Author + information
- Received January 18, 2018
- Revision received March 4, 2018
- Accepted March 21, 2018
- Published online June 11, 2018.
- Anne S. Chin, MDa,
- Martin J. Willemink, MD, PhDa,
- Aya Kino, MDa,
- Virginia Hinostroza, BSa,
- Anna M. Sailer, MD, PhDa,
- Michael P. Fischbein, MD, PhDb,
- R. Scott Mitchell, MDb,
- Gerald J. Berry, MDc,d,
- D. Craig Miller, MDb and
- Dominik Fleischmann, MDa,d,∗ (, )@StanfordMed@SUMedicine@StanfordCV
- aDepartment of Radiology, Stanford University School of Medicine, Stanford, California
- bDepartment of Cardiothoracic Surgery, Stanford University School of Medicine, Stanford, California
- cDepartment of Pathology, Stanford University School of Medicine, Stanford, California
- dStanford Cardiovascular Institute, Stanford University School of Medicine, Stanford, California
- ↵∗Address for correspondence:
Dr. Dominik Fleischmann, Department of Radiology, Stanford University School of Medicine, 300 Pasteur Drive, S-072, Stanford, California 94305-5105.
Background Limited intimal tears (LITs) of the aorta (Class 3 dissection variant) are the least common form of aortic pathology in patients presenting with acute aortic syndrome (AAS). LITs are difficult to detect on imaging and may be underappreciated.
Objectives This study sought to describe the frequency, pathology, treatment, and outcome of LITs compared with other AAS, and to demonstrate that LITs can be detected pre-operatively by contemporary imaging.
Methods The authors retrospectively reviewed 497 patients admitted for 513 AAS events at a single academic aortic center between 2003 and 2012. AAS were classified into classic dissection (AD), intramural hematoma, LIT, penetrating atherosclerotic ulcer, and rupturing thoracic aortic aneurysm. The prevalence, pertinent risk factors, and detailed imaging findings with surgical and pathological correlation of LITs are described. Management, early outcomes, and late mortality are reported.
Results Among 497 patients with AAS, the authors identified 24 LITs (4.8% of AAS) in 16 men and 8 women (17 type A, 7 type B). Patients with LITs were older than those with AD, and type A LITs had similarly dilated ascending aortas as type A AD. Three patients presented with rupture. Eleven patients underwent urgent surgical aortic replacement, and 2 patients underwent endovascular repair. Medial degeneration was present in all surgical specimens. In-hospital mortality was 4% (1 of 24), and in total, 5 patients with LIT died subsequently at 1.5 years (interquartile range [IQR]: 0.3 to 2.5 years). Computed tomography imaging detected all but 1 LIT, best visualized on volume-rendered images.
Conclusions LITs are rare acute aortic lesions within the dissection spectrum, with similar presentation, complications, and outcomes compared with AD and intramural hematoma. Awareness of this lesion allows pre-operative diagnosis using high-quality computed tomography angiography.
Among the spectrum of lesions presenting with acute aortic syndrome (AAS), so-called “limited intimal tears” (LITs), “limited dissections,” or “intimal tears without hematoma” all refer to one of the most elusive of acute aortic pathologies. Murray and Edwards (1) observed as early as 1973 in an autopsy series that spontaneous lacerations of the aortic intima and subjacent media could result in varying degrees of dissection of blood within the media. In the majority of cases, extensive intramedial dissection of blood resulted in the double-barrel lumen of classic dissection; in other cases, however, there was no significant dissection of blood into the media, which Murray termed “incomplete dissection.” Following the intimal-medial tear, the lacerated edges of an incomplete dissection retract and the residual “bare area,” which usually contains some medial tissue, showed localized aneurysmal dilatation or bulging (1). In the years following its original description, only a handful of case reports—typically discovered post mortem—have been published in the English language published reports using various terms including “spontaneous,” “nontraumatic,” or “incomplete” rupture or tearing of the aorta (2–8).
In 1999, Svensson et al. (9) rediscovered this neglected lesion, identifying 9 patients with “limited intimal tears” in 109 patients undergoing ascending aortic repair. Of note, all 9 LITs were missed on pre-operative imaging, despite the use of 3 or more imaging modalities and a high clinical index of suspicion. Svensson et al. (9) proposed “limited intimal tears” as a “Class 3 dissection variant” in a new classification of aortic dissections, which has since been widely adopted: In 2001, the Task Force on Aortic Dissection of the European Society of Cardiology, classified this lesion as a “Class 3 intimal tear (subtle or discrete dissection)” (10), and in 2010 the multispecialty American College of Cardiology Foundation/American Heart Association Task Force document classified this lesion as a “Class 3 dissection variant (limited dissection)” (11). Despite this international, multispecialty endorsement, only a few practitioners appear to be familiar with this entity (12,13), and numerous case reports corroborate that these lesions remain notoriously difficult to detect on imaging (14–22).
The purpose of this retrospective examination of 497 patients with AAS presenting to a single academic aortic center over a 10-year period was to: 1) evaluate the frequency, clinical characteristics, and risk factors of patients with acute LITs of the aorta; 2) report the spectrum of computed tomography (CT) findings with surgical and pathological correlation; and 3) describe the management and outcomes of patients with LITs, compared with other causes of AAS. To our knowledge, this is the largest series of LITs to date, and the first to report on the diagnostic capabilities of CT angiography (CTA) to diagnose this obscure lesion.
This HIPAA (Health Insurance Portability and Accountability Act of 1996)-compliant, retrospective, observational study was approved by the institutional review board. The requirement for informed consent was waived. In an effort to identify all patients with the possible diagnosis of an LIT, we reviewed the medical records and imaging data of all patients admitted or transferred to our center with a final diagnosis of AAS between January 2003 and December 2012. Individuals were identified by a combination of electronic searches by diagnostic and discharge codes in the electronic medical records (EMR), text searches within the radiology information system, and expert review of all candidate imaging data in the picture archiving and communication system to confirm the presence of positive imaging findings or verify the absence of an acute aortic lesion in reported negative studies. Specifically, subtle aortic contour abnormalities were sought to identify possible LITs that might have been initially missed.
AAS was defined as an acute aortic lesion presenting within 14 days of symptom onset. We categorized AAS based on the 5 classes of the American College of Cardiology Foundation/American Heart Association classification of aortic dissection, expanded by a sixth category for patients with rupturing thoracic aortic aneurysms: Class 1 = classic aortic dissection (AD); Class 2 = intramural hematoma (IMH); Class 3 = limited intimal tear (LIT); Class 4 = penetrating atherosclerotic ulcer (PAU); Class 5 = iatrogenic dissection; and Class 6 = rupturing thoracic aortic aneurysm (RTA), defined as thoracic diameter >4 cm in patients presenting with aortic pain and imaging findings suggestive of transmural leakage, intraclot hemorrhage, periaortic hematoma, or active contrast medium extravasation. Diagnosis of AAS was based on presentation and initial imaging, intraoperative inspection, pathology review, follow-up imaging, and clinical outcome.
CT image evaluation
All CT scans were reviewed in consensus by 2 cardiovascular radiologists (A.S.C. and D.F.) with 8 and 17 years of experience, respectively, for the presence, extent, and distribution of aortic lesions as well as pertinent extravascular findings. Radiological classification of aortic lesions was based on transverse CT images in cases with classic findings. Multiplanar reformations were routinely obtained and reviewed. Interactive post-processing techniques including 3-dimensional (3D) volume rendering with endoluminal views were utilized in all but classic dissections to search for subtle aortic contour abnormalities, and to measure lesion diameters (AquariusNet Viewer, TeraRecon, Foster City, California). The criteria for LIT classification consisted of a linear or stellate luminal contour abnormality without a dissection flap separating a true from a perfused false lumen. Associated findings could include focal linear filling defects (corresponding to undermined edges of the tear), a small amount of local intramural hematoma under the lesion edges, and an aortic bulge representing the outpouching of the residual aortic wall at the base of the intimomedial defect. Recorded extravascular findings included pleural effusion, pericardial effusion, mediastinal hematoma, hemothorax, hemopericardium, periaortic stranding or hematoma, pulmonary artery subadventitial hematoma, and esophageal abnormality. CT evidence of acute complications, such as frank rupture with active contrast medium extravasation and branch vessel malperfusion was also noted.
Clinical management and early outcomes
The protocol for patients with AAS presenting or transferred to our institution included admission to the cardiovascular surgery intensive care unit for blood pressure and pain control. Patients with lesions in the ascending aorta (Stanford type A) established by in-house CTA, or confirmed by pre-operative transesophageal echocardiography (TEE), were considered for urgent surgical repair. Patients without involvement of the ascending aorta (Stanford type B) were treated medically, unless complications such as rupture, organ or limb malperfusion, aortic diameter expansion during admission, uncontrollable hypertension, or intractable pain required consideration for open surgical or endovascular intervention.
EMRs were reviewed for demographic data and risk factors, such as hypertension, hypercholesterolemia, atherosclerosis (defined as a diagnosis of coronary artery disease, cerebrovascular disease, and peripheral vascular disease), smoking (defined as current or former smoker), and presence of connective tissue disease (Marfan syndrome or other connective tissue diseases). Clinical presentation, course of disease, interventions performed, and survival during the hospital course and up to 30 days were recorded. Surgical and pathology reports were reviewed in detail for intraoperative and pathological findings. Surgical pathology specimens were retrieved from the archives, and histological review was performed by an experienced cardiovascular pathologist (G.J.B.). The aortic segments were classified according to the recent nomenclature proposed by the Society for Cardiovascular Pathology and Association for European Cardiovascular Pathology (23). Gross pathology was correlated with intraoperative photographs and CTA findings where available.
Follow-up and long-term survival
Inpatient and outpatient EMRs were reviewed after hospital discharge for long-term survival (up to 14 years). Mortality data were updated until August 2017 using EMR data and our institutional Translational Research Integrated Database (STRIDE), which contains Social Security Death Index data until 2011.
Analysis and statistics
Anatomic distribution of AAS (type A vs. type B), age, sex, hypertension, hypercholesterolemia, atherosclerosis, smoking, connective tissue diseases, and prior AAS were compared between LIT and other AAS (AD, IMH, PAU, and RTA) using the Fisher-Freeman-Halton exact test. Age and maximum aortic diameter were compared between all AAS groups using the Kruskal-Wallis test. If variables differed significantly between AAS types, post hoc Mann-Whitney U tests were conducted between LIT and other types of AAS (AD, IMH, PAU, and RTA). p Values for post hoc analyses were corrected for multiple testing using the Bonferroni method. Kaplan-Meier analysis was used to evaluate 30-day and 5-year survival for the different types of AAS. Survival curves were compared with the log-rank test. Iatrogenic dissections were excluded from statistical analyses because of the different pathogenesis and low occurrence. A p value ≤0.05 was considered statistically significant for initial testing, and a Bonferroni-corrected p value ≤0.0033 was considered statistically significant for post hoc analyses.
Descriptive data were presented as frequencies and percentages for categorical variables and as medians (interquartile range [IQR]) for continuous variables. Statistical analyses were performed using MedCalc version 17.9.7 (MedCalc, Ostend, Belgium) and SPSS Statistics version 24 (IBM Corp., Armonk, New York).
Frequency of LITs among AAS
Over 10 years, 513 AAS events occurred in 497 patients (66% men, 34% women), with a median age of 61 years (range: 16 to 91 years). Sixteen patients experienced 2 separate events and admissions during the study period, within a median interval of 150 days (range 24 to 1,880 days), none of which was an LIT. Due to dependency of data, the second events were not included in statistical testing. One patient with an acute LIT in the aortic arch and with simultaneous descending thoracic AD was categorized as LIT for analysis. Another 37 of 497 patients (7.4%) had a history or imaging evidence of prior aortic dissection—one of which was consistent with a chronic type A LIT. Forty-three of 497 patients (8.6%) had a history of ascending aorta, with or without aortic root and/or hemiarch replacement (n = 36), or thoracic aortic endovascular repair (n = 7). One-half of the patients were white (50.7%), the remainder of the cohort identified as Asian (14.5%), Hispanic (9.3%), black (6.0%), Pacific Islander (1.2%), or other (1.8%); in 16.5% of patients, ethnicity was unknown. Forty-five patients (9.1%) had a diagnosis of connective tissue disease, including 29 of 497 (5.8%) patients with the Marfan syndrome.
LITs occurred in 24 patients, representing 4.8% of 497 patients. The most common AAS categories were classic AD (56.1%, 279 of 497) and IMH (24.3%, 121 of 497), together accounting for 80.5% (400 of 497) of AAS. The remaining AAS included 36 (7.2%) RTA and 29 (5.8%) PAU events. Only 8 (1.6%; 8 of 497) iatrogenic dissections were recorded, 2 post-surgical and 6 complications of endovascular procedures (Table 1).
Imaging for AAS
CTA was the initial imaging modality performed in 97% (503 of 513) of all AAS events. A total of 124 patients were transferred directly to the intensive care unit and operating room without onsite CT imaging, and diagnosis was made at operation. Pre-operative outside imaging included 8 transthoracic or transesophageal echocardiograms, 1 conventional angiography, and 115 CTAs, 63 of which were available for review. There were 388 CTAs acquired onsite using 8-, 16-, or 64-slice multidetector CT technology. One patient died during work-up before any imaging was obtained. All onsite CTA datasets were reconstructed with section thickness of 1.25 mm or smaller; one-half of CTAs (48%; 188 of 388) were acquired with electrocardiogram (ECG) gating.
Demographics and risk factors in patients with LITs
Two-thirds of patients (16 of 24) with LIT were men, similar to AD and all other AAS (p = 0.093). The youngest and oldest LIT patients were 35 and 91 years of age, respectively. At a median age of 71 years (IQR: 55 to 79 years), LIT patients were significantly older than patients with classic AD at a median of 54 years (IQR: 44 to 66 years; p = 0.001), but not significantly older than patients with IMH (median age 66 years; IQR: 54 to 76 years; p = 0.432). The percentage of elderly (>70 years) individuals among LIT patients was also greater (50.0%; 12 of 24) than in patients with AD (17.2%; 48 of 279; p = 0.001) (Table 2).
The most common associated condition in patients with LIT was arterial hypertension (91.7%; 22 of 24), which was slightly less common in classic AD (86.7%; 242 of 279) and slightly more common in IMH (95.0%; 115 of 121). Differences in prevalence of hypertension were significant (p = 0.045). Other risk factors, such as atherosclerosis (58.3%; 14 of 24), current or previous smoker (50.0%; 12 of 24), and hypercholesterolemia (45.8%; 11 of 24) also differed significantly between AAS types (p ≤ 0.016). The 35-year-old patient was diagnosed with a nonclassified inheritable connective tissue disease (4.2%; 1 of 24) (Table 2).
Presentation of LITs
Three LIT patients presented with hypotension (systolic blood pressure <100 mm Hg), the remaining patients were hemodynamically stable. Twelve patients were hypertensive (systolic blood pressure >140 mm Hg) at presentation, 6 of whom with systolic blood pressure >180 mm Hg. Three patients had imaging signs of rupture (2 type A, 1 type B). One type B LIT with continued pain progressed on imaging during the admission. One patient with a type B LIT in the arch and concomitant descending AD had mesenteric and lower extremity ischemia.
The majority of LITs affected the ascending aorta (70.8% type A; 17 of 24), similar to classic AD (59.9% type A; 167 of 279). Seven patients had type B LITs (29.2%). Of type B LITs, 3 involved the transverse aorta and 4 involved the descending thoracic aorta. Other AAS groups (IMH, PAU, and RTA) were predominantly type B.
The affected aortic segment was aneurysmal (>40 mm) in 20 of 24 LIT cases (83.3%) and in 16 of 17 type A LITs (94.1%). This is again similar to classic AD with a large proportion of ascending aneurysms (type A: 103 of 118, 87.3%). Median diameters of the affected ascending aorta were also similar between type A LIT (50.0 [IQR: 46.5 to 58.0] mm) and type A AD (50.0 [IQR: 44.8 to 60.0] mm). Aortic diameters of both LIT and AD were slightly larger than IMH, PAU, and iatrogenic dissection, but smaller than RTA (p < 0.001) (Table 2).
CT imaging features of LIT
Anatomically, an LIT is a partial thickness linear or stellate tear of the luminal aortic wall with retracted edges that may be undermined to a variable degree, and an outward bulge of the outer aortic contour. Historically, we have referred to these lesions as “mushroom cap lesions” on the basis of their angiographic appearance, or “stretch marks.” The CT typical appearance consists of an oval or band-like luminal contour defect, with subtle step-offs representing the retracted edges. When undermined, the edges are easier to identify on CT as linear filling defects (“focal flaps”) with or without associated subtle wall hematoma. The base of the defect typically bulges outward (Figure 1).
The 24 LITs were oriented longitudinally (parallel to the long axis of the aorta, n = 10) or circumferentially (perpendicular to the long axis of the aorta, n = 14). The lesions were between 17 mm and 77 mm in length, on average 41 to 43 mm for either orientation, and for both, type A and type B LITs. The widths of the tears in vivo under systemic pressure were between 9 and 39 mm. The edges of the lesions were undermined to a variable degree, up to 20 mm, and either filled with contrast-enhanced blood or fresh clot. The aortic bulge on the outer aortic contour thus could be larger than the luminal dimensions of the tear (Figure 1).
CT imaging—LIT-associated findings
The most common associated CT finding was pericardial effusion in 9 of 24 (38%), followed by pleural effusion in 7 of 24 (29%), and hemopericardium in 6 of 24 (25%). Nine patients (38%) had an abnormal mediastinal finding (mediastinal hematoma, periaortic fluid, or pulmonary artery periadventitial hematoma). One-half of the patients (12 of 24) had a small amount of intramural hematoma present at the edges of the tear. Importantly, a quarter (25%; 6 of 24) of LITs had no other associated CTA findings at presentation besides the LIT itself (Figure 1).
Missed and “Misclassified” limited intimal tears on CT imaging
Four cases of LIT were missed prospectively: 2 type A and 2 type B lesions. One type A LIT consisted of a subtle contour abnormality of the ascending aorta. This lesion progressed slightly on follow-up imaging, becoming more conspicuous; however, it was visible in retrospect on the initial exam. No associated CT findings were present in this individual. There were no hemodynamic complications, and the patient had an uneventful hospital discharge. A second sizable type A LIT was not recognized at presentation at an outside hospital in a patient with acute chest pain, a mildly aneurysmal ascending aorta, and a small pericardial effusion; however, the lesion was identified upon our subsequent review of the initial CTA. An ECG-gated CTA was performed upon transfer, which clearly demonstrated the extent of the tear (Figure 2). This patient also experienced no hemodynamic compromise and underwent successful valve-sparing aortic root and ascending and proximal arch replacement.
Two type B LITs were missed initially: 1 lesion involved the aortic arch only, and the other involved the proximal descending thoracic aorta (Figure 3). Both patients were transferred to our institution for ongoing aortic pain, and the diagnosis of LIT was made upon transfer. Both patients were managed medically and survived an uneventful hospital course.
Six cases of LIT were misclassified, although the presence of an acute aortic lesion was recognized and treated accordingly. Two cases were initially classified as IMH, 2 cases were classified as PAU, and 2 cases were classified as RTA. In the cases misclassified as IMH, very little IMH was actually present, and the bulging contour abnormality was initially not appreciated. In the cases of PAU misclassification, there was no significant atherosclerosis of the aorta, making this diagnosis unlikely. In the cases of RTA misclassification, the aorta was indeed aneurysmal, and signs of rupture were present; however, the linear tear was not appreciated on initial review but clearly displayed using 3D volume rendering.
Management and early outcomes
Eleven of 17 type A LITs were treated surgically, including 2 cases with suspected rupture on CTA, which was confirmed at operation. There were no in-hospital deaths in surgically treated patients, and all 14 patients with follow-up survived to 30 days. Six of the 17 patients with type A LIT were managed medically. One patient with multiple comorbidities refused surgery and died in-hospital. The other 5 medically managed patients survived to discharge: 2 elderly patients (ages 81 and 82 years) refused surgery, 2 patients were considered nonsurgical candidates (ages 89 and 91 years), and in 1 patient, the lesion was initially missed, but remained stable once detected at follow-up imaging. All 4 of the 5 medically treated patients with type A LITs discharged alive with available follow-up survived to 30 days. In total 3 of 17 patients with type A LIT were lost to follow-up after discharge within 30 days.
Two type B LITs underwent successful endovascular aortic repair, including 1 patient with CTA signs of rupture and 1 patient with persistent aortic pain and lesion progression on follow-up CTA. The patient with an arch LIT and concomitant type B dissection underwent successful branch vessel stenting, fenestration, and surgical femoral cross-over bypass for mesenteric and lower extremity ischemia (classified as endovascular treatment for analysis); the remaining 4 of 7 patients with uncomplicated type B LIT were managed medically. All type B LITs survived to hospital discharge, and all 4 patients with follow-up survived to 30 days. Three of 7 type B LITs were lost to follow-up after discharge within 30 days). Although the total number of events in LIT patients is too small for a formal analysis, in-hospital mortality (4.2%; 1 of 24) fell within the same order of magnitude as classic AD (5.4%; 15 of 279) and IMH (4.1%; 5 of 121). As expected, RTA was associated with the highest in-hospital mortality (25%; 9 of 36).
Pathological specimens from 10 of 11 patients who underwent surgical repair of the ascending aorta were available for histological review. All 10 specimens showed an acute intimal-medial tear, and a focal/limited dissection plane in the medial layer, similar to classic acute aortic dissection. All 10 specimens showed medial degeneration with 8 of 10 cases graded moderate or severe. One-half of the specimens additionally demonstrated lamellar medial collapse and band-like smooth muscle cell loss. Seven of 10 specimens showed no associated atherosclerosis, whereas 3 of 10 showed mild or moderate atherosclerosis. Gross specimen photographs were available in 3 of the patients, showing excellent correlation with preoperative CTA findings (Figures 1 and 2).
The majority of patients treated for AAS at our center are referred from other institutions, and survivors of the acute phase are often transferred back to the referring facilities for continued care and follow-up. Long-term follow-up data are thus sparse.
None of the 23 LIT patients who survived the initial hospitalization was readmitted at our institution. Four patients died after 94 days (3.1 months), 542 days (1.5 years), 919 days (2.5 years), and 1,953 days (5.4 years) of unknown cause. Six patients were lost to follow-up after discharge. The remaining 13 patients were alive at a median of 1,298 (IQR: 532 to 2,156) days (3.6 [IQR: 1.5 to 5.9] years) after the event, with the longest follow-up being 3,956 days (10.8 years).
Kaplan Meier analysis showed that the 30-day and 5-year survival estimates for LIT patients falls within a similar range as other AAS patients, except for RTAs who have the worst prognosis of all patients presenting with AAS (p < 0.001 for both 30-day and 5-year survival) (Central Illustration).
First and foremost, this study confirms that LITs are a rare, but unquestionable, cause of AAS—representing approximately 5% of AAS in our series. This is only slightly less common than PAUs and RTAs. Coincidentally, this finding implies that LITs—despite being recognized as “Class 3 dissection variant” in the published reports (10,11)—are grossly underdiagnosed. The glaring absence of LITs in all but the Svensson et al. (9) surgical series, as well as in the largest AD registry, IRAD (International Registry of Aortic Dissection), which includes >5,000 patients (24), suggests that patients with an LIT are either misclassified or completely missed, with potentially serious consequences (25). Our best explanation for this paradox is that, even today, most physicians remain unfamiliar with this lesion (26), and that LITs were deemed undetectable by imaging (9). The second and more encouraging main finding of our study is that LITs can in fact be detected with contemporary CT. By familiarizing those involved in the diagnosis and care of patients with aortic diseases, with the epidemiology, pathology, imaging appearance, and outcomes of LITs, we hope to break this vicious cycle and enable the timely diagnosis, accurate classification, and treatment of patients affected by this overlooked lesion in the future.
The frequency of LITs in our series was 4.8% of all patients (24 of 497) presenting with AAS to an academic aortic center over 10 years. This includes not only “traditional” causes of AAS (Classes 1 to 5) in the denominator but also patients with RTAs because these are clinically indistinguishable from other AAS at presentation, and might be confused or overlap morphologically with LITs. Our frequency is similar to the Svensson et al. (9) 5.0% (9 of 181) in the largest surgical series published in 1999, although 3 of the Svensson et al. (9) 9 LITs were subacute/chronic. Considering only type A Class 1 to 3 lesions, as in the Svensson et al. (9) surgical series, our percentage is slightly higher (7.3%; 17 of 232).
The frequency of LITs in the only other published series, by Chirillo et al. (12) in 2007, was also 7.3% (8 of 109 patients with acute type A AD diagnosed prospectively using TEE), although this study differs considerably from both the Svensson et al. (9) and our own investigations. In the Chirillo et al. (12) series, TEE identified all LITs on the posterior aspect of the ascending aorta, 1 to 40 mm above the left main coronary artery ostium. The Svensson et al. (9) and our series demonstrate that LITs are not restricted to the posterior wall of the ascending aorta, although this is undeniably the best visualized wall by TEE. Our series is the first to describe type B LITs affecting the transverse arch and descending aorta. Although the type B lesions in our series are not histologically proven, we included only cases with a classic appearance where we had high confidence in the diagnosis. Equivocal cases in which a significant amount of atherosclerosis was present on imaging were classified as PAU.
Of note, the Chirillo et al. (12) lesions were invariably smaller (2.8 to 12.3 mm) than any of the LITs in our (17 to 75 mm) or in the Svensson et al. (9) (20 to 100 mm) series, and pathologically, the dominant finding in 5 of 8 cases in the Chirillo et al. (12) series was atherosclerosis, as one might expect in PAU, rather than cystic media degeneration. Atherosclerosis was not a dominant finding in our series.
The pathological results in our patients and in virtually all case reports with pathological correlation support the notion put forward originally by Murray and Edwards (1) that LITs represent one of the possible manifestations of cystic media degeneration.
Why some patients with cystic media degeneration develop AD or IMH and others develop LIT eludes us, but our data suggest that aortic size and age appear to play a role. All LIT patients in the Svensson et al. (9) surgical series had ascending aneurysms. In our series, 94% of patients with type A LITs had aortic diameters >40 mm, and an average diameter of 50 mm. This was similar to classic AD. Type B LITs had aneurysmal aortas as well, which was neither the case in type B AD nor in type B IMH. Patients with LITs were significantly older than patients with classic AD (also after excluding patients with connective tissue disorders) but not significantly older than patients with IMH. Finally, LITs occurred predominantly in the ascending aorta—similar to AD, but unlike IMH and all other causes of AAS, which preferentially involve the descending aorta. Hypertension is an important risk factor in all 3 groups, and connective tissue disease may play a role as well.
Although the sensitivity of CT has been reported at 100% for the detection of classic AD since the late 1990s (27,28), none of the 6 patients who underwent single-slice CT in the Svensson et al. (9) 1999 series (9) and none of the 6 patients in the Chirillo et al. (12) 2007 series undergoing 4-slice CT were accurately diagnosed with an LIT pre-operatively. This is not surprising, given that the defining feature of an LIT may be no more than a subtle aortic contour irregularity, easily confused and obscured by cardiac pulsation artifacts transmitted to the ascending aorta (Figure 2).
In the present series, using 8- to 64-row CT, all but 4 LITs (20 of 24, 83%) were diagnosed on initial CT, with 3 of the 4 missed on outside CT imaging but correctly identified at the time of transfer by readers familiar with this condition. Only a single LIT (1 of 24, or 4%) was initially missed completely even by readers familiar with this pathology, reminding us that the sensitivity of CT for all AAS is not 100%. Although we used ECG gating in almost one-half of our CTAs, it is important to appreciate that most patients with LITs had additional findings alluding to an acute aortic abnormality, and only 6 of 24 patients had isolated subtle tears as the sole finding on CTA. Although ECG gating can eliminate pulsation artifacts and should always be considered in equivocal studies (13), we believe that most LITs can be diagnosed or at least suspected on non-gated CT. Multiplanar reformations improve confidence, and we found that interactive 3D volume-rendered images display the lesions best. The main limitation to detect and correctly classify LITs is not CT or post-processing technology but the familiarity of the interpreting physician and surgeon with this obscure lesion.
Six of 24 LITs in our series—and presumably many more worldwide—were misclassified over the 10-year study period. In 2 patients, the LIT itself was confused with a PAU, which has also been observed in a recent case report (29). In the 2 patients with very focal IMH, and in the 2 patients thought to represent RTA due to periaortic stranding, the luminal tear and aortic out-bulging were not appreciated. Although completely missing a potentially fatal aortic lesion may have serious consequences (25,30), misclassifying an LIT as another acute aortic lesion is arguably of less clinical importance because management of AAS is predominantly dictated by lesion location (type A vs. type B), the presence or absence of complications, and comorbidities rather than the type of lesion.
In-hospital management and outcomes
As with AD, we consider patients with type A LITs for urgent surgical repair, under the presumption that this strategy prevents deaths from aortic rupture. We had no operative deaths in the 11 surgically treated type A LITS, 2 of whom presented with rupture. At the same time, only 1 of the 6 medically managed type A LITs died in hospital, a reminder that the true biology of these lesions remains incompletely understood. LITs—even in the ascending aorta—can heal. One patient with an acute type B AD in our series had evidence of a chronic type A LIT, and 3 LITs in the Svensson et al. (9) series were subacute/chronic as well. If incidentally detected on imaging, chronic type A LITs may have a confusing appearance (26). Chronic type B LITs have also been observed in patients with Marfan syndrome (31). We refer to these lesions informally as “aortic stretch-marks” at our institution. The limited follow-up data available substantially limit any inference concerning long-term survival. It seems fair to say, however, that LIT survival curves generally resemble those of AD and IMH, and that all 3 of these lesions fare much better than RTAs.
In addition to the limited follow-up beyond the initial hospital admission, the main shortcomings of this study are the retrospective design and the small overall number of LITs diagnosed. Although this is the largest series of LITs yet reported, it is unlikely that we captured the entire spectrum of this obscure disease. By design, we excluded patients who did not survive to hospital admission—a limitation of all hospital-based versus population-based studies of AAS (32,33)—and we also excluded patients who may have presented with mild or atypical symptoms, thus not prompting a work-up for AAS.
LITs are rare, subtle lesions within the dissection spectrum, occurring in approximately 5% of patients presenting with AAS. Although subtle, most of these lesions can be detected using contemporary CTA. Post-processing techniques, particularly 3D volume-rendered slabs and endoluminal views, can help identify and display the extent of these lesions, which can occur anywhere in the thoracic aorta, including the transverse and descending aorta. Imagers and physicians involved in the diagnosis and care of patients with AAS must be aware of this dissection variant and remain vigilant when an obvious aortic lesion is not found but is clinically suspected. The underlying pathology, natural history, and risk of rupture imply a treatment approach similar to AD and IMH with urgent surgical repair of type A lesions, and medical management of type B lesions unless complications occur. Given a similar underlying pathology (cystic medial degeneration) and presumed comparable long-term outcomes compared with AD and IMH, lifelong surveillance is warranted.
COMPETENCY IN MEDICAL KNOWLEDGE: Limited intimal tears are an uncommon, but potentially lethal, variant of aortic dissection. Contemporary high-quality computed tomography imaging can detect these subtle lesions pre-operatively.
TRANSLATIONAL OUTLOOK: Multicenter prospective and observational studies are required to further elucidate the natural history and long-term outcomes of limited intimal tears.
This work was supported by National Institutes of Health Clinical and Translational Science Award UL1 RR025744. Dr. Willemink has served on the speakers bureau of Philips Healthcare. Dr. Fischbein has received a speaking honorarium from Abbott. Dr. Miller has been a principal of coinvestigator on trials for Abbott Vascular, Medtronic, and Edwards Lifesciences; and has received (modest) fees as a consultant for Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute aortic syndrome
- aortic dissection
- computed tomography
- computed tomography angiography
- electronic medical records
- intramural hematoma
- interquartile range
- limited intimal tear
- penetrating atherosclerotic ulcer
- rupturing/leaking thoracic aortic aneurysm
- transesophageal echocardiography
- Received January 18, 2018.
- Revision received March 4, 2018.
- Accepted March 21, 2018.
- 2018 The Authors
- Murray C.A.,
- Edwards J.E.
- Svensson L.G.,
- Labib S.B.,
- Eisenhauer A.C.,
- Butterly J.R.
- Hiratzka L.F.,
- Bakris G.L.,
- Beckman J.A.,
- et al.
- Kalogerakos P.D.,
- Kampitakis E.,
- Pavlopoulos D.,
- Chalkiadakis G.,
- Lazopoulos G.
- Halushka M.K.,
- Angelini A.,
- Bartoloni G.,
- et al.
- Evangelista A.,
- Maldonado G.,
- Gruosso D.,
- Teixido G.,
- Rodriguez-Palomares J.,
- Eagle K.
- Sarraj A.,
- Munoz D.E.,
- Calle Valda C.M.,
- Monguio E.,
- Reyes G.
- Shiga T.,
- Wajima Z.,
- Apfel C.C.,
- Inoue T.,
- Ohe Y.
- Mitsomoy M.F.,
- Alexoiu V.,
- Kirsch M.
- Nagra K.,
- Coulden R.,
- McMurtry M.S.
- Howard D.P.,
- Banerjee A.,
- Fairhead J.F.,
- et al.