CMR reclassifies the majority of patients with suspected MINOCA and non MINOCA

Abstract

Aims

In ∼5–15% of all cases of acute coronary syndromes (ACS) have unobstructed coronaries on angiography. Cardiac magnetic resonance (CMR) has proven useful to identify in most patients the underlying diagnosis associated with this presentation. However, the role of CMR to reclassify patients from the initial suspected condition has not been clarified. The aim of this study was to assess the proportion of patients with suspected MINOCA, or non-MINOCA, that CMR reclassifies with an alternative diagnosis from the original clinical suspicion.

Methods and results

A retrospective cohort of patients in a tertiary cardiology centre was identified from a registry database. Patients who were referred for CMR for investigation of suspected MINOCA, and a diagnosis pre- and post-CMR was recorded to determine the proportion of diagnoses reclassified. A total of 888 patients were identified in the registry. CMR reclassified diagnosis in 78% of patients. Diagnosis of MINOCA was confirmed in only 243 patients (27%), whilst most patients had an alternative diagnosis (73%): myocarditis n = 217 (24%), Takotsubo syndrome n = 115 (13%), cardiomyopathies n = 97 (11%), and normal CMR/non-specific n = 216 (24%).

Conclusion

In a large single-centre cohort of patients presenting with ACS and unobstructed coronary arteries, most patients had a non-MINOCA diagnosis (73%) (myocarditis, Takotsubo, cardiomyopathies, or normal CMR/non-specific findings), whilst only a minority had confirmed MINOCA (27%). Performing CMR led to reclassifying patients’ diagnosis in 78% of cases, thus confirming its important clinical role and underscoring the clinical challenge in diagnosing MINOCA and non MINOCA conditions.

Graphical Abstract

The role of cardiac magnetic resonance (CMR) imaging in reclassification of diagnosis in patients investigated for suspected acute coronary syndromes (ACSs) with non-obstructive coronary arteries.

Open in new tabDownload slide

See the editorial comment for this article ‘With a single ‘wave of the wand': cardiovascular magnetic resonance transforms the diagnostics of acute coronary syndrome’, by G. Markousis-Mavrogenis et al., https://doi.org/10.1093/ehjci/jead215.

Introduction

In ∼5–15% of presentations of acute coronary syndromes (ACS),^1–5 patients have unobstructed coronary arteries, or myocardial infarction with non-obstructive coronary arteries (MINOCA). Whilst recognition and clinical management of ACS with obstructive coronary disease is well established, there is no current widely accepted diagnostic pathway for patients presenting with ACS and unobstructed coronary arteries. Recent studies have highlighted the importance of identifying the underlying aetiology of a presentation that can have an heterogenous aetiology, from myocardial infarction to acute myocarditis and Takotsubo syndrome (TS), to mention the most frequent causes.^6,7 Appropriate differentiation of underlying aetiologies is important in both guiding downstream investigations and subsequent clinical management.

Previous studies have sought to outline clear diagnostic pathways utilising invasive and non-invasive imaging.^6,8 The role of cardiac magnetic resonance (CMR) imaging has been widely accepted to provide accurate diagnosis and prognosis of both ischaemic and non-ischaemic pathologies.⁹ This is particularly relevant in the differentiation of MINOCA in identifying those who have a true presentation of myocardial infarction with underlying coronary aetiology.^10,11 This can be assessed with intra-coronary imaging [optical coherence tomography (OCT) or intravascular ultrasound (IVUS)], although the expertise and availability for these diagnostic tests is not always widely available, or routinely undertaken in clinical practice. Whilst similar limitations may occur with CMR, it remains an easily accessible test for many clinicians and provides an alternative non-invasive test to assess suspected MINOCA.

Mimics of MINOCA such as TS, acute myocarditis, and other cardiomyopathic processes have specific myocardial tissue characterisation changes distinct from ischaemic patterns,^12–14 and thus can be easily differentiated on CMR without the need for further invasive investigation.

The importance of performing CMR early following presentation has previously been highlighted.^15–17 This is not only to facilitate accurate diagnosis of MINOCA, but also to enhance the diagnostic yield of identifying other pathologies that may mimic this presentation. Furthermore, earlier recognition of patients will guide downstream diagnostic investigations and facilitates tailored clinical management. In combination, routine clinical tests and CMR (performed <14 days from presentation), results in a diagnostic yield as high as 74%.¹⁵ Further studies have compared more acute scanning (<3 days) demonstrating higher diagnostic yields.^18,19 Regardless of timing of scan, CMR is recognised in both the European and American guidelines^20,21 as an important diagnostic tool in investigating patients with suspected MINOCA and guiding appropriate clinical practice. However, the extent of which CMR is reclassifying patients with suspected MINOCA or non MINOCA from the initial clinical suspicion is not completely established.

Objectives

The aim of this study was to assess the proportion of patients with suspected MINOCA or non MINOCA that CMR reclassifies with an alternative diagnosis.

Methods

Study population

The cohort is formed from the ‘Bristol CMR Registry’. The database includes consecutive patients, over the age of 18, seen at the Bristol Heart Institute for suspected MINOCA between September 2011 and December 2021. Patients were identified from the electronic clinical reporting system. All patients included had undergone a diagnostic invasive coronary angiogram, followed by CMR at our tertiary cardiology centre. Those with obstructive coronary disease (≥50% stenosis) on angiography, and those with no documented troponin result or a negative troponin on review on clinical notes, were excluded from the database (Figure 1). Data was collected by the clinical team and collated into an anonymised database for use by the research team. As this was an analysis of an anonymised clinical database, individual patient consent was not required. The ‘Bristol CMR Registry’ has received ethics approval (REC 21/WS/0072).

Clinical characteristics

Data was collected on demographics, body surface area, body mass index, risk factors, and co-morbidities. Cardiac risk factors were recorded as part of routine clinical documentation from clinical referral details and available electronic notes.

All-cause mortality was tracked to December 2022, 12 months post last data entry into the registry.

Diagnostic investigations

Comprehensive CMR imaging was performed on a 1.5 Tesla Scanner (Magnetom Avanto, Siemens Healthineers) with previously described standard CMR protocols.^22,23 In brief, the CMR imaging protocol included steady state free precession, T2 short tau inversion recovery image acquisition and post-contrast imaging following gadolinium-based contrast agent injection. Early gadolinium enhancement was acquired 2 min after contrast administration (gadolinium dose 0.15 mmol/kg) and late gadolinium enhancement (LGE) was performed 10–15 min post contrast administration. Images were acquired in standard long and short axis views.

Image analysis was performed using CVi 42 (Circle Cardiovascular Imaging, Canada). All studies were reviewed and assessed by a Level 3 European Association of Cardiovascular Imaging accredited consultant.

Diagnostic definitions

Working diagnosis was based on clinical referral details provided on the CMR request and recorded in the database, with ‘uncertain’ diagnosis defined by those without a clear clinical diagnosis from the clinical history or presenting diagnostic investigations performed at the index admission. The cohort was grouped by final CMR diagnoses—myocardial infarction, myocarditis, TS, cardiomyopathy (including hypertrophic cardiomyopathy, dilated cardiomyopathy, cardiac amyloidosis, and other non-ischaemic pathologies), and normal/non-specific.

The CMR diagnosis considered clinical information and presence of myocardial oedema and/or myocardial inflammation. Myocardial oedema was assessed using T2-weighted imaging sequences, and myocardial inflammation with T1-weighted imaging (predominantly gadolinium enhancement). Other sequences relevant to the clinical diagnosis, e.g. abnormal gadolinium kinetics on inversion time (TI scout) in cardiac amyloidosis were also considered where the relevant diagnoses were suspected.

Myocardial infarction was defined by either a subendocardial or transmural late enhancement pattern¹⁴ in a coronary distribution, with or without regional wall motion abnormalities and with and without corresponding myocardial oedema. Acute myocarditis was defined using the 2018 modified Lake Louise Criteria.²⁴

TS was defined as myocardial oedema not pertaining to a specific coronary artery distribution and absence of LGE, as outlined in the 2018 International Takotsubo registry diagnostic criteria.²⁵ If myocardial oedema was found to be present without LGE, but in the distribution of a typical coronary artery, this was considered a myocardial infarction, presumably aborted.

Normal/non-specific patients were those without any regional wall motion abnormalities (excluding dys-synchrony), and absence of myocardial oedema and LGE (non-specific enhancement of the LV/RV insertion points not considered pathological). Cardiomyopathies, e.g. hypertrophic cardiomyopathy, dilated cardiomyopathy, and cardiac amyloidosis, etc., were defined by prior established CMR tissue characterisation and clinical features of disease.¹⁴

Other parameters

Blood results collected routinely as part of the patient’s clinical admission were recorded in the database. This included admission bloods such as full blood count, C-reactive protein (CRP), lipid profile and peak troponin.

Statistical analysis

Statistical analysis was performed using Stata version 17 (StatCorp LLC, Texas United States). Continuous data are presented as mean ± SD or median [interquartile range (IQR)], depending on the normality of their distribution, as assessed by visual inspection of the histograms. Comparisons of continuous data were performed using the student’s t-test or the Mann–Whitney U test.

Categorical data are presented as counts and proportions. Comparisons between categorical data were performed using chi-square test. Comparisons across multiple groups were performed using the Kruskal–Wallis with Dunn’s post hoc test (Bonferroni method).

Results

A final cohort of 888 patients was included in this study (Figure 1). The mean age across the cohort was 57 years (SD ±15.9 years) with 51% female patients. Demographic data and cardiac risk factors across the cohort are summarised in (Table 1). Hypertension and dyslipidaemia were different across diagnostic groups although other cardiac risk factors were generally similar. Overall all-cause mortality was 8.0% with a median follow-up time of 1496 days (4.1 years). Median scan interval across the cohort was 21 days (IQR 6–55). The highest peak troponin and CRP in those with a CMR diagnosis of myocarditis (Table 2).

Reclassification of diagnosis by CMR

All patients underwent CMR to determine the final diagnosis, which was not clear to the referring physician based on the available clinical history and clinical investigations (ECG, blood biomarkers, echocardiography, and coronary anatomy). In 45.6% of patients, there was a clinical suspicion that warranted confirmation, and in 54.4% clinicians were uncertain regarding the underlying aetiology of presentation based on clinical referral details, confirming the clinical challenges of diagnosing these patients clinically without a CMR. Performing CMR led to an overall reclassification of the diagnosis in 77.8% patients (Graphical Abstract).

Pre-CMR, 74 patients had an initial working diagnosis of MINOCA, which was confirmed only in 29 patients (39%) whilst 61% of these were reclassified to a non-MINOCA alternative diagnosis (Table 3). Patients with an initial working diagnosis of myocarditis, CMR reclassified 53.5% of cases (Table 3). Re-classification of diagnosis was highest in those thought to have an underlying cardiomyopathy (16 patients, 69.6%), and lowest in TS (17 patients, 26.2%) (Table 3). In those referred for CMR due to uncertain working diagnosis, CMR identified an underlying aetiology in 362 (74.9%) patients (Figure 2): the majority were found to have MINOCA (n = 151 patients, 31.3%), n = 91 myocarditis, n = 56 TS and n = 64 cardiomyopathy (Table 3).

Reclassification of diagnosis was high both in patients in whom CMR was performed <14 days and in patients in whom CMR was performed ≥14 days (71.8% vs. 82.6%, respectively, P = 0.0851) (Table 4). Patients undergoing CMR earlier (<3 days from admission) had lower reclassification (n = 92, 60.9%) vs. those scanned slightly later (<14 days from admission) (n = 282, 71.8%) (P < 0.001), i.e. reduced diagnostic uncertainty (see Supplementary data online, Table S1). A normal CMR was more commonly found in patients with CMR performed ≥14 days (n = 167, 33.7%) vs. patients with CMR < 14 days (n = 49, 12.4%), P < 0.001 (see Supplementary data online, Table S2). In the whole population with suspected ACS and unobstructed coronary arteries, only 243 patients (27.4%) were diagnosed with MINOCA, whilst in the majority of patients (n = 645. 72.6%) non-MINOCA diagnoses were established by CMR.

Discussion

In this single tertiary centre study, we present a large cohort of patients with suspected ACS and unobstructed coronary arteries investigated by CMR. This cohort extends on previous data published^15,26 from our group, focussing on the incremental value of CMR in the diagnostic work up of these patients, underscoring the clinical challenge of identifying the correct diagnosis without the use of CMR. The study also unveils that the percentage of MINOCA diagnosis is small (27.4%) compared to the diagnosis of non-MINOCA aetiologies (72.6%), despite the emphasis on international guidelines on the diagnosis of MINOCA.^20,21

We have demonstrated that the use of CMR leads to a reclassification of diagnosis in nearly 80% of patients. This supports the results of a recent meta-analysis, but their pooled data could not provide insight on the initial suspected diagnosis and how CMR provided a diagnostic reclassification in each group.²⁷

Reclassification was highest in those thought to have an underlying cardiomyopathy, followed by myocardial infarction and myocarditis. The number of patients thought to have infarction prior to CMR was far lower than those following CMR and demonstrates the role of CMR in correctly identifying these patients. Whilst information on the specific coronary aetiology was not available in this cohort, this demonstrates many patients who could potentially be missed as an infarction if not further evaluated with downstream diagnostic tests. Authors have previously commented on the role of CMR as part of a non-invasive diagnostic pathway for suspected MINOCA, particularly in combination with intravascular coronary imaging such as OCT and IVUS.^6,28 Furthermore, whilst CMR confirmed the diagnosis in Takotsubo in as many as 73.8% of all cases, it also highlighted that some of these patients may indeed have an underlying MINOCA, as reflected in 10 of our patients. The differences in a final diagnosis of infarction vs. myocarditis, or infarction vs. TS, have major implications on both ongoing clinical management and outcomes.²⁹ The current ESC guidelines²⁰ focus on using CMR to understand the aetiology of suspected MINOCA without an obvious cause. Yet in our cohort, we demonstrate that even when there is a suspected non-MINOCA diagnosis, CMR can still alter the diagnosis, including to MINOCA itself, and so its use should be considered more widely in all patients presenting with myocardial injuries and unobstructed coronary arteries. It is crucial to utilise the correct diagnostic pathways to establish the underlying diagnosis. Whilst the SMINC studies^18,19 highlight the value of higher diagnostic yield in very early scanning (<3 days), we demonstrate that it remains important to scan patients at any time point, recognising that CMR can reclassify diagnosis at any given time, whether performed earlier (<3 days or <14 days) or later ≥14 days, even as diagnostic yield decreases with increased scan interval.

It should be noted that in more than half of patients when referred for CMR had an uncertain diagnosis, based on clinical referral details. This represents a significant proportion of patients where the clinical certainty of MINOCA is unclear. In our cohort, this reflects in part the maturity our clinical CMR service, both in terms of experience and ease of access, and therefore CMR being more frequently used in suspected cases of ACS with unobstructed coronaries. Nonetheless, it is clear from our results that CMR is important in reclassifying these patients to a final diagnosis and therefore impacting their clinical management.

Whilst it is important to not miss a diagnosis of myocardial infarction, it is equally important to not misdiagnose these patients and expose patients to unnecessary harm from side effects of antiplatelet therapy, such as gastrointestinal bleeding. CMR not only confirms or identifies those with MINOCA, but importantly those with an alternative non-MINOCA diagnosis. In our cohort, around three-quarters of patients had an alternative final diagnosis to MINOCA by CMR. Similar to previous studies,⁹ those with other cardiomyopathies in this cohort had the highest all-cause mortality, thus highlighting the importance to adequately identify these patients adequately and treating them accordingly. The focus in current ESC guidelines²⁰ is on the MINOCA diagnosis, whilst with this study we aim at shifting the attention of the non-MINOCA diagnosis which represented the largest cohort of patients presenting with ACS and unobstructed coronary arteries that clinicians should be aware of. Moreover, the role of CMR in this patient cohort extends beyond the diagnosis to the prognosis.^9,27 This confirms that CMR has an important clinical impact and powerful clinical tool.

By reclassifying patients with CMR, clinicians can accurately determine the underlying aetiology of suspected MINOCA cases. This will allow tailored clinical management and facilitate appropriate risk stratification for these patients, including modifying and improving secondary prevention medication and downstream clinical follow-up. It is also particularly important in recognising those with diagnoses where prognosis may be markedly different, allowing early detection of their disease process and thereby facilitating appropriate and specific clinical treatment, and where possible improve longer-term prognosis.

Study limitations

The Bristol CMR registry is derived from patients who have undergone CMR for investigation with myocardial injuries and non-obstructed coronary arteries over a period of 10 years. Clinical practice has evolved over that time, and so some may not have been considered for CMR and thus not included in the registry. The time period also covered the COVID-19 pandemic and thus may not have recruited all potential patients, therefore potentially underestimating the prevalence of each diagnosis. Nonetheless, this cohort still represents a large number of patients from a single tertiary referral centre covering a large geographical area with high volumes of CMR performed, and with a long follow up time.

Unfortunately, parametric mapping was not included as standard imaging protocol for all patients and thus information derived from novel tissue characterisation could not be assessed from this dataset. Nonetheless, this is likely to be more representative of many patients undergoing CMR as mapping is not largely available to all centres depending on access to resources and/or expertise. Diagnosis by CMR can still be made without parametric mapping and therefore its absence does not preclude accurate CMR diagnosis. Likewise, other cardiac biomarkers such as NTproBNP have not been routinely performed in the cohort and so its value alongside CMR cannot be assessed.

Specific underlying coronary pathology of those with myocardial infarction demonstrated on CMR was not available in this cohort. In most patients, intra-coronary imaging was not performed with only a small number of patients undergoing IVUS/OCT. The combined role of these tests in this cohort cannot therefore be assessed but should be considered in future studies.

Conclusion

In a large single-centre cohort of patients presenting with ACS unobstructed coronary arteries, CMR was able to reclassify diagnosis in nearly 80% of patients. In addition, the majority of patients with suspected ACS with unobstructed coronary arteries have an alternative non-MINOCA diagnosis (73%) (myocarditis, Takotsubo, cardiomyopathies, or normal CMR/non-specific findings), whilst only a minority had confirmed MINOCA (27%). We highlight the strength of CMR in the diagnostic pathway of patients with suspected ACS with unobstructed coronary arteries, both in reclassifying diagnosis and identifying alternative underlying aetiologies with potential implications for downstream clinical management.

Acknowledgements

The authors thank the team of cardiovascular magnetic resonance radiographers in the Bristol Heart Institute for their help in acquiring the cardiovascular magnetic resonance studies.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

None declared.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes