Spontaneous coronary artery dissection: Acute findings on coronary computed tomography angiography

Abstract

Background

The coronary computed tomography angiography features of acute spontaneous coronary artery dissection, an important cause of acute coronary syndrome in young women, have not been assessed.

Methods

The “Virtual” Multicenter Mayo Clinic Spontaneous Coronary Artery Dissection Registry was established in 2010 and includes retrospective and prospective patient data. Retrospective assessment of acute coronary computed tomography angiography images was performed for 14 patients (16 vessels) who had images performed within two days of invasive coronary angiography diagnosis of acute spontaneous coronary artery dissection.

Results

Four pertinent diagnostic coronary features of acute spontaneous coronary artery dissection were observed in order of prevalence: 1) abrupt luminal stenosis (64%); 2) intramural hematoma (50%); 3) tapered luminal stenosis (36%); and 4) dissection (14%). Additional findings include epicardial fat stranding (42%), coronary tortuosity (29%), and coronary bridge (14%). Fifty percent of patients had myocardial hypoperfusion in the myocardial distribution of the dissected coronary artery.

Conclusions

We define key coronary computed tomography angiography features of acute spontaneous coronary artery dissection, the most common of which are abrupt luminal stenosis and intramural hematoma. Importantly, intramural hematoma appears similar to noncalcified atherosclerotic plaque, emphasizing the importance of invasive coronary angiography for acute diagnosis of spontaneous coronary artery dissection until the sensitivity and specificity of coronary computed tomography angiography is better understood.

Introduction

Acute spontaneous coronary artery dissection (SCAD) is a nonatherosclerotic cause of sudden cardiac arrest and myocardial infarction, often affecting young persons without traditional cardiac risk factors.^1,2 It is the etiology of myocardial infarction in as many as 35% of women <50 years³ and a common cause of myocardial infarction in pregnant women.^1,4 The mechanism of myocardial infarction due to SCAD is different from that of atherosclerosis; instead of atherosclerotic plaque, the presence of an intramural hematoma with or without visible coronary intima/media tear leads to obstruction of coronary blood flow.^1,5–7 Percutaneous coronary intervention is less successful in SCAD as compared to atherosclerotic disease with complications such as dissection propagation, loss of flow, or worsened stenosis.^8,9 Moreover, the majority of conservatively managed SCAD vessels demonstrate partial or complete healing upon follow-up.^8,10–12 Therefore, unless patients have loss of coronary blood flow or are hemodynamically unstable, conservative therapy is preferred when feasible.^8,10,13

The increased complication rate with interventions, preference for conservative management in stable patients, and increasing concern for iatrogenic coronary dissection¹⁴ in these patients calls for the consideration of noninvasive coronary computed tomography angiography (CCTA) for the diagnosis and assessment of SCAD, including follow-up evaluations.¹⁵ There are no studies describing the features of acute SCAD on CCTA with confirmatory invasive coronary angiography (CA), nor is there CCTA terminology for description of the acute changes of SCAD vessels. Due to the paucity of published literature and knowledge regarding the appearance of acute SCAD on CCTA, the objective of this study was to describe the CCTA findings of patients with acute SCAD, strictly including only those patients whose CCTA images were within two days of confirmatory CA.

Methods

This retrospective study was approved by the Mayo Clinic Institutional Review Board. Written informed consent was obtained for participation in the “Virtual” Multicenter Mayo Clinic SCAD Registry and DNA Biorepository as described elsewhere.^16,17 Patients are included in the Mayo Clinic SCAD Registry if the CA images are consistent with a diagnosis of SCAD upon review by experienced interventional cardiologists (RG, PJMB). Of 812 Mayo Clinic SCAD Registry patients, 104 patients had CCTA images available for analysis. Patients with non-gated computed tomography (CT) angiography (e.g. pulmonary embolism protocol) were not included due to insufficient imaging of the coronary arteries.

Since SCAD can heal within days to weeks from time to presentation¹¹ and in order to focus on the acute SCAD cohort, only patients with CCTA within two days (before or after CA demonstrating SCAD) were included (Figure 1). Of those 16 patients, two had uninterpretable coronary images due to suboptimal ECG-gating and poor image quality. Six CCTA studies were performed before CA, and the others were subsequent to CA. All studies were reviewed by three interpreters with dedicated experience in interpreting CCTA (PAA, NJA, MST) in conjunction with the CA and clinical history. CCTA images were reconstructed and analyzed using post-processing software (Aquarius, TeraRecon, San Mateo, California, USA). Details regarding initial image interpretations, CT scanners and slice thickness were recorded. Patient information including demographics and clinical details was collected from the medical records and Mayo Clinic SCAD Registry database.

Figure 1.

Study patient selection process. CA: Invasive coronary angiography; CCTA: coronary computed tomography angiography; ECG: electrocardiogram; SCAD: spontaneous coronary artery dissection.

Results

A total of 14 patients and 16 vessels were included in the study with clinical characteristics outlined in Table 1. Although some patients received their acute SCAD diagnosis and care elsewhere, all were evaluated for further SCAD-specific care at the Mayo Clinic in Rochester, Minnesota, USA.

Table 1.

Baseline clinical characteristics of patients with coronary computed tomography angiography at time of acute spontaneous coronary artery dissection (SCAD).

Clinical characteristics
Female, n (%)	14 (100)
Mean age (range), years	44 (32–59)
Mean body mass index (range), m/kg2	26.5 (20–35)
White, n (%)	13 (93)
Postpartum SCAD, n (%)	1 (7)
Recurrent SCAD, n (%)	2 (14)
Hypertension, n (%)	0
Hyperlipidemia, n (%)	2 (14)
Diabetes mellitus, n (%)	0
Active tobacco, n (%)	0
Connective tissue disease, n (%)	1 (7)
Extracoronary vascular abnormalities, n (%)	8 (57)
Fibromuscular dysplasia, n (%)	5 (36)
SCAD associated with exertion, n (%)	5 (36)
SCAD associated with stress, n (%)	2 (14)

All patients were women with mean age of 44 years with few co-morbidities who presented with acute coronary syndrome; two patients had an initial normal troponin level. Six of the patients underwent CCTA before CA as part of a chest pain evaluation. The SCAD diagnosis was subsequently determined by CA (Table 2). Only two of the original CCTA reports mentioned coronary dissection as a differential diagnosis. Eight patients had a CCTA after CA (Table 3).

Table 2.

Details of patients with acute spontaneous coronary artery dissection (SCAD) who underwent coronary computed tomography angiography (CCTA) before subsequent invasive coronary angiography.

Patients with CCTA before coronary angiogram
Pt	Age	Brief history	Original CCTA Report	CT details	Slice (mm)	SCAD vessel	Coronary features	Vessel patency	Distal flow	Comments
1	45	Chest pain while running, elevated troponin, NSTEMI.	Thrombotic subtotal occlusion of first marginal.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.5	OM	Abrupt luminal stenosis. No calcium, intramural hematoma, epicardial fat stranding or dissection.	Occluded	None	Myocardial hypoperfusion in LV lateral wall.
2	38	Chest pain climbing stairs, first troponin negative then increased, NSTEMI.	Normal coronary arteries.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.75	LAD	Abrupt luminal stenosis with occlusion at segment of LAD bridging and epicardial fat stranding. No calcium, intramural hematoma or dissection.	Occluded	None	Myocardial hypoperfusion and regional wall motion abnormalities. Vessel recanalized by time of CA.
3	33	Nocturnal chest pain postpartum, EKG changes.	Hazy narrowing of proximal LAD with 60% stenosis; dissection cannot be excluded.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.6	LAD	Abrupt luminal stenosis. LAD bridging which is not adjacent to SCAD. No calcium, intramural hematoma, epicardial fat stranding or dissection.	Patent	Present	Appearance similar to noncalcified atherosclerotic plaque.
4	56	Chest pain on treadmill, elevated troponin, NSTEMI.	Focal occlusion or severe narrowing of the distal LAD is suggestive of SCAD.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.75	LAD	Abrupt luminal stenosis with intramural hematoma, epicardial fat stranding and occlusion. No calcium or dissection.	Occluded	Present	Myocardial hypoperfusion and regional wall motion abnormalities. Distal LAD fills with contrast on CCTA whereas it does not on CA.
5	59	Awoke from chest pain, elevated troponin, NSTEMI.	Mild (<50%) narrowing in the mid left circumflex secondary to low attenuation plaque.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.75	OM	Abrupt luminal stenosis. No calcium, intramural hematoma, epicardial fat stranding or dissection.	Patent	Present	Vessel irregularity which may be subtotal occlusion with reconstitution.
6	45	Chest pain, elevated troponin, NSTEMI.	Normal coronary arteries without dissection, aneurysm, or significant stenosis.	Siemens SOMATOM Definition Flash dual source 1280-slice.	0.75	RCA	Long tapered luminal stenosis change with intramural hematoma. No calcium, epicardial fat stranding or dissection.	Patent	Present	No myocardial hypoperfusion or regional wall motion abnormalities.

CA: invasive coronary angiography; EKG: electrocardiogram; LAD: left anterior descending coronary artery; LV: left ventricle; NSTEMI: non-ST-elevation myocardial infarction; OM: obtuse marginal; pt: patient; RCA: right coronary artery.

Table 3.

Details of patients with acute spontaneous coronary artery dissection (SCAD) who underwent coronary computed tomography angiography (CCTA) after subsequent invasive coronary angiography.

Patients with CCTA after coronary angiogram
Pt	Age	Brief history	CT details	Slice thickness (mm)	SCAD vessel	Coronary features	Vessel patency	Distal flow	Comments
7	40	Chest pain, elevated troponin, NSTEMI.	Siemens SOMATOM Definition 64-slice.	0.75	LAD	Intramural hematoma. No abrupt or tapered luminal stenosis calcium, epicardial fat stranding or dissection.	Patent	Present	Myocardial hypoperfusion and regional wall motion abnormalities. Appearance similar to scattered noncalcified atherosclerotic plaque.
8	39	Chest pain at rest, cardiac arrest in ED.	Siemens SOMATOM Definition Flash dual source 128-slice.	0.5	LAD	Combination abrupt and long tapered luminal stenosis with epicardial fat stranding. No calcium, intramural hematoma or dissection.	Patent	Present	No myocardial hypoperfusion.
9	53	Chest pain, STEMI.	GE 64-slice.	0.63	LAD	Long tapered luminal stenosis with intramural hematoma and epicardial fat stranding. No calcium or dissection.	Patent	Present	Myocardial hypoperfusion and regional wall motion abnormalities. Suboptimal resolution, has the appearance of noncalcified atherosclerotic plaque.
10	50	Chest pressure, NSTEMI.	GE Light speed 64-slice.	0.75	LAD	No abrupt or tapered luminal stenosis intramural hematoma, calcium, epicardial fat stranding or dissection.	Patent	Present	No myocardial hypoperfusion. Initial CA had mild abnormality in the mid LAD. Seven days later she had progressive SCAD on CA.
11	39	Chest pain, NSTEMI.	Siemens SOMATOM Definition 64-slice.	0.75	LAD	Abrupt luminal stenosis with intramural hematoma and epicardial fat stranding. No calcium or dissection.	Patent	Present	Myocardial hypoperfusion. Changes extend into diagonal but motion artifact obscures further assessment.
12	32	Cardiac arrest.	Siemens SOMATOM Definition 64-slice.	0.75	LM into LCx and LAD	Long tapered luminal stenosis with intramural hematoma and dissection. No calcium or epicardial fat stranding.	Patent	Present	No myocardial hypoperfusion or regional wall motion abnormalities.
13	53	Chest pain, first troponin negative then increased.	GE Revolution 320-slice.	0.63	LAD	Abrupt luminal stenosis. No calcium, intramural hematoma, epicardial fat stranding or dissection.	Patent	Present	No myocardial hypoperfusion or regional wall motion abnormalities.
14a	39	Chest pain, NSTEMI.	Siemens SOMATOM Definition 64-slice.	0.75	LCx	Abrupt luminal stenosis with dissection. No calcium, intramural hematoma or epicardial fat stranding.	Patent	Present	LV lateral wall myocardial hypoperfusion.
14b					LAD	Long tapered luminal stenosis with intramural hematoma and epicardial fat stranding. No calcium or dissection.	Occluded	Present	LV apical myocardial hypoperfusion.
14c					RPLA	Beaded (alternating luminal stenosis and dilatation) appearance suggestive of fibromuscular dysplasia.	Patent	Present	Small vessel with limited resolution.

CA: invasive coronary angiography; ED: emergency department; LAD: left anterior descending coronary artery; LCx: left circumflex; LM: left main; LV: left ventricle; NSTEMI: non-ST-elevation myocardial infarction; pt: patient; RPLA: right posterolateral artery; STEMI: ST-elevation myocardial infarction.

There were four primary SCAD-related coronary features observed (Figure 2): (a) intramural hematoma defined as discrete thickening within the wall of the vessel (and not within the lumen, consistent with a Type 2 or 3 SCAD)¹⁸ (Figures 3–5); (b) dissection visualized as a linear hypodensity as would be seen with contrast within the arterial wall consistent with Type 1 SCAD¹⁸ (Figure 4); (c) abrupt luminal stenosis (>50% diameter change over a length of 0.5 mm) (Figure 5); (d) tapered luminal stenosis (>50% diameter change over a length of 5 mm) (Figures 3 and 4).

Figure 2.

Features of acute spontaneous coronary artery dissection on coronary computed tomography angiography.

Figure 3.

Intramural hematoma and long tapered luminal stenosis ((a) and (b), arrows) on coronary computed tomography angiography (CCTA) of acute left anterior descending (LAD) coronary artery spontaneous coronary artery dissection. Corresponding invasive coronary angiography demonstrating poor LAD distal blood flow (c) and three-dimensional CCTA reconstruction (d).

Figure 5.

Intramural hematoma and abrupt luminal stenosis ((a)–(d), arrows) on coronary computed tomography angiography (CCTA) of acute left anterior descending (LAD) coronary artery spontaneous coronary artery dissection. Corresponding three-dimensional CCTA reconstruction (d).

Figure 4.

Intramural hematoma and long tapered luminal stenosis ((a), arrows) with epicardial fat stranding ((b) and (c), arrows) on coronary computed tomography angiography (CCTA) of acute left anterior descending (LAD) coronary artery spontaneous coronary artery dissection (SCAD) with corresponding invasive coronary angiography ((d), arrow). Dissection ((e) and (f), ovals) due to SCAD of the left circumflex coronary artery on CCTA with corresponding invasive coronary angiography ((g), oval).

Additional features observed include epicardial fat stranding, an ill-defined hyperdensity adjacent to the coronary artery (Figure 4); beading, defined as >50% luminal change which increases and decreases in sequence; coronary tortuosity, as discussed elsewhere;¹⁹ coronary bridging and the location of dissection in relation to the bridging segment; presence of vessel occlusion and distal flow. Myocardial hypoperfusion, defined as hypodensity in a subendocardial distribution in locations of reduced perfusion, and regional wall motion abnormalities were also noted.

The most common CCTA finding was the abrupt luminal stenosis of the coronary lumen in 64% of patients (Figures 5 and 6). Coronary intramural hematoma was visualized in 50% of patients, and dissection was visualized in 14%. Most patients had a combination of features, particularly the patient with multi-vessel SCAD (Patient 14, Figure 4) who had differing features in each coronary. One patient (Patient 10) did not have any CCTA changes. Side-by-side review of her corresponding CA showed only mild, but suspicious, abnormalities of the mid left anterior descending artery with small focal aneurysm that was not well demonstrated on CCTA. That patient’s SCAD diagnosis was confirmed seven days later on CA when she presented with worsening chest pain and SCAD progression. There were no findings to suggest that she was at risk for SCAD progression on CCTA. Myocardial hypoperfusion was present in only half of patients despite all having ongoing ACS. None of the patients had coronary calcification.

Figure 6.

Frequency of primary and secondary features of acute spontaneous coronary artery dissection on coronary computed tomography angiography.

Discussion

Although the limited literature is generally aspirational regarding the use of CCTA in SCAD patients, there are no series to our knowledge describing acute SCAD features on CCTA. Additional strengths of this study include: (a) strict criteria of short time frame between CCTA and CA (in some cases hours) since SCAD can evolve rapidly; (b) ability to review original CCTA reports created before diagnosis was known; (c) access to all ancillary patient data. The present study proposes four primary features that should be assessed among SCAD patients undergoing CCTA and also defines secondary features.

We found that most common features are abrupt luminal stenosis of the coronary lumen and intramural hematoma. These are consistent with Type 2 SCAD, which is the most common CA finding.²⁰ A coronary intramural hematoma could occur in conjunction with either abrupt or tapered coronary luminal stenosis but was not always present. As discussed in a recent case report,²¹ the coronary intramural hematoma with a tapered luminal stenosis can distinguish SCAD from coronary vasospasm. However, we also observed that the intramural hematoma can mimic noncalcified atherosclerotic plaque on CCTA. Patients may appear to have “intramural hematoma” on CCTA which with intravascular imaging during cardiac catheterization is noncalcified atherosclerotic plaque (Figure 7). These observations support the need for invasive coronary angiography±intravascular imaging to distinguish early atherosclerotic coronary disease from SCAD, since proper diagnosis has critical ramifications for patient management including decisions about revascularization, medication decisions, physical activity recommendations, and reproductive counseling.

Figure 7.

Appearance of intramural hematoma in the left circumflex coronary artery ((a)–(c), arrows) suggestive of spontaneous coronary artery dissection (SCAD) with corresponding renal fibromuscular dysplasia (FMD; (d), oval) in a patient with history of non-ST-elevation myocardial infarction who did not undergo invasive coronary angiography as an initial strategy. Subsequent invasive coronary angiography with optical coherence tomography demonstrated that the patient had noncalcified atherosclerotic plaque, not SCAD, ((e), arrow; (f), asterisk) despite confirmation of FMD ((g), arrow).

The finding of a discrete dissection on CCTA, which correlates with Type 1 SCAD, was the most obvious finding but occurred in only 14%. A case report about a 34-year-old postpartum SCAD describes multi-vessel SCAD diagnosed acutely with CCTA managed conservatively with avoidance of invasive catheterization, and published images were consistent with findings of dissection.²² While that entirely non-invasive approach is of interest, based on the observations of this study we caution against relying on CCTA alone to diagnose SCAD, since only a minority of patients presented with such straight-forward images. CCTA could potentially delay diagnosis or lead to misdiagnosis and inappropriate care, particularly if the involved vessel is small or has some of the more common features that can appear similar to atherosclerosis. Further, Patient 4 had distal filling/reconstitution of her SCAD vessel, but she had coronary occlusion without sufficient collaterals on immediate, subsequent CA and underwent percutaneous intervention which led to notable improvement in distal coronary blood flow. The CCTA reconstitution finding may be related to collateral flow as the peripheral contrast has time to circulate. However, it may also give a false impression that the coronary is not occluded and/or receiving adequate collateral blood flow, affecting important management decisions for the patient.

The patient without SCAD features on CCTA emphasizes that a normal CCTA does not exclude the potential for SCAD. This important finding has been observed elsewhere including a case in which the CCTA was normal three days prior to the patient’s cardiac arrest and dissected right coronary artery.²³ These authors hypothesize that a small dissection may have been present but not detected due to CT spatial resolution limitations and the static properties of CT.²⁴ Eleid et al. found that acute SCAD was missed on three “triple rule out” studies (which permit simultaneous assessment of the pulmonary arteries, coronary arteries and aorta) prior to CA including a case in which SCAD was clearly present on retrospective review, emphasizing that the diagnosis of SCAD on CCTA can be interpreter dependent.²⁵

Roura et al. reported a series of 20 conservatively managed SCAD patients studied on CCTA in 3–6 months of follow-up and found healing of SCAD in 18 patients.¹⁵ While this study shows utility in follow-up CCTA in patients, it does not correlate those findings with concurrent CA. Further, the features of CCTA are not always obvious and resolution of CCTA is limited; one may speculate whether SCAD healing may be overestimated particularly in those with involvement of smaller, distal vessels.

Therefore, perhaps one of the most important observations of this study is that if a patient presents with acute coronary syndrome (ACS) which is suspicious for SCAD, CCTA may be informative regarding the underlying etiology, but CA remains an important and first-line diagnostic strategy. In young patients (especially women) with chest pain in whom pulmonary embolism is on the differential, a gated, “triple rule-out” protocol may be of more use than a non-ECG-gated study assessing only for pulmonary embolism. Even if the features are challenging to distinguish from atherosclerosis (e.g. hematoma), an ECG-gated, “triple rule-out” study may appropriately refine the differential diagnoses if SCAD is identified pre-emptively on CCTA.

We found that not all studies were of sufficient quality for adequate coronary assessment, usually related to increased heart rates and motion artifact. Therefore, if a “triple rule-out” approach is used, provisions such as adequate ECG-gating, cautious use of nitroglycerin to dilate the coronaries, and beta blockade to slow the heart rate are critical to study optimization. Although not apparent in all SCAD patients, myocardial hypoperfusion when present was correlated to the affected coronary distributions. Therefore, it is also of utility for cardiac imagers to integrate assessment of myocardial perfusion in their scan pattern.

This study is limited by its retrospective nature. We were intentionally not blinded to the patient diagnosis or history; therefore, we are unable to assess the sensitivity or specificity of detecting SCAD with CCTA compared to CA. Similarly, we did not compare CCTA of patients with SCAD to patients with atherosclerotic plaque and therefore cannot provide information on how reliable CCTA is for distinguishing SCAD from noncalcified atherosclerosis. Rather, the purpose of this study was to describe the CCTA features of SCAD in order to facilitate future studies and caution against sole use of CCTA for diagnosis of acute SCAD until more is understood. Particularly as CT technology advances, with ever increasing spatial and temporal resolution and decreasing radiation dose, there is great potential for CCTA in assessing acute and chronic SCAD patients. We need further studies to refine SCAD-specific CCTA terminology, assess the sensitivity and specificity of CCTA for SCAD, and better understand the role of CCTA in this patient population.

Conclusions

SCAD has four unique coronary features that can be classified on CCTA including, in order of prevalence: 1) abrupt luminal stenosis; 2) intramural hematoma; 3) tapered luminal stenosis; and 4) dissection. Diagnosing acute SCAD on CCTA is challenging and most vessels have tapered luminal stenosis and/or have intramural hematoma without the pathognomonic visible dissection in the vessel wall; therefore, invasive coronary angiography is the recommended first-line technique to assess a patient with ACS and possible SCAD. However, CCTA remains a useful technology particularly in those presenting early with chest pain and normal initial troponin level. Further work in this field is necessary, particularly as CCTA technology continues to advance with aims to improve resolution and reduce radiation.