Multi-modality imaging to Determine Underlying Causes of MINOCA in Women and Men

Abstract

Background:

Myocardial infarction with non-obstructive coronary arteries (MINOCA) has several underlying causes, including mimicking conditions in some cases. Imaging is recommended to identify MINOCA etiologies, but it remains unclear which patients are most likely to have abnormal findings. We characterized MINOCA mechanisms, analyzed predictors of imaging abnormalities and explored sex differences.

Methods:

We enrolled patients with clinical diagnosis of MI in an international, prospective, diagnostic study at 28 sites in US, Canada and UK. After a women-only phase, we included both sexes. Individuals with ≥50% diameter stenosis or coronary dissection on angiography, or alternate causes for the clinical presentation, were excluded. Participants had multi-vessel coronary optical coherence tomography (OCT) during index coronary angiography and cardiac magnetic resonance imaging (CMR) within one week. Independent core laboratories interpreted imaging, blinded to other results.

Results:

Among 754 patients enrolled, 389 had MINOCA and 336 with MINOCA underwent OCT (270 women and 66 men); CMR was completed in 284 (85%). An OCT-defined culprit lesion was identified in 45% (116/270 women [43%] and 35/66 men [53%], p=0.18). CMR demonstrated an ischemic pattern in 114/284 (40%), similar by sex (96/225 women [43%] vs. 18/59 men [31%], p=0.12). A non-ischemic pattern was observed in 23% (23% of women, 25% of men, p=0.78). We identified a cause of the clinical presentation in 79% of patients with both tests completed: 59% had an ischemic cause of MINOCA and 20% had a non-ischemic mimicking condition. OCT alone found a MINOCA etiology in 151/336 (45%) and CMR alone in 180/284 (63%). Predictors of an OCT culprit lesion included age, abnormal angiogram, and number of vessels imaged, but 27% of normal angiograms harbored a culprit lesion. Predictors of abnormal CMR were peak troponin, shorter time to CMR, and non-Asian race, but CMR was abnormal in 40% when troponin was <4-fold above the upper reference limit.

Conclusions:

The combination of multi-vessel coronary OCT and CMR in patients with a clinical diagnosis of MINOCA confirmed MI in 59% and identified an alternate cause (MINOCA mimic) in 20%. Clinical factors had limited utility to predict imaging abnormalities. No sex differences in imaging results were detected.

Clinical Trial Registration:

NCT02905357 https://clinicaltrials.gov/ct2/show/NCT02905357

Introduction

Up to 15% of patients presenting clinically with spontaneous myocardial infarction (MI) have no major narrowing on coronary angiography, and are given a working diagnosis of MI with no obstructive coronary arteries (MINOCA).¹ Patients with MINOCA experience major adverse cardiovascular events at a rate of approximately 10% at 1 year, and 24% at 4 years, with 5-year mortality ^~11%.^2–4 Acute coronary syndrome guidelines recommend imaging to determine the underlying cause of MINOCA, recognizing that some cases initially diagnosed as MINOCA have vascular and cardiac abnormalities diagnostic of MI while others have non-ischemic mimicking conditions, such as myocarditis or takotsubo syndrome.^5–8

We previously reported that the combination of intracoronary imaging and cardiac magnetic resonance imaging (CMR) identified a cause of the MINOCA presentation in 85% of women in a prospective cohort study.⁹ Prediction of coronary culprit lesions and of abnormal CMR using clinical variables was limited, suggesting all patients would require both imaging tests to define the underlying cause of the MINOCA presentation, but sample size was limited.

MINOCA is three times more common among women as compared to men with MI, yet because men are over twice as likely to experience MI, MINOCA is also frequently observed in men.^10,11 There are further reasons to suspect sex differences in mechanisms of MINOCA: men with a provisional diagnosis of MINOCA have higher likelihood of myocarditis than women, and in general, women have less severe atherosclerosis than men.^12,13,14

We built on our prior prospective diagnostic observational cohort study, more than doubling the prior sample size and enrolling men, with the objectives of 1) improving precision in prediction of underlying causes of MINOCA; and 2) investigating sex differences in MINOCA mechanisms.

Methods

Anonymized data and materials will be made publicly available at Zenodo.org. The Heart Attack Research Program was a prospective, observational, multi-center diagnostic study applying standardized imaging protocols to patients with a clinical diagnosis of MINOCA between 2016–2025. The primary objective was to determine the proportion of patients with vascular causes of MINOCA, as identified by coronary optical coherence tomography (OCT), and the proportion of patients with CMR findings that could explain the clinical presentation. A secondary objective was to determine whether results of imaging differed by sex. The study was initially designed as a single sex, women-only study. The results of this phase were previously reported.⁹ After a renewal of funding, we expanded enrollment to include both sexes (“extension phase”). We also included 15 participants (10 women, 5 men) from a pilot prospective phase conducted 2014–2015.

Eligible participants with no ≥50% stenosis in a major epicardial vessel were to undergo OCT during the diagnostic angiogram performed to evaluate clinically diagnosed MI. The goal of OCT was to image all major epicardial arteries. CMR was to be conducted within 1 week of the acute presentation using a study-specified protocol. Patients with MINOCA could be enrolled after diagnostic angiography if clinically indicated testing included multi-vessel coronary optical coherence tomography (OCT) and CMR according to the study protocol.

Images from OCT, CMR and angiography were interpreted at core laboratories, blinded to sex, detailed clinical, ECG and laboratory information, and results of the other imaging studies.

OCT culprit lesions were characterized as plaque rupture, plaque erosion, intraplaque hemorrhage, healed plaque, calcified nodule, or spontaneous coronary artery dissection, according to published recommendations, by the core laboratory.^15,16 We selected OCT for the study rather than intravascular ultrasound due to superior resolution of OCT for plaque rupture and thrombus.¹⁷

CMR results were categorized as ischemic if there was evidence of: i) MI based on subendocardial late gadolinium enhancement (LGE) with/without associated rest perfusion abnormality and T1 mapping; or ii) regional ischemic injury (e.g., myocardial edema in one coronary territory, typically with an associated perfusion abnormality and/or regional wall motion abnormality).^18–20 CMR results were categorized as non-ischemic when there was either i) evidence of myocarditis based on revised Lake Louise criteria²¹; pericarditis based on pericardial edema and/or LGE;²² ii) non-ischemic cardiomyopathy with reduced LV systolic function;²³ iii) constellation of findings consistent with takotsubo syndrome (e.g., circumferential myocardial edema without LGE and a typical wall motion pattern, particularly with right ventricular involvement);²⁴ or iv) evidence of infiltrative cardiomyopathy (based on abnormal gadolinium kinetics, elevated global native T1 and ECV values, and typical patterns of global/circumferential LGE with/without RV and atrial involvement)^25,26 or hypertrophic cardiomyopathy (regional left ventricular hypertrophy with maximal wall thickness of ≥15 mm, with associated abnormalities, e.g., systolic anterior motion of the mitral valve, resting LV outflow tract flow acceleration, patchy LGE).²⁷

The study was approved by the NYU Grossman School of Medicine Institutional Review Board, and by each site’s local institutional review board or ethics committee. All participants provided informed consent.

Participant Eligibility

Clinical Eligibility.

Patients aged ≥21 years with MI meeting the 4^th universal definition²⁸ who were referred for clinically indicated coronary angiography were considered for enrollment. Participants were enrolled after testing either contemporary or high sensitivity cardiac troponin assays. Exclusion criteria: alternate explanation for troponin elevation (e.g., acute heart failure, tachyarrhythmia, hypertensive urgency, pulmonary embolism, renal failure); prior history of obstructive CAD including history of percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG); recent use of vasospastic agents, such as cocaine, triptans, or ergot alkaloids (≤1 month); estimated glomerular filtration rate <45 mL/min or contraindication to additional contrast needed for OCT imaging; pregnancy; takotsubo syndrome as the most likely clinical diagnosis; and fibrinolytic therapy for the qualifying MI event. Patients with MI thought clinically to be due to supply-demand mismatch were not enrolled.

Angiographic Eligibility.

Patients with ≥50% stenosis in any major coronary artery as determined by the treating physicians or with contraindication to OCT after diagnostic angiography (e.g., contrast allergy, excessive coronary tortuosity, angiographically evident coronary dissection) were excluded from research imaging.

OCT Image Acquisition

OCT imaging was performed using a commercially available OCT system (Dragonfly OPTIS or OpStar Imaging Catheters, Abbott Vascular, Santa Clara, CA). Intracoronary nitroglycerin and anticoagulation were given prior to imaging. Physicians were requested to image at least the proximal 60 mm of the left anterior descending, left circumflex, and right coronary arteries, using automated pullback. The goal was to image all major arteries supplying the left ventricle. If the left circumflex was small in caliber, for example, a ramus branch, obtuse marginal branch, or appropriate diagonal branch was to be imaged instead.

CMR Image Acquisition

Sites utilized 1.5 or 3 Tesla (T) magnetic resonance imaging systems with phased-array receiving coils and cardiac gated pulse sequences. The study protocol included cine imaging for left and right ventricular function, late gadolinium enhancement (LGE) with phase-sensitive inversion recovery and inversion time selected to null normal myocardium, qualitative T2-weighted imaging, T2 mapping, resting first pass myocardial perfusion, and T1 mapping pre and post gadolinium contrast injection using the modified look-locker inversion recovery (MOLLI) sequence, or a similar sequence depending on MRI vendor and system.²⁹ See Supplemental Methods for details.

Determination of the cause of MINOCA based on OCT and CMR

A diagnosis of MI was assigned if: i) CMR identified infarction or regional injury in a coronary territory; or ii) OCT identified a culprit lesion. A non-ischemic cause was assigned if CMR demonstrated myocarditis, pericarditis, non-ischemic cardiomyopathy, infiltrative disease, takotsubo syndrome, pulmonary embolism, or hypertrophic cardiomyopathy with evidence of acute myocardial edema. In the small number of cases with a coronary culprit lesion on OCT in which CMR revealed a non-ischemic cause, the CMR and OCT core laboratories reviewed all images together to arrive at a consensus final diagnosis, incorporating clinical data at this stage if needed. If no cause was identified using either imaging test, MINOCA was classified as idiopathic.

Statistical Analyses

We computed descriptive statistics of clinical and laboratory characteristics as well as findings of OCT and CMR. They are presented as median and inter-quartile range (IQR) for continuous variables, or frequency and percentage for categorical variables. Comparisons between patients with and without OCT culprit lesions, with or without abnormal CMR, and by sex used the chi-square or Fisher’s exact test in the case of categorical variables as appropriate and the Mann-Whitney U test for continuous variables. Comparison between OCT and CMR findings was performed using McNemar’s test for the 2×2 comparison and chi-square test for 3×2 comparison. We performed multivariable logistic regression modeling to assess the odds of an OCT-defined culprit lesion, and the odds of having abnormal CMR, adjusting for demographics, clinical characteristics, and laboratory values. Sex was included as the primary independent variable. Variables with p-values of <0.1 in univariate analyses were selected as candidate predictors for entry into multivariable models.³⁰ Variables with sparse distribution was omitted in the multivariable analyses. All analyses were performed in R (Version 4.4.1). A two-sided p-value <0.05 was considered to be statistically significant.

Power for sex differences was projected at 80% based on planned enrollment of 250 women and 100 men, with OCT culprit lesion prevalence estimated at 30% among women and 46% among men.

Results

Participant Characteristics

A total of 754 patients were enrolled at 28 sites, including 539 women (244 during the extension phase) and 215 men (Figure 1). Among these, 389 participants had MINOCA (317 women and 72 men). OCT was not performed in 48 MINOCA participants, most commonly due to contraindications such as tortuous vessels or contrast allergy. Three OCT image sets were not interpretable for the presence or absence of a culprit lesion, and two were lost. See Table S1 for clinical characteristics of patients with MINOCA who were excluded due to no OCT or uninterpretable OCT.

Figure 1. Recruitment and reasons for screen failure by sex.

Participants were enrolled from 2016–2025 (initial women-only phase, 2016–2020; extension phase with women and men, 2020–2025). 15 patients (10 women, 5 men) enrolled prospectively during a pilot phase in 2014–2015 are also included here. Reasons for ineligibility are shown. Some sites screened exclusively after angiography and enrolled patients who had OCT conducted for clinical indications.

MINOCA screen failures did not have OCT due to a contraindication that was discovered during angiography such as small arteries or excessive tortuosity, risk of additional contrast (e.g., high contrast load, allergy or volume overload), or unexpected technical difficulties (e.g., OCT console failure). Patients with takotsubo syndrome were not to be enrolled in the extension phase, thus some patients in whom this was the most likely diagnosis after angiography were excluded. Angiographic evidence of spontaneous coronary artery dissection was also an exclusion criterion. Three patients provided consent before angiography and ultimately did not undergo angiography. Two OCT image sets were not included in archiving at the site due to technical failure and could not be transmitted to the core laboratory for evaluation.

*includes 10 pilot patients recruited at NYU, during a time when screening logs were not maintained.

CMR = cardiac magnetic resonance imaging; OCT = coronary optical coherence tomography; MI-CAD = myocardial infarction with ≥stenosis in a major epicardial vessel; MINOCA = myocardial infarction with non-obstructive coronary arteries (<50% stenosis in all major epicardial vessels); SCAD = Spontaneous coronary artery dissection.

Women were less likely than men to be excluded based on ≥50% angiographic stenosis due to atherosclerotic coronary artery disease (MI-CAD) discovered on the index angiogram, 37% vs. 61%, p<0.001.

The final study cohort included 336 patients with a clinical diagnosis of MINOCA and interpretable OCT imaging (270 women and 66 men). The median age was 58 years (interquartile range [IQR] 50–67); 63% identified as White, and 20% as Hispanic (Table 1). The median peak cardiac troponin was 32-fold greater than the upper reference limit (IQR 11–111 fold). ST-segment elevation was reported on the electrocardiogram in 5%. Echocardiography demonstrated a segmental wall motion abnormality in 94/270 (35%). Sites rated the coronary angiogram as normal in 156/336 participants (46%). The median maximal stenosis was 10% (IQR 0–30%) as assessed at sites, and 30% (IQR 24–38%) as assessed by the angiographic core laboratory.

Table 1:

Demographic and clinical characteristics of the study cohort overall, by sex, and by OCT and CMR findings.

	Overall N = 336	Women N = 270	Men N = 66	p-value	OCT Culprit lesion N = 151	No OCT culprit lesion N = 185	p-value	CMR abnormal N=180	CMR Normal N=104	p-value
Age, years	58 (50, 67)	59 (51, 68)	55 (46, 64)	0.034	63 (53, 71)	55 (46, 64)	<0.001	59 (49, 68)	56 (50, 65)	0.218
Sex				NA			0.141			0.182
Women					116/151 (77%)	154/185 (83%)		147/180 (82%)	78/104 (75%)
Men					35/151 (23%)	31/185 (17%)		33/180 (18%)	26/104 (25%)
Race				0.211			0.865			0.056
Asian	20/336 (6.0%)	18/270 (6.7%)	2/66 (3.0%)		8/151 (5.3%)	12/185 (6.5%)		6/180 (3.3%)	12/104 (12%)
Black	56/336 (17%)	45/270 (17%)	11/66 (17%)		26/151 (17%)	30/185 (16%)		31/180 (17%)	15/104 (14%)
Other	47/336 (14%)	42/270 (16%)	5/66 (7.6%)		19/151 (13%)	28/185 (15%)		24/180 (13%)	13/104 (13%)
White	213/336 (63%)	165/270 (61%)	48/66 (73%)		98/151 (65%)	115/185 (62%)		119/180 (66%)	64/104 (62%)
Ethnicity				0.084			0.249			0.672
Hispanic/Latino	67/335 (20%)	59/270 (22%)	8/65 (12%)		26/151 (17%)	41/184 (22%)		54/283 (19%)	33/180 (18%)
Troponin
Initial troponin ratio	7 (2, 33)	9 (3, 33)	3 (2, 15)	0.007	6 (2, 40)	7 (2, 26)	0.726	12 (3, 65)	4 (2, 15)	<0.001
Peak troponin ratio	32 (11, 111)	33 (12, 123)	24 (5, 60)	0.013	31 (8, 108)	33 (12, 115)	0.581	57 (16, 176)	16 (7, 39)	<0.001
Angiography result (site)				0.201			<0.001			0.021
Abnormal	180/336 (54%)	140/270 (52%)	40/66 (61%)		109/151 (72%)	71/185 (38%)		87/180 (48%)	65/104 (63%)
Normal	156/336 (46%)	130/270 (48%)	26/66 (39%)		42/151 (28%)	114/185 (62%)		93/180 (52%)	39/104 (38%)
Worst stenosis severity	10 (0, 30)	10 (0, 30)	15 (0, 30)	0.226	30 (0, 35)	0 (0, 20)	<0.001	0 (0, 30)	20 (0, 30)	0.069
Worst stenosis severity (core laboratory)	0.30 (0.24, 0.38) N=332	0.30 (0.25, 0.37) N=266	0.31 (0.24, 0.41)	0.291	0.33 (0.26, 0.41)	0.28 (0.23, 0.36); N=181	<0.001	0.30 (0.24, 0.39); N=179	0.31 (0.25, 0.38); N=102	0.950
Echocardiography
Wall motion abnormalities	94/270 (35%)	78/216 (36%)	16/54 (30%)	0.371	48/119 (40%)	46/151 (30%)	0.091	60/145 (41%)	22/80 (28%)	0.038
LVEF	60 (53, 65); N=272	60 (52, 65); N=216	60 (55, 65) ;N=56	0.553	60 (51, 65); N=120	60 (54, 65); N=152	0.868	58 (49, 65); N=149	60 (55, 65); N=78	0.017
Medical history
Diabetes	60/336 (18%)	50/270 (19%)	10/66 (15%)	0.522	38/151 (25%)	22/185 (12%)	0.002	32/180 (18%)	18/104 (17%)	0.920
Hypertension	160/336 (48%)	126/270 (47%)	34/66 (52%)	0.480	86/151 (57%)	74/185 (40%)	0.002	84/180 (47%)	45/104 (43%)	0.580
Dyslipidemia	128/335 (38%)	104/270 (39%)	24/65 (37%)	0.812	73/150 (49%)	55/185 (30%)	<0.001	65/180 (36%)	43/103 (42%)	0.348
Smoking				0.288			0.469			0.454
Current	50/333 (15%)	39/269 (14%)	11/64 (17%)		24/150 (16%)	26/183 (14%)		27/179 (15%)	16/102 (16%)
Former	78/333 (23%)	59/269 (22%)	19/64 (30%)		39/150 (26%)	39/183 (21%)		38/179 (21%)	28/102 (27%)
Never	205/333 (62%)	171/269 (64%)	34/64 (53%)		87/150 (58%)	118/183 (64%)		114/179 (64%)	58/102 (57%)
Prior MI	30/335 (9.0%)	28/270 (10%)	2/65 (3.1%)	0.064	16/151 (11%)	14/184 (7.6%)	0.341	16/180 (8.9%)	9/104 (8.7%)	0.946
Depression	71/329 (22%)	60/264 (23%)	11/65 (17%)	0.308	30/148 (20%)	41/181 (23%)	0.601	31/175 (18%)	26/102 (25%)	0.123
Anxiety	74/329 (22%)	62/265 (23%)	12/64 (19%)	0.424	32/148 (22%)	42/181 (23%)	0.732	38/177 (21%)	27/101 (27%)	0.319
ECG findings (according to site)				0.342			0.742			0.037
NSTEMI	316/333 (95%)	256/268 (96%)	60/65 (92%)		143/150 (95%)	173/183 (95%)		164/178 (92%)	102/104 (98%)
STEMI	17/333 (5.1%)	12/268 (4.5%)	5/65 (7.7%)		7/150 (4.7%)	10/183 (5.5%)		14/178 (7.9%)	2/104 (1.9%)
Enrollment period				NA			0.797			0.002
Women only	145/336 (43%)	145/270 (54%)	0/66 (0%)		64/151 (42%)	81/185 (44%)		86/180 (48%)	30/104 (29%)
Women and men	191/336 (57%)	125/270 (46%)	66/66 (100%)		87/151 (58%)	104/185 (56%)		94/180 (52%)	74/104 (71%)
Number of vessels with OCT				0.281			0.032			0.672
1 or 2	121/336 (36%)	101/270 (37%)	20/66 (30%)		45/151 (30%)	76/185 (41%)		63/180 (35%)	39/104 (38%)
3 or 4	215/336 (64%)	169/270 (63%)	46/66 (70%)		106/151 (70%)	109/185 (59%)		117/180 (65%)	65/104 (63%)
Timing
Days from MI to CMR	6 (3, 10; N=285	6 (3, 9; N=226	7.5 (4, 11; N=59	0.046	6(3, 10 ;N=123	6(4, 11); N=162	0.452	6 (3, 9); N=177	8 (5, 13); N=101	<0.001
Days from MI to OCT	2 (1, 3); N=335	2 (1, 3)	2 (1, 3); N=65	0.557	2 (1, 4)	2 (1, 3); N=184	0.058	2 (1, 3); N=180	2 (1, 3); N=104	0.136
Duration of symptoms on day of admission, hours	2 (1, 7); N=320	3 (1, 11); N=255	1 (0, 3); N=65	0.028	2 (1, 11); N=144	2 (1, 6); N=176	0.889	2 (1, 8); N=172	2 (1, 6); N=99	0.153
Symptoms
Chest pain	299/336 (89%)	238/270 (88%)	61/66 (92%)	0.320	19/151 (13%)	18/185 (9.7%)	0.406	159/180 (88%)	91/104 (88%)	0.835
Arm pain	127/335 (38%)	102/269 (38%)	25/66 (38%)	0.995	94/150 (63%)	114/185 (62%)	0.845	74/179 (41%)	37/104 (36%)	0.338
Back pain	70/334 (21%)	58/268 (22%)	12/66 (18%)	0.536	117/149 (79%)	147/185 (79%)	0.835	33/179 (18%)	24/104 (23%)	0.348
Jaw pain	70/335 (21%)	58/269 (22%)	12/66 (18%)	0.545	122/150 (81%)	143/185 (77%)	0.366	39/179 (22%)	22/104 (21%)	0.901
Dyspnea	156/336 (46%)	125/270 (46%)	31/66 (47%)	0.922	84/151 (56%)	96/185 (52%)	0.494	77/180 (43%)	56/104 (54%)	0.072
Vital signs on admission	N=330	N=264			N=150	N=180		N=175	N=103
Heart rate, bpm	73 (66, 87)	73 (66, 87)	73 (63, 88)	0.360	76 (65,88)	72 (66,85)	0.459	74 (67,86)	72 (63, 85)	0.105
Systolic blood pressure, mmHg	133 (120, 153)	133 (119, 153)	137 (122, 152)	0.592	135 (122, 158)	132 (118, 152)	0.163	134 (121, 152)	131 (113, 152)	0.277
Diastolic blood pressure, mmHg	79 (69, 89)	77 (68, 88)	84 (72, 92)	0.028	77 (68, 88)	79 (69, 89)	0.269	79 (70, 89)	76 (66, 87)	0.072

Values are median (IQR) or N (%). Troponin values are a multiple of the upper reference limit.

LVEF: left ventricular ejection fraction; MI: myocardial infarction; NSTEMI: non-ST-elevation myocardial infarction; STEMI: ST-elevation myocardial infarction; OCT: optical coherence tomography; CMR: cardiac magnetic resonance imaging

Women were older than men (59 years [IQR 51–68] vs. 55 years [IQR 47–64], p=0.034) and had higher peak cardiac troponin concentrations (33-fold the upper reference limit [IQR 12–122] vs. 24 [IQR 5–59], p=0.013). Women presented with a longer duration of ischemic symptoms on the day of admission (median 3 hours vs 1 hour, p=0.006) and had lower diastolic blood pressure than men. There were no other sex differences in clinical and angiographic characteristics.

OCT Findings

Cardiac catheterization with OCT imaging was performed a median of 2 calendar days from the onset of MI [IQR 1–3 days]. OCT was performed in all 3 major epicardial coronary arteries in 215 (64%) participants, 2 arteries in 99 (29%), and 1 artery in 23 (7%). Reasons OCT was not performed in all vessels included small or tortuous arteries (N=32), coronary spasm (N=9), thrombus noted (N=2), iatrogenic dissection or wire entanglement (N=3), operator preference or not reported (N=76).

An OCT-defined culprit lesion was identified in 151/336 (45%) of participants, including 116/270 women (43%) and 35/66 men (53%), p=0.141 (Table 2). Culprit lesion types included intraplaque hemorrhage (17% of the cohort), layered plaque (12%), plaque rupture (7%), plaque erosion (5%), eruptive calcified nodule (2%), thrombus without plaque (1%) and spontaneous coronary artery dissection (0.6%). There were no differences in this distribution by sex (Figure 2). Representative OCT culprit lesions are shown in Figures S1-S6. An additional 61 participants had non-culprit intraplaque hemorrhage or healed plaque (examples in Figure S7). These do not appear as intraplaque hemorrhage or healed plaque in data tables.

Table 2.

OCT and CMR findings by sex.

	Overall (n=336)	Women (n=270)	Men (n=66)	p-value
OCT findings
Days from MI to OCT	2 (1, 3)	2 (1, 3)	2 (1, 3)	0.148
Culprit lesion	151/336 (45%)	116/270 (43%)	35/66 (53%)	0.141
Culprit lesion type				0.095
Calcified nodule	6/336 (1.8%)	2/270 (0.7%)	4/66 (6.1%)
SCAD	2/336 (0.6%)	2/270 (0.7%)	0/66 (0%)
Healed plaque	40/336 (12%)	31/270 (11%)	9/66 (14%)
Intra plaque hemorrhage	58/336 (17%)	49/270 (18%)	9/66 (14%)
Plaque erosion	17/336 (5.1%)	13/270 (4.8%)	4/66 (6.1%)
Plaque rupture	24/336 (7.1%)	16/270 (5.9%)	8/66 (12%)
Thrombus without plaque	4/336 (1.2%)	3/270 (1.1%)	1/66 (1.5%)
CMR performed	284 (84.5%)	270 (83.3%)	59 (89.4%)
Days from MI to CMR	6.0 (4.0, 11.0); N=285	6.0 (4.0, 10.0); N=226	7.0 (4.0, 12.0); N=59	0.289
CMR abnormal				0.223
Ischemic CMR finding	114/284 (40%)	96/225 (43%)	18/59 (31%)
Non-ischemic CMR finding	66/284 (23%)	51/225 (23%)	15/59 (25%)
Normal CMR	104/284 (37%)	78/225 (35%)	26/59 (44%)
CMR detailed findings				0.044
Infarction	79/284 (28%)	64/225 (28%)	15/59 (25%)
Regional injury	35/284 (12%)	32/225 (14%)	3/59 (5.1%)
HCM	6/284 (2.1%)	3/225 (1.3%)	3/59 (5.1%)
Myocarditis	28/284 (9.9%)	22/225 (9.8%)	6/59 (10%)
Non-ischemic Cardiomyopathy	16/284 (5.6%)	10/225 (4.4%)	6/59 (10%)
Takotsubo Syndrome	9/284 (3.2%)	9/225 (4.0%)	0/59 (0%)
Other	7/284 (2.5%)	7/225 (3.1%)	0/59 (0%)
Normal	104/284 (37%)	78/225 (35%)	26/59 (44%)
Final diagnosis based on OCT and CMR	N=284	N=225	N=59	0.472
MI	167/284 (59%)	131/225 (58%)	36/59 (61%)
Mimic (non-ischemic alternate diagnosis)	56/284 (20%)	44/225 (20%)	12/59 (20%)
Idiopathic	61/284 (21%)	50/225 (22%)	11/59 (19%)

Values are n (%) or median (IQR)

MI: myocardial infarction; OCT: optical coherence tomography; CMR: cardiac magnetic resonance imaging

Figure 2. Findings of OCT and CMR in patients with a clinical diagnosis of MI and no obstructive CAD in the HARP study, overall and by sex.

Top panel, OCT core laboratory determined findings. A schematic illustrating the lesion is shown above an example of each culprit lesion type and a normal vessel. P values are not provided due to the small sample size in several cells. P value for overall distribution by sex = 0.095. Bottom panel, CMR core laboratory determined findings. An example of each diagnosis is shown. The representative image for other diagnoses shows the case of pulmonary embolism. P values are not provided due to the small sample size in several cells. P value for overall distribution by sex = 0.044.

CM = cardiomyopathy; CMR = cardiac magnetic resonance imaging; MINOCA = myocardial infarction with no obstructive coronary arteries; OCT = optical coherence tomography

The likelihood of a culprit lesion was similar in the women-only phase of the study and the phase with both sexes enrolled (64/145 [44%] vs. 87/192 [45%], p=0.843). The likelihood of a culprit lesion was similar when participants were enrolled before or after cardiac catheterization with OCT (Table S2).

Three coronary dissections resulted from OCT (0.9%), all requiring stenting for treatment. One was associated with a small hemopericardium that did not require drainage.

Participants with an OCT culprit lesion were older than those without a culprit lesion (median 63 years [IQR 53–71] vs 55 [46–64], p<0.001) and more likely to have diabetes, hypertension, and dyslipidemia (Table 1). The presence of a culprit lesion was not associated with peak troponin or left ventricular ejection fraction. There was a culprit lesion on OCT in 27% of participants with site-determined normal coronary angiography vs. 61% of those with angiography reported as showing any degree of atherosclerosis, p<0.001. The maximal diameter stenosis was higher in those with vs. without a culprit lesion on OCT as determined by sites (median 30% vs. 0%, p<0.001) and by the core laboratory (median 33% vs. 28%, p<0.001).

Multivariable analysis of predictors of OCT culprit lesion

On multivariable analysis, older age, site-determined abnormal angiography, and performance of 3-vessel OCT imaging were associated with higher odds of identifying a culprit lesion by OCT (Table 3).

Table 3.

Multivariable analysis of OCT culprit lesion.

	OR [95% CI]	p
Age, per 10 years	1.63 [1.29, 2.08]	<0.001
Abnormal angiogram per site	3.69 [2.25, 6.14]	<0.001
Number of vessels imaged by OCT (3 vs. 2 or 1)	2.07 [1.23, 3.53]	0.006
Female vs male sex	0.63 [0.33, 1.19]	0.156

Variables included in the model in addition to above: diabetes, dyslipidemia, hypertension, days from MI to OCT.

OCT; optical coherence tomography

Findings on Cardiac Magnetic Resonance Imaging

CMR was performed in 289 and interpretable in 284 of the 336 participants (85%; 225 women and 59 men). CMR was performed a median of 6 days from MI onset (IQR, 4–11). An ischemic pattern of LGE (i.e., infarction) was identified in 79 participants (28%), including 64 women (28%) and 15 men (25%), p=0.87. Among participants with infarction evident on CMR (N=79), the median infarct size was 4.26 grams (IQR 2.44–6.70 grams). Regional injury (i.e., myocardial edema in a single coronary territory) was observed in an additional 35 participants (12%; 14% of women and 5% of men). Non-ischemic findings were identified in 66 participants (23%; 23% of women and 25% of men). Women were more likely than men to have regional injury and less likely to have nonischemic cardiomyopathy (p=0.044 for distribution). Representative CMR images are shown in Figures S8-S13. See Table S3 for the completion rate for each part of the CMR protocol and Table S4 for reasons CMR was not done.

Abnormal CMR was more likely when CMR was done earlier after MI (median 6 days [IQR 3–9] vs. 8 [5–12.5] with normal CMR, p<0.001). Abnormal CMR was also associated with higher peak troponin, lower left ventricular EF on echocardiography (58 [IQR 49–65] vs. 60 [IQR 56–65], p=0.01), and STEMI (88% of STEMI cases vs. 62% of NSTEMI cases, p=0.037), as well as Asian race. There was a trend toward higher likelihood of segmental wall motion abnormalities on echo with abnormal vs. normal CMR (73% vs. 59%, p=0.055). Among participants with echocardiography done, CMR was more likely to be normal when echo was normal (56/131 [43%] vs. 24/96 [25%] with reduced ejection fraction or segmental wall motion abnormalities, p=0.006), but CMR was abnormal in most cases with normal echocardiography. CMR was more frequently abnormal in the women-only phase of the study (86/116 [74%], vs. 94/168 [56%] of women in the extension phase, p=0.002, Table S5).

There was no association between the presence of an OCT culprit lesion and the frequency of abnormal CMR (65% with vs. 62% without a culprit lesion, p=0.111, Table 5) but ischemic CMR findings were more common in association with an OCT culprit lesion (52% vs. 32% of those without a culprit lesion on OCT, p<0.001). See Figure 3 for the frequency of abnormal CMR based on peak troponin ratio when divided into deciles. In the lowest decile (described by peak troponin ratio <4 as a multiple of the upper reference limit), the likelihood of abnormal CMR was 38%.

Table 5.

Relationship between OCT and CMR findings.

OCT and CMR performed (n=284)	OCT culprit lesion present (n=124)	No culprit lesion on OCT (n=160)	P value
Abnormal CMR (n=180)	81 (65.3%)	99 (61.9%)	0.111*
Ischemic CMR findings (n=114)	64 (51.6%)	50 (31.3%)	<0.001 †
Non-ischemic CMR findings (n=66)	17 (13.7%)	49 (30.6%)
Normal CMR (n=104)	43 (34.7%)	61 (38.1%)

^*P value for abnormal vs normal CMR;

^†P value for Ischemic CMR vs. non-ischemic CMR vs. normal CMR

Figure 3. Frequency of Abnormal CMR by Cardiac Troponin Decile.

Cardiac troponin concentration as a multiple of the upper reference limit by decile: lowest decile, <4 fold elevation; second decile, 4–9.2 fold; third decile, 9.3–14.3 fold; fourth decile, 14.6–22.55 fold; fifth decile, 22.57–34 fold; sixth decline, 36.7–55.4 fold; seventh decile, 56.4–86.6 fold; eighth decile, 90.8–162.9 fold; ninth decile, 164.6–351.5 fold; top decile, >352 fold.

Multivariable analysis of predictors of abnormal CMR

On multivariable analysis, abnormal CMR was associated with higher peak troponin, shorter time from MI to CMR, and non-Asian race (Table 4). Myocarditis on CMR tended to be more likely among patients with abnormal angiography, White race, or younger age, but not by sex. (Table S6).

Table 4.

Multivariable analysis of abnormal CMR.

	OR [95% CI]	p
Asian vs White	0.14 [0.03, 0.54]	0.008
Black or African American vs White	0.79 [0.34, 1.85]	0.586
Other vs White	0.70 [0.27, 1.87]	0.462
Peak troponin ratio (log transformed)	1.47 [1.20, 1.83]	<0.001
Days from MI to CMR, per day	0.95 [0.91, 0.99]	0.015
Female vs male sex	1.33 [0.61, 2.86]	0.472

Variables included in the model in addition to above: abnormal vs. normal angiogram per site, diastolic blood pressure, abnormal wall motion on echocardiography, dyspnea. Race was collected by self-report. When the presence of an OCT culprit lesion was added to the model, it was not a significant predictor of abnormal CMR and other associations did not change.

CMR = cardiac magnetic resonance imaging; MI = myocardial infarction.

Combined Imaging Findings (OCT and CMR)

Among 284 participants with OCT and CMR, 223 had an abnormality on one or both studies (79%). The mechanism of the MINOCA presentation was determined to be vascular (i.e., final diagnosis remained MI) in 167 (59%). The final diagnosis was changed from MINOCA to a non-ischemic alternative diagnosis (MINOCA mimic) in 56 (20%). No mechanism was identified in the remaining 61 (21%) (Figure 4). Patients with a final diagnosis of idiopathic MINOCA were younger than those with MI or a MINOCA mimic, had lower peak troponin, and had longer time from MI onset to CMR (Table S7). Those with a final diagnosis of MI were least likely to have a normal angiogram. Patients with a final diagnosis of a MINOCA mimic had lower EF on echocardiography, when performed, than those with a final diagnosis of MI or idiopathic MINOCA.

Figure 4. Synthesis of OCT and CMR findings to a final diagnosis.

Other non-ischemic CMR findings included pericardial disease in 5 (2%), concentric remodeling with a mild increase in T1 in 1 (0.4%), pulmonary embolism in 1 (0.4%).

CM = cardiomyopathy; CMR = cardiac magnetic resonance imaging; OCT = coronary optical coherence tomography; MI-CAD = myocardial infarction

The proportion of patients with an identified cause of the MINOCA presentation with OCT and CMR, 79%, (95% CI 73–83%), was higher than with OCT alone (124/284 = 44%, 95% CI 38–49%; p <0.001) or CMR alone (180/284 = 63%, 95% CI 58–69%; p <0.001). The likelihood of a culprit lesion on OCT when both CMR and angiography were normal was 21% (8/39).

The association between OCT and CMR findings is shown in Table 5 and Figure 5. There was CMR-detected infarction or regional injury in 64 participants of 124 with an OCT culprit lesion (52%). A total of 17 participants with an OCT culprit lesion (14%) had a non-ischemic abnormality on CMR. After consensus review, the final diagnosis was a MINOCA mimic in 7 cases and MI in 10 of these cases (Table S8). In cases with MI on CMR and OCT imaging of the vessel serving the MI territory, the culprit vessel by OCT matched the myocardial territory with MI 81% of the time (Table S9).

Figure 5. Detailed CMR findings according to the presence or absence of a coronary culprit lesion.

CMR = cardiac magnetic resonance imaging; OCT = coronary optical coherence tomography; SCAD = spontaneous coronary artery dissection. Obstructive CAD was defined as ≥50% coronary stenosis in a major epicardial vessel at angiography.

Among patients with an OCT culprit lesion, abnormal CMR was more common in women (81% vs. 65% of men, p=0.043) and was associated with higher troponin (55-fold upper reference limit [IQR 16–168) vs. 18-fold [IQR 7–39], p<0.001), earlier CMR (6 days [IQR 3–9) vs. 7.5 [IQR 4–11], p=0.046), and longer duration of symptoms on the day of admission (3 hours [IQR 1–11] vs 1 [IQR 0–3], p=0.046, lower EF (57 [IQR 45–65] vs. 62 [IQR 60–65], p=0.036), as well as race (p=0.036). Abnormal CMR was most common when the culprit lesion type was intraplaque hemorrhage (CMR abnormal in 79% of cases) and less common with plaque rupture (48% of cases) or calcified nodule (0%), p=0.006 for distribution (Table S10). When there were ischemic findings on CMR and the related vessel was imaged by OCT, the location of the CMR abnormality matched the independently determined culprit vessel in 81% (Table S10).

Representative cases of MI are shown in Figures 6 and 7.

Figure 6. Plaque rupture with regional ischemia in the LAD distribution.

Panel A – Coronary angiography with no obstructive CAD. The culprit lesion is located in the proximal LAD (white arrow).

Panel B - A low-intensity mural thrombus with an irregular surface is observed (white triangles) and a part of the thrombus protrudes into the lumen. The underlying plaque shows strong attenuation, consistent with the lipidic plaque (asterisk). The rupture site is marked with an orange arrow. Overall, these findings suggest plaque rupture within a relatively small amount of lipidic plaque in a large lumen, followed by mural thrombus formation.

Panel C - LV SAX LGE imaging demonstrates focal subendocardial infarct (red arrow) in mid anterior segment. The remaining segments demonstrate no infarction.

Panel D - LV short axis STIR images demonstrate increased signal within basal anterior, basal anteroseptal, basal inferoseptal, mid anterior, mid anteroseptal, mid inferoseptal, apical septal, apical anterior and apical inferior segments consistent with diffuse edema.

Panel E - AHA 16 segment native T1 bull's-eye representing corresponding increase in T1 mapping ranging from 1141 ms in basal, 1264 ms in mid and 1231 ms in the apical segments (normal T1 mapping < 1100 ms at 1.5 T scanner).

Figure 7. Healed plaque in the LAD leading to acute myocardial infarction.

Panel A - Coronary angiography revealed a large LAD coronary artery that wraps around the apex to supply the inferior wall. A healed plaque culprit lesion was identified in the mid LAD.

Panel B- A layered structure with very low and heterogeneous signal intensity (white triangles) is observed overlying the lipid-rich plaque (asterisk), consistent with a thrombus in the healing stage.

Panel C: LV SAX PSIR LGE shows transmural subendocardial LGE in the apical septal, inferior and lateral segments suggestive of Acute MI.

Panel D- T2 mapping shows (arrows) edema in the apical septal, inferior and lateral segments.

Panel E - T1 Native Mapping bull’s eye shows corresponding high T1 signal.

Discussion

In this prospective, international study of patients with a working diagnosis of MINOCA, 59% had evidence of MI on intracoronary OCT imaging, CMR, or both, when both tests were performed. Approximately 1 in 5 patients undergoing CMR had a mimicking condition rather than MINOCA; among remaining patients, 68% had imaging evidence of MI. This included an OCT-defined coronary culprit lesion in 50%, and CMR-determined infarction or regional injury in 52%, with frequent co-occurrence. Coronary and cardiac imaging with OCT and CMR provided complementary diagnostic information. For example, 44% of those with evidence of MI on CMR had no coronary culprit lesion, and among patients with normal CMR, 41% had a coronary culprit lesion. As expected, ischemic CMR findings were more common among participants with an independently identified OCT-defined culprit lesion and conversely, non-ischemic findings were associated with the absence of a culprit lesion.

We found that delineation of the underlying cause of a MINOCA presentation requires multi-modality imaging, an approach that is consistent with clinical practice guidelines.^1,5,7 Clinical characteristics showed limited ability to predict coronary culprit lesions on OCT. OCT imaging of all three coronary arteries doubled the odds of identifying a culprit lesion. The severity of nonobstructive coronary stenosis on angiography and older age were each independent predictors of a culprit lesion on OCT. Still, one-quarter of patients with normal angiography had a coronary culprit lesion, including 21% of patients with normal angiography and normal CMR. This highlights the known inaccuracy of diagnostic angiography for the presence of any atherosclerosis in patients with MINOCA.³¹ Coronary computed tomography angiography is more sensitive for atherosclerosis but does not have the necessary image resolution to determine which coronary plaques are disrupted or harbor thrombus.

Prediction of abnormal CMR was also challenging. The degree of cardiac troponin elevation was strongly associated with CMR abnormalities, yet nearly 40% of CMR studies were abnormal in patients in the lowest decile of peak cardiac troponin. The earlier CMR was performed, the higher the diagnostic yield, and 70% of incomplete CMR tests in our cohort were attributable to the inconvenience of returning to the hospital, with fewer missed due to claustrophobia or other patient factors.

Sex differences

MINOCA is more common among women with MI, leading to the hypothesis that underlying causes of MINOCA might differ by sex. We found no difference by sex in the likelihood of an OCT culprit lesion, abnormal CMR, or the final diagnosis as MINOCA, a nonischemic MINOCA mimic or idiopathic. However, power was limited based on frequent screen failure of enrolled men due to MI-CAD. The proportion of women with no obstructive coronary arteries by coronary angiography was nearly double that of men. Across a spectrum of disease from subclinical atherosclerosis to severe ischemia, men have more severe atherosclerosis than women, yet our women and men had similar severity of nonobstructive atherosclerosis with MINOCA.

We did not confirm the previous finding of higher likelihood of myocarditis among men than women with MINOCA in our prospective, multicenter study.¹² This is likely due to selection based on referral for cardiac catheterization based on strong suspicion of MI in our study, with more than half of coronary angiograms showing nonobstructive atherosclerosis.

Our deeply phenotyped cohort was too small to permit assessment of sex differences in outcomes. However, several large studies have indicated similar outcomes by sex after a MINOCA event.^32,33 This provides support for our finding that underlying mechanisms of the MINOCA presentation were similar by sex.^34,35

Coronary culprit lesions on OCT

Coronary culprit lesions were identified by OCT imaging in 45% of study participants and 50% of those without a mimicking non-ischemic condition on CMR. This prevalence is consistent with most prior prospective and retrospective series of MINOCA patients and higher than in the PROMISE trial, which included single vessel OCT by protocol.^36–41 Most coronary culprit lesions were atherothrombotic, with only two spontaneous coronary artery dissections identified. MINOCA with a coronary atherothrombotic culprit lesion is type 1 MI, and guidelines recommend treating accordingly, with antiplatelet and lipid lowering therapy.^1,28 Atherothrombotic culprit lesions are associated with worse prognosis in MINOCA.³⁶

MI due to atherothrombotic disease can occur with a spectrum of angiographic severity of disease ranging from none to mild or moderate, as demonstrated in this study, to complete occlusion. Further research is needed to investigate why some patients with smaller plaques that rupture or erode present with MINOCA, while others generate larger thrombi prompting diagnosis with MI-CAD, and still others are clinically silent.^42,43 Variability between individuals in platelet aggregation, coagulation, and the reaction of vascular endothelium and smooth muscle in the setting of plaque instability are plausible explanations for this spectrum of disease.

The most common coronary culprit lesion types in our study were intraplaque hemorrhage (38% of culprit lesions), healed plaque (26%), plaque rupture (16%), and plaque erosion (11%). Calcified nodule (4%), lone thrombus (3%), and spontaneous coronary artery dissection (1%) were each rare culprit lesion types. The morphology of these culprit lesions is similar to MI-CAD culprit lesion morphology but plaques are smaller in MINOCA patients, with lower thrombus burden, rendering imaging findings more subtle as published previously.^9,44 Intraplaque hemorrhage is a well-described coronary culprit lesion in acute coronary syndrome and sudden cardiac death, and the OCT appearance is similar to the culprit intraplaque hemorrhage lesions in our cases of MINOCA, other than the size of the culprit plaque.^45–47 The difference in OCT appearance between a culprit intraplaque hemorrhage in our study and a plaque rupture can be subtle and subject to interobserver variability. The absence of associated thrombus in intraplaque hemorrhage culprit lesions on OCT is consistent with pathologic studies. In a series of culprit lesions found at autopsy in sudden death victims, 63 of 103 ruptured plaques with intraplaque hemorrhage and severe associated stenosis had no intraluminal thrombus.⁴⁷ Healed plaque represents a slightly later healing stage of plaque rupture than intraplaque hemorrhage and is a vulnerable coronary lesion type.⁴⁸ We observed healed plaque in 27% of MINOCA patients, similar to the prevalence of this culprit lesion type on pre-intervention OCT in ACS patients with obstructive CAD.⁴² Intraplaque hemorrhage and healed plaque lesions with stable appearance were not classified as culprit lesions by the core laboratory.

The CMR findings from this study provide strong support for our conclusion that OCT-detected culprit lesions caused MINOCA, particularly when considering that OCT and CMR were interpreted by independent core laboratories, blinded to the findings of other imaging tests. For example, 62% of patients with OCT evidence of intraplaque hemorrhage with CMR completed had ischemic CMR findings. Culprit lesions that do not narrow the coronary artery severely may cause MI when there is superimposed coronary artery spasm, distal embolization of atherothrombotic material, or transient complete thrombosis that improves by the time of angiography. The latter seems less likely given the relatively low thrombus burden observed on OCT. The co-occurrence of coronary spasm and plaque rupture or erosion has been described.⁴⁹

Some patients with coronary culprit lesions had normal CMR. This is expected because sensitivity of CMR for smaller injuries can be limited, quality of CMR studies in acute coronary syndrome patients is not uniformly optimal, and CMR was completed at a median of 6 days from MI onset in our study; earlier CMR has better sensitivity.^50,51 It is unexpected that plaque rupture, the culprit morphology most likely to cause MI-CAD, was associated with a lower likelihood of abnormal CMR in our study; further investigation may shed light on the reason for this finding.

Prevalence of CMR abnormalities

The prevalence of abnormal CMR was 63% overall in this study, identical to a meta-analysis of CMR studies focused on MINOCA but lower than in our prior report limited to a smaller sample of women.³⁴ There are several potential explanations for this. First, sites in the extension phase used high sensitivity troponin assays, while most used a contemporary troponin assay in the women-only phase. Higher peak troponin was independently associated with abnormal CMR. We observed a lower frequency of regional ischemia and myocarditis in the extension phase. These findings may be the most sensitive to the magnitude of injury because they are characterized by detection of myocardial edema. Second, unlike other series investigating CMR in MINOCA, we excluded patients with a presumptive diagnoses of takotsubo syndrome or with an alternate cause of troponin elevation.

Limitations

We encountered high rates of screen failure among men due to MI-CAD, and we did not meet our projected enrollment of 100 men with MINOCA. Few patients with STEMI were included. Not all patients had 3-vessel OCT performed, 15% did not complete CMR, and imaging was technically limited in some cases. We did not perform provocative testing for coronary artery spasm as part of the research protocol because our investigators found the procedural duration in combination with multivessel OCT to be unacceptable. Prior studies have demonstrated positive coronary reactivity testing for spasm in 24–70% of MINOCA cases.^{38,41,52–58} We identified 17 cases with a culprit lesion on OCT but non-ischemic findings on CMR. We resolved these cases by consensus based on the confidence level of each imaging finding and, when needed, clinical context. This highlights the value of multidisciplinary team discussion for complex MINOCA cases. We were unable to include a control group of stable patients without obstructive CAD, due to practical and ethical challenges. Images were interpreted by highly skilled core laboratories; interpretation in local clinical practice may be subject to greater variability. The shift from contemporary to high sensitivity cardiac troponin assays at many centers over the study period may have resulted in smaller degrees of detectable acute myocardial injury and a higher rate of normal CMR in the extension phase, limiting our ability to correlate OCT and CMR findings. We did not include coronary CT angiography for comparison with OCT. This will be the subject of a future study. Some sites enrolled patients after OCT and CMR were completed, and it remains possible that selective enrollment at these sites resulted in an inflated estimate of diagnostic yield. We did not collect blood for centralized determination of troponin levels; different assays may not be fully comparable. We permitted some sites to enroll after clinically indicated OCT and did not collect a comprehensive screening log of all patients with MI at sites.

Conclusions

In this study using multi-modality imaging in patients with suspected MINOCA, the majority had imaging evidence of MI, including a coronary culprit lesion, evidence of infarction or regional ischemic injury, or both, without marked differences in these findings by sex. A mimicking nonischemic condition rather than MINOCA was present in one fifth of suspected MINOCA patients. The combination of OCT and CMR performed better than either test alone to provide information about the cause of a MINOCA presentation. The mechanisms of MINOCA are most commonly vascular in origin, with contributions from atherothrombotic disease and presumed coronary artery spasm. Our findings indicate that sex differences in the prevalence of MINOCA are more prominent than sex differences in underlying mechanisms. Our findings support current guidelines that recommend imaging to guide selection of treatment after a MINOCA presentation.

Supplementary Material

Supplemental Methods

Tables S1-S10

Figures S1-S13

HARP Research Group

Clinical Perspective.

What is New?

• In this study, the largest prospective study of myocardial infarction and no obstructive coronary arteries (MINOCA) to date (n=336), the combination of coronary optical coherence tomography and cardiac magnetic resonance imaging led to a diagnosis in 79% of patients.

• Most patients clinically diagnosed with MINOCA had a final diagnosis of myocardial infarction (59%), and an alternate diagnosis mimicking MI was present in 20%.

• Clinical factors had limited utility to predict imaging abnormalities, and no sex differences in imaging results was detected.

What are the Clinical Implications?

• The mechanisms of MINOCA are most commonly vascular in origin, with contributions from atherothrombotic disease and presumed coronary artery spasm.

• Our findings support current guidelines that recommend imaging to guide selection of treatment after a MINOCA presentation.

• Sex differences in the prevalence of MINOCA are more prominent than sex differences in underlying mechanisms.

Non-standard Abbreviations and Acronyms

CAD

coronary artery disease

CMR

cardiac magnetic resonance imaging

eGFR

estimated glomerular filtration rate

IQR

interquartile range

LGE

late gadolinium enhancement

LVEF

left ventricular ejection fraction

MI

myocardial infarction

MINOCA

myocardial infarction with no obstructive coronary arteries

NA

not applicable

NSTEMI

non-ST-elevation myocardial infarction

OCT

coronary optical coherence tomography

STEMI

ST-elevation myocardial infarction

URL

Upper reference limit

Appendix

HARP Research Group Members:

Harmony Reynolds, MD

Nathaniel Smilowitz, MD

Anais Hausvater, MD

Sripal Bangalore, MD

Judith S. Hochman, MD

Michael Attubato, MD

Lori Vales Lay, MD

Sunil Rao, MD

Binita Shah, MD

Louai Razzouk, MD

Rafael Harari, MD

Stylianos Papadakos, MD

Claudia Serrano Gomez, MD

Jeffrey Trost, MD

Alair Holden

Amanda Joa

Manuela Plazas Montana

Lillian Na

Michelle Shum

Ellen Hada, MHA, PMP

Michele Fisher, LPN,RCS

Ehtisham Mahmud, MD

Bahman Ghannadian

Catalin Toma, MD

Janet Wei, MD

Tara Sedlak, MD

Jacqueline Saw, MD

Ngaire Meadows

Laxmi Mehta, MD

Kevin Marzo, MD

Juan Gaztanaga, MD

Nisha Jhalani, MD

Dwithiya Thomas, MD

Caitlin Giesler, MD

Puja K. Mehta, MD

Fauzia A. Rashid, PhD

Kevin Bainey, MD

Bryan Har, MD

Bobak Heydari, MD

Norma Hogg

Suzanne Welsh

Jennifer A Tremmel, MD, MS

Shuangbo Liu, MD

Odayme Quesada, MD

Hayder Hashim, MD

Tej Sheth, MD

Aun Yeong Chong, MD

Derek So, MD

Natalia Pinilla Echeverri, MD

Atul Sharma, MD

Ashley Gutierrez

Yader Sandoval, MD

Akl C. Fahed, MD, MPH

Brian Case, MD

Tomas Cieza, MD

Colin Berry, MD

Akiko Maehara, MD

Mitsuaki Matsumura, MD

Raymond Y. Kwong, MD, MPH

Mobeen Ahmed, MD

Ayako Seno, MD, PhD

Yuhe Xia, MS

Chang Yu, PhD

Hua (Judy) Zhong, PhD

Bina Ahmed, MD