Author + information
- Received January 22, 2016
- Revision received April 5, 2016
- Accepted April 14, 2016
- Published online September 21, 2016.
- Michael A. Bolen, MDa,b,∗ (, )
- Ellen Brinza, BSb,
- Rahul D. Renapurkar, MDa,
- Esther S.H. Kim, MDb and
- Heather L. Gornik, MDb
- aImaging Institute, Cleveland Clinic, Cleveland, Ohio
- bHeart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio
- ↵∗Reprint requests and correspondence:
Dr. Michael A. Bolen, Cardiovascular Imaging Laboratory, J1-4, Imaging Institute, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, Ohio 44195.
Objectives This study sought to evaluate the diagnostic yield of a dedicated computed tomography angiography (CTA) protocol of the chest, abdomen, and pelvis in patients with fibromuscular dysplasia (FMD).
Background FMD is an uncommon vascular disease that may result in stenosis, dissection, or aneurysm of nearly all arterial distributions, typically affecting medium-sized arteries. Findings from the United States Registry for Fibromuscular Dysplasia have suggested the potential need to perform screening imaging of the aorta and medium branch vessels.
Methods A total of 113 consecutive patients enrolled in our institutional FMD registry who received a tailored CTA protocol at our institution between March 2013 and June 2015 were included in this study. Arterial phase contrast-enhanced images were obtained on a dual-source scanner using high pitch and electrocardiogram trigger. Images were analyzed by 2 readers.
Results Abnormalities including beading, aneurysm, dissection, and stenosis/occlusion were noted in aortic, renal, mesenteric, and iliac distributions. The most commonly affected vessels were the renal arteries (n = 76 [67%]), followed by the lower extremity/iliac arteries (n = 37 [32%]). Aortic abnormalities were less frequently encountered (n = 3 [3%]), including 1 case with mild dilation (4.2 cm) of the ascending aorta and 2 cases of dissection involving the descending aorta, 1 with mild dilation (4.4 cm). Incremental findings beyond those known at patient intake were commonly noted, including new areas of arterial beading (n = 55 [49%]), new aneurysms (n = 21 [19%]), and new dissections (n = 3 [3%]). Reformatted images were crucial, affecting final assessment in 56% of cases evaluated by reader 1 and 36% evaluated by reader 2.
Conclusions Screening chest, abdomen, and pelvis CTA in patients with FMD showed substantial and incremental diagnostic yield. Reformatted images should routinely be included in imaging analysis. Abnormalities in the aorta were not common, so screening of the thoracic aorta may not be indicated.
Fibromuscular dysplasia (FMD) is a vascular disease that may result in stenosis, dissection, or aneurysm of nearly all arterial distributions, with the renal and extracranial carotid and vertebral arteries most commonly affected (1). Medium-sized arteries are typically affected, and the disease often involves multiple vascular territories within an individual patient. The etiology of this disease is unclear. It is believed to be neither an inflammatory nor an atherosclerotic process, but genetic and environmental risk factors may play a role in the condition (2). The clinical presentation is driven by the vascular beds affected. Renal involvement usually presents with hypertension, and carotid artery involvement is often heralded by pulsatile tinnitus, headache (generally the migraine type), and transient ischemic attack or stroke (often due to carotid or vertebral artery dissection). FMD may also be discovered incidentally in an asymptomatic patient who is imaged for another clinical indication.
Histologically, FMD is classified depending on which layer of the vessel wall is predominantly involved. Medial fibroplasia is the most common subtype, seen in >80% of patients (2). Angiographically, the appearance of medial fibroplasia resembles a string of beads, with beaded areas larger in caliber than the unaffected portion of the artery (3). Less common subtypes of FMD include intimal fibroplasia, which is a more focal process characterized by smooth or bandlike areas of narrowing, and perimedial and adventitial fibroplasia (4). A recent American Heart Association scientific statement proposed a simplified angiography-based FMD classification, with FMD lesions characterized as either multifocal (previously primarily medial fibroplasia) or focal (2).
To further the understanding of the clinical characteristics, epidemiology, management, and outcomes of patients with FMD, a multicenter registry was instituted in 2008 (1), and initial data from the first 447 patients entered into the United States Registry for Fibromuscular Dysplasia were reported in 2012 (1). Notable findings in this group included aneurysm in 17% of patients (with 20% of these cases involving the aorta) and dissection in 20%. Furthermore, because the entire aortic and visceral branch circulation was not imaged in all cases, the prevalence of aneurysms may have been underestimated. Based on these initial findings, in 2013 our institution began imaging all patients with a diagnosis of FMD to screen for aortic and/or visceral branch vessel and pelvic arterial abnormalities (5). A specialized computed tomography angiography (CTA) protocol is used for this one-time screening of patients who had not previously undergone comprehensive assessment of the aorta or visceral and pelvic branch vessels. For this study, we reviewed the first 113 patients imaged with this protocol to evaluate the imaging technique and analysis, categorize pathological findings, and assess for possible effects on clinical management.
This retrospective study was approved by the local institutional review board, with a waiver of individual consent. From March 2013 to June 2015, a total of 360 patients with an established diagnosis of FMD who were enrolled in our institutional FMD registry were seen for new patient or follow-up visits. Among these patients, a total of 113 consecutive patients (31%) who had not previously undergone tomographic imaging of the aorta or visceral and pelvic branch vessels underwent the specialized CTA protocol and were included in this study.
Imaging was performed on a commercially available second-generation dual-source computed tomography (CT) scanner (Siemens Somatom Flash, Siemens Healthcare, Erlangen, Germany). The full scan range extended from the thoracic inlet to the ischial tuberosities. High-pitch (pitch = 3.2) imaging was performed during the arterial phase. Patient electrocardiogram was used to trigger initiation of table motion as well as detector rotation, timed for table positioning at the level of the aortic root and proximal ascending aorta during the diastolic phase. The average estimated dose-length product was 273.1 ± 91 mGy·cm, corresponding to an effective dose of approximately 4 mSv. Contrast (111 ± 12 ml) (Omnipaque 350, GE Healthcare, Chicago, Illinois) was administered intravenously according to patient weight as determined by manufacturer-provided software (P3t Cardiac version 1.0, MEDRAD Inc., Pittsburgh, Pennsylvania), after a test bolus scan (20 ml) was obtained at the midabdominal level to assess time to peak enhancement. Techniques for radiation dose minimization included lower tube potential and tube current settings based on patient size (Table 1). In all cases, 18g intravenous access was obtained. Images were reconstructed to a slice thickness of 1 mm (increment = 1 mm) and 3 mm (increment = 3 mm), with medium smooth kernel (B31f).
All cases were analyzed on a dedicated workstation using manufacturer’s software (Agfa Healthcare, Mortsel Belgium) by a cardiovascular radiologist with fellowship training and 7 years’ CT reading experience (M.B.). Aortic size was assessed with the use of multiplanar reformatted images oriented orthogonal to the aorta centerline. Images were analyzed for the presence of beading, aneurysm, or dissection, first with axial thin-slice images, then with coronal and sagittal reformatted images in both thin slice (1 mm, 1-mm increment) and maximum intensity projection (8 mm, 1-mm increment). The investigators noted whether reconstructed images changed their initial interpretation and led to new findings (e.g., an aneurysm or new region of beading) or increased the reader’s confidence in the finding (i.e., changed assessment from possible to convincing beading). A second reader with cardiovascular imaging fellowship training and 4 years’ CT experience (R.R.) analyzed 25 randomly selected cases to assess inter-reader variability.
Aneurysm was diagnosed when the arterial diameter measured ≥150% the diameter of the adjacent normal artery. FMD (multifocal, string-of-beads type) was diagnosed when characteristic sequential areas of dilation greater in dimension than the adjacent normal vessel and separated by areas of narrowing were identified; these areas were graded as possible, convincing, or absent. Segments of focal concentric stenosis were also diagnosed as FMD (focal type). Dissection was diagnosed when the contrast-filled lumen was separated into true and false components by the intimal flap.
Clinical Data Collection
All patients included in this retrospective imaging study had been seen in the FMD clinic at our institution and had previously consented and been enrolled in the United States Registry for Fibromuscular Dysplasia (1). Information about past medical history, presenting signs and symptoms, and physical examination findings was collected at the time of patients’ enrollment into the registry and was extracted from the local registry database for this study. Regarding vascular bed involvement, all previous imaging studies performed before the time of the study CTA were included for analysis, including data captured at baseline and during previous follow-up visits at our center. Heart rate was obtained from the office visit closest to the date of the CTA, and all incremental findings from the imaging study were reported separately.
Baseline characteristics at the time of imaging are summarized in Table 2. Of the 113 patients included in the study, 112 (99%) were women, with a mean age of 55.6 ± 10.1 years (range 29 to 76 years) at the time of the imaging study. Common signs and symptoms included headache (77%), hypertension (54%), pulsatile tinnitus (55%), dizziness (47%), and neck pain (46%). Multifocal FMD was by far the most common type of FMD (96% of patients). A history of dissection was noted in 42% of patients, and 16% had a known aneurysm. Among patients who had previously undergone renal imaging, FMD was present in 64%. Carotid involvement was found in 92% of patients who underwent carotid artery imaging, and vertebral artery involvement was noted in 46% of patients. The makeup of this patient cohort may have been skewed by the imaging protocol. Performing screening thoracoabdominal aorta and branch vessel CTA in patients without prior tomographic imaging of this region may have led to a larger fraction of patients with cerebrovascular involvement by FMD. Conversely, patients with renal or other thoracoabdominal aortic branch vessel involvement at the time of diagnosis more likely would have undergone prior imaging of this area and therefore would not have undergone the screening imaging evaluated in this study. The relatively high frequency of upper extremity, lower extremity, and coronary artery involvement on imaging studies likely reflects the small number of patients who underwent dedicated imaging studies of these vascular territories, generally for investigation of signs and symptoms, increasing the likelihood of vascular bed involvement.
Frequency and Distribution of Vascular Abnormalities
Abnormalities including beading, aneurysm, dissection, and stenosis/occlusion were noted in aortic, renal, mesenteric, iliac, and coronary distributions (Table 3). The most commonly affected vessels were the renal arteries (n = 76 [67%]), followed by the lower extremity/iliac arteries (n = 37 [32%]). Aneurysms of the aortic branch vessels ranged in size from 5 to 14 mm. Aortic abnormalities were less frequently encountered (n = 3 [3%]), including 1 case with mild dilation (4.2 cm) of the ascending aorta and 2 cases of dissection involving the descending aorta, with mild dilation (4.4 cm) in 1 of these cases. Aortic valve morphology suggested a trileaflet valve in all but 1 case, which appeared bicuspid (confirmed at the time of transthoracic echocardiography). Examples of the various vascular abnormalities are shown in Figures 1 to 4⇓⇓⇓⇓.
Incremental findings yielded by CTA imaging were commonly noted, including beading (multifocal FMD) (n = 55 [49%]) or focal stenosis (focal FMD) (n = 1 [1%]) in a new vascular territory, and new aneurysms (n = 21 [19%]) or new dissections (n = 3 [3%]).
Effect of Reconstructions
The use of reformatted images affected final assessment in 63 of the 113 cases (56%) evaluated by reader 1 and 9 of the 25 cases (36%) evaluated by reader 2.
The inter-reader agreement (Table 4) was reasonably good, with the readers showing exact agreement in grading of beading (not present, possible, or convincing) in 80% to 88% of cases. When the “possible” and “convincing” groups were classified as positive, the readers showed 84% to 96% agreement in diagnosis. All cases of dissection showed concordant analysis, and only 3 subcentimeter aneurysms were discrepant, with the remaining 22 aneurysms concordant (88% agreement). The renal arteries showed the lowest exact agreement (80%) between readers, although the agreement was excellent (96%) when the scores of “possible” and “convincing” were combined.
A high prevalence of aneurysms has been reported in patients with FMD (1). At our institution, we use CTA to assess a wide expanse of vascular territory in patients with documented FMD, and this study is the first to our knowledge to report the outcomes of this screening approach. A significant number of abnormal vascular findings were made, many of these not known at the time of patient intake, although a surprisingly low number of aortic abnormalities were identified.
Imaging Techniques to Screen for Vascular Abnormalities in Patients With FMD
Invasive techniques such as digital subtraction angiography (DSA) would likely improve the detection of subtle vascular abnormalities in this patient population. A review of DSA in potential renal donors demonstrated an FMD incidence of 3.8% (6) versus 2.6% with CTA (7), suggesting that DSA may be an attractive option when there is high suspicion that intervention will be required. However, the risk, time, and cost associated with this approach make it a less appealing option for the assessment of large vascular territories. Some investigators have reported robust performance with CTA versus DSA in the assessment of FMD (8) and renal artery stenosis, even with slightly older-generation CT scanners (9,10). Alternatively, ultrasonography with color flow Doppler can be used to identify areas of turbulence and some cases of abnormal arterial morphology. However, Doppler ultrasonography is strongly dependent on user skill and satisfactory sonographic windows, which can be challenging in the assessment of renal and visceral branch arteries (2).
Magnetic resonance angiography (MRA) is another possible technique for the detection of vascular abnormalities across multiple vascular territories. This modality offers the additional benefit of increased functional (e.g., flow) measurements, all without the need for ionizing radiation. However, MRA is more time intensive than CTA, and its spatial resolution is slightly inferior to that of CTA, which might limit its ability to detect abnormalities. In studies of patients with spontaneous coronary artery dissection, MRA detected a 25% prevalence of extracoronary abnormalities (11), whereas CTA detected a 69% prevalence (12). Additionally, magnetic resonance imaging is not effective for evaluating calcium deposition and may lead to difficulties in differentiating between atherosclerosis and FMD or in evaluating mixed processes (e.g., atherosclerosis plus FMD).
Frequency of Abnormalities
Because this was the first study to evaluate the diagnostic yield of screening aortic and branch vessel angiography in patients with FMD diagnosis, no direct comparison can be drawn with previous investigations. In a study of patients with spontaneous coronary artery dissection, Liang et al. (12) used spiral CTA of the neck, chest, abdomen, and pelvis to assess for incidental vascular abnormalities. The investigators identified abnormalities in the aorta and mesenteric arteries (10% each), renal arteries (26%), and lower extremity arteries (36%). The slightly differing (typically lower) frequencies of vascular territory involvement found in this investigation may reflect differences in the patient groups, as FMD is only one of several conditions that predispose patients to spontaneous coronary artery dissection (13). In contrast to this investigation, our protocol did not fully image the arch branch vessels and circle of Willis because most of the patients referred had already undergone imaging of this region, and the primary indication was evaluation of aortic and visceral and pelvic arterial distribution.
Compared with the initial report of the entire United States Registry for Fibromuscular Dysplasia, the prevalence of aortic findings in our cohort was low. For example, whereas 20% of aneurysms noted in the registry were aortic, only 2 cases of mild thoracic aortic dilation were observed in our study cohort (4.2 to 4.4 cm), 1 in the setting of descending aortic dissection. Renal artery involvement in our investigation was similar to that of the overall United States Registry for Fibromuscular Dysplasia but slightly less common (67% in the current study vs 80% in the entire registry). Mesenteric involvement also was similar in the 2 investigations (22% in this study vs. 26% in the entire registry). As discussed earlier, the potential skewing of this patient cohort based on scanning patients without prior thoracoabdominal tomographic imaging may have influenced these results.
Effect of Reconstruction
The effect of reconstruction on final image interpretation may be attributable to the relatively small size (approximately 3- to 15-mm outer diameter) of the aortic branch vessels evaluated, as well as the course of these vessels, as they often pass outside of the axial plane. Use of multiplanar reformatted and maximum intensity projection imaging can increase the length of the vessel assessable in a single imaging plane and increase the conspicuity of findings such as smaller aneurysms or subtle beading. Careful assessment of vascular morphology is warranted in the evaluation of FMD because other processes such as atherosclerosis and arteritis have a potentially overlapping imaging appearance. Although the typical string-of-beads appearance is very specific to FMD, challenges may occur in the setting of subtle imaging findings, as well as in cases with concomitant atherosclerosis and FMD, as can often be seen in older patients.
This study is limited by several factors. The investigation reported on a cohort of patients seen within a specialized referral clinic and therefore may not be generalizable to the entire population of patients with FMD. The imaging protocol described within this report assessed the aorta and visceral and pelvic branches and did not include imaging of the extracranial (e.g., carotid/vertebral arteries) or intracranial (e.g., cerebral, basilar arteries) cerebrovasculature. The makeup of this patient cohort may have been skewed by the imaging protocol, as mentioned earlier. Complete clinical data for all clinical parameters were not available for all patients included in this series, and all vascular territories were not assessed in all patients. Additionally, not all patients with FMD seen at our institution were included in this study because the dedicated screening CTA protocol was generally not ordered for patients who had previously undergone aortic and visceral and pelvic branch imaging. We acknowledge that, although coronary findings occasionally may be identified using our imaging protocol, which includes high pitch and electrocardiogram triggering, the lack of chronotropic and vasodilatory medications results in nonoptimized CT assessment. This precluded precise assessment of coronary abnormalities and formalized analysis in this investigation. The CT scanner technology used in this study, although commercially available, is not present in all clinical settings. Additionally, patients scanned elsewhere within our institution using different CT scanners and protocols could not be included in this analysis. Finally, although efforts were made to characterize arterial lesions as aneurysms or dissections, it is possible that some small aneurysms actually represented pseudoaneurysms that formed during arterial remodeling after an arterial dissection.
Use of a specialized screening CTA imaging protocol of the aorta and medium branch vessels in patients with FMD yielded useful incremental findings. The overall frequency of abnormalities observed was within the expected range for visceral branch and lower extremity arteries. Infrequent abnormalities in the thoracic aorta suggest that screening imaging in this patient population may be more appropriately limited to abdomen and pelvic aorta and branch vessel CTA. Confidence in this finding may be increased by correlating these results with findings from other centers. This screening CTA protocol was performed with a reasonable estimated radiation dose and contrast volume. Reformatted image processing had a notable effect on image analysis.
COMPETENCY IN MEDICAL KNOWLEDGE: FMD is a diffuse arteriopathy that involves medium-sized arteries in multiple vascular territories throughout the body, most commonly the renal and cerebrovascular arteries. A high prevalence of arterial aneurysm and dissection has been previously demonstrated among patients with FMD. Use of a specialized screening CTA protocol of the chest, abdomen, and pelvis in a single-center group of patients with FMD diagnosis provided important incremental information about aortic branch vessels; however, aortic abnormalities were uncommon in this group (3%). Reformatted images had a significant effect on final interpretations.
TRANSLATIONAL OUTLOOK: Additional studies are needed to validate the findings and determine whether screening CTA can be limited to the abdomen and pelvis in patients with a diagnosis of FMD.
The authors appreciate the editorial contribution of Megan M. Griffiths and the CT scanning expertise of Stacie Kuzmiak.
Dr. Kim is a paid consultant for Philips Ultrasound. Dr. Gornik is a member of the medical advisory board of the FMD Society of America, a nonprofit organization. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- computed tomography
- computed tomography angiography
- digital subtraction angiography
- fibromuscular dysplasia
- magnetic resonance angiography
- Received January 22, 2016.
- Revision received April 5, 2016.
- Accepted April 14, 2016.
- American College of Cardiology Foundation
- Olin J.W.,
- Froehlich J.,
- Gu X.,
- et al.
- Olin J.W.,
- Gornik H.L.,
- Bacharach J.M.,
- et al.,
- American Heart Association Council on Peripheral Vascular Disease,
- American Heart Association Council on Clinical Cardiology,
- American Heart Association Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation,
- American Heart Association Council on Cardiovascular Disease in the Young,
- American Heart Association Council on Cardiovascular Radiology and Intervention,
- American Heart Association Council on Epidemiology and Prevention,
- American Heart Association Council on Functional Genomics and Translational Biology,
- American Heart Association Council for High Blood Pressure Research,
- American Heart Association Council on the Kidney in Cardiovascular Disease,
- American Heart Association Stroke Council
- Harrison E.G. Jr..,
- McCormack L.J.
- O’Connor S.C.,
- Gornik H.L.