personalised approach, with treatment strategies designed to benefit the individual patient subgroups that emerge from thorough phenotypic characterisation.

Co-morbidities and lifestyle

In the individualised treatment of individuals with DCM, it is important to also manage comorbidities such as coronary artery disease, hypertension, diabetes, thyroid disease, anaemia, and obesity (Reichart et al., Reference Reichart, Magnussen, Zeller and Blankenberg2019; Verdonschot et al., Reference Verdonschot, Hazebroek, Ware, Prasad and Heymans2019; Zeppenfeld et al., Reference Zeppenfeld, Tfelt-Hansen, de Riva, Winkel, Behr, Blom, Charron, Corrado, Dagres, de Chillou, Eckardt, Friede, Haugaa, Hocini, Lambiase, Marijon, Merino, Peichl, Priori, Reichlin, Schulz-Menger, Sticherling, Tzeis, Verstrael, Volterrani and Group2022). It is likely that such co-morbidities interact with intrinsic susceptibility to develop contractile impairment. Whether more intensive treatment and stricter control of these issues improves outcomes remains unclear. Special attention should also be paid to the impact of alcohol and cardiotoxic chemotherapy, such as anthracyclines (Ware et al., Reference Ware, Amor-Salamanca, Tayal, Govind, Serrano, Salazar-Mendiguchia, Garcia-Pinilla, Pascual-Figal, Nunez, Guzzo-Merello, Gonzalez-Vioque, Bardaji, Manito, Lopez-Garrido, Padron-Barthe, Edwards, Whiffin, Walsh, Buchan, Midwinter, Wilk, Prasad, Pantazis, Baski, O’Regan, Alonso-Pulpon, Cook, Lara-Pezzi, Barton and Garcia-Pavia2018; Andersson et al., Reference Andersson, Schou, Gustafsson and Torp-Pedersen2022; Tayal et al., Reference Tayal, Verdonschot, Hazebroek, Howard, Gregson, Newsome, Gulati, Pua, Halliday, Lota, Buchan, Whiffin, Kanapeckaite, Baruah, Jarman, O’Regan, Barton, Ware, Pennell, Adriaans, Bekkers, Donovan, Frenneaux, Cooper, Januzzi, Cleland, Cook, Deo, Heymans and Prasad2022). Whilst it is clear that excessive amounts of alcohol may be harmful, it is debatable whether low or moderate levels of consumption lead to adverse remodelling and unclear whether abstinence should be recommended (Andersson et al., Reference Andersson, Schou, Gustafsson and Torp-Pedersen2022). It is possible that specific genotypes may lead to increased susceptibility to cardiotoxins (Ware et al. 2018). Individualised exercise prescription is another important factor to consider. Patients with symptomatic DCM or features of increased risk should avoid engaging in high-intensity or competitive sports (Pelliccia et al., Reference Pelliccia, Solberg, Papadakis, Adami, Biffi, Caselli, La Gerche, Niebauer, Pressler, Schmied, Serratosa, Halle, Van Buuren, Borjesson, Carre, Panhuyzen-Goedkoop, Heidbuchel, Olivotto, Corrado, Sinagra and Sharma2019). There is particular concern for those with high-risk genotypes.