Imaging and Surveillance of Chronic Aortic Dissection: A Scientific Statement From the American Heart Association

Abstract

All patients surviving an acute aortic dissection require continued lifelong surveillance of their diseased aorta. Late complications, driven predominantly by chronic false lumen degeneration and aneurysm formation, often require surgical, endovascular, or hybrid interventions to treat or prevent aortic rupture. Imaging plays a central role in the medical decision-making of patients with chronic aortic dissection. Accurate aortic diameter measurements and rigorous, systematic documentation of diameter changes over time with different imaging equipment and modalities pose a range of practical challenges in these complex patients. Currently, no guidelines or recommendations for imaging surveillance in patients with chronic aortic dissection exist. In this document, we present state-of-the-art imaging and measurement techniques for patients with chronic aortic dissection and clarify the need for standardized measurements and reporting for lifelong surveillance. We also examine the emerging role of imaging and computer simulations to predict aortic false lumen degeneration, remodeling, and biomechanical failure from morphological and hemodynamic features. These insights may improve risk stratification, individualize contemporary treatment options, and potentially aid in the conception of novel treatment strategies in the future.

Chronic aortic dissection has received little public or scientific attention compared with the dramatic events associated with acute aortic dissection.^1,2 An increasing number of patients survive the acute phase of the disease and outlive the initial hospitalization for many years.^3–7 Late complications, driven predominantly by chronic false lumen degeneration and aneurysm formation, often require surgical, endovascular, or hybrid interventions to treat or prevent aortic rupture. The requirement for continued, lifelong surveillance and specialized medical and surgical care for these complex patients represents a growing health care burden.⁸

Imaging plays a central role in the surveillance of patients with chronic aortic dissection. Monitoring the inevitable degeneration and aneurysmal dilatation of any residual aortic false lumen is critical for clinical decision-making and timing of interventions. Accurate aortic diameter measurements and systematic documentation of diameter changes over time and across imaging modalities pose technical and practical challenges that have not been adequately addressed in the literature. In this scientific statement, we systematically examine those challenges and present state-of-the-art imaging and measurement techniques that can be assembled into a standardized reporting system for patients with chronic aortic dissection.

In addition to its established role in the surveillance of patients with chronic dissection, modern imaging provides new and potentially important insights into the pathophysiology, hemodynamics, and mechanobiology of chronic aortic dissection. Imaging and computer simulations of blood flow allow patient-specific estimates of luminal pressures and wall tension.^9,10 Wall tension directly drives false lumen expansion, ultimately leading to rupture, if it exceeds the remodeling capability and mechanical strength of the cellular and extracellular matrix constituents of the aortic wall.¹¹ Understanding the complex interplay among morphological, hemodynamic, and biomechanical factors is a prerequisite for developing risk prediction models, for personalizing contemporary treatment options, and for conceiving and developing new treatment strategies.

EPIDEMIOLOGY AND ANATOMIC SPECTRUM

Although aortic dissection typically presents as an acute event with high mortality, aortic dissection is inherently a chronic disease. Medial degeneration is a prerequisite for aortic dissection to occur in most if not all patients and likely precedes the acute event by many years.¹² Survivors of the acute event may live with the long-term sequelae of a dissected aorta for many years.

Originally, chronic aortic dissection was defined as the period after the first 14 days after onset of symptoms¹³ because most dissection-related complications occur within this time.^14,15 A refined, contemporary classification subdivides the phases of aortic dissection into a hyperacute phase (first 24 hours) within the acute phase (up to 14 days after onset), followed by a subacute phase (15–90 days after onset) and chronic phase (>90 days after onset).^16,17

Prevalence and Economic Impact

The prevalence of patients living with chronic aortic dissection in the United States is unknown and difficult to ascertain. This is an important gap in our knowledge and limits our ability to estimate the long-term quality of life and cost-effectiveness of evolving, personalized treatment or prevention strategies. In the first postevent year alone, the costs of care for the relatively small population of patients with chronic aortic dissection are substantial compared with the general population.⁸ The cumulative expenses for long-term specialty care, rehabilitation, surveillance, and preoperative and postoperative imaging, combined with frequent readmissions for elective or emergency surgical or endovascular repair, are likely to further escalate over time. Unfortunately, actual health care cost data beyond the first year after index hospitalization are not readily available.⁸

A rough approximation of the prevalence of chronic aortic dissection, assuming a conservative incidence rate of 4 to 6 per 100 000 per year^18–21 and conservatively assuming an average survival of 7 years for all subgroups, suggests a prevalence of 28 to 42 per 100 000 or 92 000 to 138 000 survivors of acute dissection in the United States (Figure 1). This corresponds well to the age-adjusted prevalence of 26 per 100 000 that is based on hospital administrative data in Manitoba, Canada, which has a reported incidence of acute dissections of 3 per 100 000 per year.²⁹ For radiologists and others who routinely interpret vascular computed tomography (CT) studies, these numbers imply that many more CT scans are performed during the chronic than the acute phase of aortic dissection. The much larger proportion of chronic versus acute phase clinical imaging studies is illustrated in a small cohort of 47 patients with uncomplicated type B dissection occurring between 2003 and 2012³⁰ who were followed up for up to 17 years (Figure 2). In addition to their 47 baseline scans, these patients underwent a total of 363 follow-up scans, or 7.7 follow-up scans per patient.

Figure 1. Derivation of the anatomic spectrum of chronic aortic dissection.

Flow diagram (left) and corresponding distribution charts (pie charts, right) illustrate how differences in survival transform the anatomic spectrum from acute to chronic aortic dissection. Patients are stratified by anatomic spectrum (type A vs type B), clinical course (for type B, complicated vs uncomplicated), and treatment (medical vs open or endovascular surgery). Approximately 60% of early survivors have surgically repaired type A dissections; 25% have uncomplicated type B dissections; and 15% are survivors of complicated type B dissections. This distribution likely persists given the comparable midterm and long-term survival between surgically repaired type A and medically or surgically managed chronic type B aortic dissections. Post-op TAAD indicates postoperative type A aortic dissection; TEVAR, thoracic endovascular aortic repair; and Tx, treatment. Flowchart based on International Registry of Acute Aortic Dissection registry (Pape et al²²) and Registry of Acute Aortic Syndrome-II registry (Evangelista et al⁶) data, differential survival of type B dissections from Charilaou et al²³ and Afifi et al,²⁴ and midterm and long-term survival data from others.^{3,4,7,24–28}

Figure 2. Computed tomography (CT) scans in 47 patients with aortic dissection.

Diagram plots all baseline (n=47; red dots) and follow-up CT scans (n=363; black dots) performed in a cohort of 47 consecutive patients (y axis) with an uncomplicated type B aortic dissection between January 2003 and December 2012 who were followed up in an aortic surveillance program until August 2020 over time (x axis). A small number of scans were performed before the index events (patients with known aneurysms or connective tissue disease [circles]).

Anatomic Spectrum of Chronic Aortic Dissection

The spectrum of patients living with chronic aortic dissection at a given point in time reflects the cumulative outcomes of multiple past events and depends on (1) the initial spectrum of the disease, (2) the differences in survival between groups, and (3) the outcomes of any surgical or endovascular interventions. Figure 1 illustrates the changing spectrum of disease from acute to chronic aortic dissection for a putative contemporary cohort of patients with chronic aortic dissection. Any recent changes in management or outcomes will be reflected in future chronic aortic dissection cohorts for years to come.

The largest group of patients with chronic aortic dissection (≈60%) are patients with surgically corrected ascending (Stanford type A) dissection with a persistent false lumen distal to the surgical repair (Supplemental Video). In the acute phase, the proportion of type A dissections is even higher, up to 75%,¹⁸ but not all patients survive. Patients with nonoperated type A dissection rarely live past the acute event; hence, chronic dissections of the ascending aorta are exceedingly rare. A small subset of patients with type A dissection limited to the ascending aorta (DeBakey type II, 10% in the International Registry of Acute Aortic Dissection³¹) are theoretically cured if the entire dissected segment is resected and no residual false lumen remains. No sustained benefit of type II morphology has been conclusively demonstrated, however.³¹ In the majority of surgically treated patients with type A dissection (DeBakey type I, 90%), the false lumen extends beyond the ascending aorta. The in-hospital mortality for acute type A dissection has been reported to be between 22% and 47%.^{3,5,6,19,22,32} In tertiary high-volume referral centers for aortic diseases, the surgical mortality for ascending and hemiarch replacement with or without root replacement is as low as 10% to 20%.^33–35 The subsequent midterm survival of surgically corrected ascending dissections at 1, 3, and 5 years is remarkably good at 95%, 93%, and 89%, respectively.^4,7,36

Approximately one-third of acute dissections are Stanford type B, defined as dissections not involving the ascending aorta. Patients with uncomplicated type B dissection, that is, those without signs of rupture, rapid expansion, or branch vessel malperfusion³⁷ (≈60%–70% of type B dissections) can be safely managed medically.^24,36 The in-hospital mortality of these patients is extraordinarily low (0%–2.2%).^23,24 Almost all patients with initially uncomplicated type B dissection therefore progress to the chronic phase, at which point they represent the second largest fraction (≈25%) of patients with chronic dissection. Complicated type B dissections (ie, those requiring urgent endovascular or surgical repair for aortic rupture, rapid expansion, or branch vessel malperfusion, as well as those patients too sick to undergo an acute intervention) represent ≈30% to 40% of type B dissections. The in-hospital mortality of these patients is unsurprisingly higher (≈20%).^6,24 The survivors of this group include patients treated successfully with surgical or endovascular approaches and represent ≈15% of patients with chronic dissection. This percentage may change as a result of the current trend of using thoracic endovascular aortic repair (TEVAR) earlier in patients with uncomplicated acute or subacute type B dissections (Figure 1).^38,39

The proportion of patients with chronic dissection who require ≥1 elective or urgent surgical or endovascular interventions increases over time in all groups. Although the occurrence of late complications and the need for emergency or elective intervention are the competing risks that each of these patients must face, many patients continue to live beyond these events. The long-term survival of patients with type A dissection (65%–89% after 5 years and 45%–65% after 10 years, respectively)^{3,4,7,25,27,28,36} is comparable to that of patients with type B dissection (84% after 5 years in patients with endovascular treatment; 71% after 5 years and 35%–69% after 10 years in medically managed patients).^26,40,41

In addition to classic aortic dissection, 2 dissection variants, intramural hematoma⁴² and limited intimal tears,^43,44 also may evolve into the equivalent of a chronic aortic dissection, leaving those patients at risk for subsequent expansion of the false lumen–equivalent, thinned outer aortic wall of these respective lesions.⁴⁵

PATHOPHYSIOLOGY

Several alterations occur over time in the dissected aorta. Above all, the dissected aorta grows more rapidly than the nondissected aorta (3.1 mm/y versus 0.5 mm/y).⁴⁶ Recent studies from multiple centers have confirmed these estimates, revealing growth rates of 2.3 to 4.1 mm/y.^47–50 A detailed study from the Aortic Institute at Yale University showed that growth during the first few days after aortic dissection is rapid but decreases substantially by 30 days to a slow but inexorable growth (Figure 3).⁵¹ In the same study, Peterss and colleagues⁵¹ revealed that the initial remodeling phase of the outer false lumen aortic wall is characterized pathologically by a high grade of cellular destruction in the aortic media, followed by substantial fibrosis in the chronic phase. In the absence of intact medial smooth muscle cells, adventitial fibroblasts remain as the primary cell population responsible for restoring and maintaining the structural integrity of the outer wall of the aortic false lumen.¹¹

Figure 3. Morphological changes in aortic dissection.

Orthogonal cross sections through the descending thoracic aorta at baseline (day 1) and 3 months, 6 months, 3 years, 8 years, and 9 years after the acute event. In the acute phase, the dissection flap (arrowheads) is highly mobile but becomes less mobile at 3 months and thicker and immobile after 6 months. The false lumen degenerates and becomes aneurysmal over time. The maximum aortic diameter increases from 39 to 43 mm in the first 6 months (annualized growth, 8 mm/y) and after 9 years becomes 55 mm (annualized growth, 1 mm/y from 6 months–9 years). False lumen thrombus (Thr.) is evident in years 8 and 9.

Similar pathological changes alter the dissection flap over time, which becomes thicker, fibrotic, and less mobile. The increasing stiffness of the dissection membrane increases pressure gradients between the true lumen and false lumen.¹⁰

Late Complications

In addition to the common but gradual development of false lumen aneurysms large enough to warrant elective repair, patients with chronic aortic dissection remain at risk for less common but acute and potentially lethal events requiring emergency intervention.³⁰

Emergency Late Complications

Approximately 6% of patients with chronic dissection require an emergency intervention to treat new limb or mesenteric ischemia.⁴¹ New dissections or redissections occurred in ≈5% of patients in 2 single-center studies.^52,53 New type A dissections in patients with chronic type B dissections occurred in 3.6% of medically managed patients, 4.3% of patients undergoing open aortic surgery, and 6.9% of patients with TEVAR.⁵² Acute redissection of a chronically dissected distal aorta, resulting in a new, third flow channel, is strongly associated with Marfan syndrome (21.5% versus 3.1%).⁵³ Ensuing further thinning of the already dissected wall requires prompt intervention (Supplemental Figure 1). The most feared complication, late aortic rupture, is reported in 3% to 18%.^54–56 However, the actual rate is difficult to establish because, in most patients undergoing surveillance who die suddenly, a cause is not identified.

Aneurysm Formation

Aneurysms develop in 34% to 38% of patients with chronic type B aortic dissection^26,55,56 and in up to 49% of patients with residual chronic aortic dissection in the distal aorta after type A aortic repair.⁵⁷ Aneurysm formation is the primary reason for reoperations in patients with chronic aortic dissection.⁵⁵ Reoperation rates in residual chronic type A aortic dissection range between 9% and 27.5%, with some patients requiring >1 reintervention.^27,54,57 Intervention-free survival in type B aortic dissection is reported to be between 41% and 66%, depending on complication status at the initial event.^24,55,56

Hemodynamics and Mechanobiology

The complex interplay among morphological, hemodynamic, and biomechanical factors contributing to false lumen degeneration and aneurysm formation is incompletely understood. This is a critical gap in our knowledge that impedes our efforts to solve important problems. First, we currently lack reliable prediction models to guide treatment in patients with uncomplicated type B aortic dissection.⁵⁸ We possess a wealth of imaging data attesting to the vast individual variation of aortic dissection manifestations. However, without understanding how an individual’s constellation of dissection features translates into false lumen pressure, wall tension, and growth, we may not be able to synthesize a reliable prediction tool that can be clinically tested. Second, only by understanding the interaction of morphology, hemodynamics, and biomechanics will we be adequately equipped to anticipate the effect of interventions on these parameters, which control the long-term fate of the false lumen and thus patient outcome. A broader interdisciplinary approach to the conundrum of the false lumen may help to develop novel therapies.

New insights are currently emerging from the combination of imaging data (3-dimensional [3D] and 4-dimensional morphological data from CT,^30,59,60 flow data from magnetic resonance imaging [MRI]⁶¹), computer simulations (computational fluid dynamics,⁶² fluid-structure interaction^9,10,63), bench-top and in vitro phantom studies,^64,65 biomechanical bench testing,⁶⁶ and numerical modeling of aortic wall constituents,^66–68 as well as from the field of mechanobiology.^11,69 A graphical synopsis of factors contributing to false lumen dilatation is shown in Figure 4. The key points are as follows:

Figure 4. Pathophysiology and biomechanics of chronic aortic dissection.

Schematic illustrates the complex interplay between morphological features derived from imaging, with hemodynamic, biomechanical, and mechanobiological factors leading to false lumen degeneration and aneurysm formation and ultimately rupture in chronic aortic dissection. See text for details. CFD/FSI indicates computational fluid dynamics/fluid structure interaction; ECM, extracellular matrix; FL, false lumen; PIT, primary entry tear; and TL, true lumen.

• The outer wall of the false lumen is a load-bearing structure derived from the aortic adventitia left intact after the initial delamination of the intima and variable amounts of the media layer of the aortic wall.

• Restoration of false lumen wall strength and homeostasis is maintained by adventitial myoblasts/fibroblasts, which are sensitive to changes in wall tension (mechanosensing), responding with proliferation and increased synthesis and cross-linking of collagen. This is noticed at late surgical intervention as a strong, leathery thickening of the adventitial layer.

• Even a marginally inadequate tissue response to wall tension can enter a vicious cycle in which any increase in false lumen diameter proportionally translates into an ever-increasing wall tension, resulting in continuous growth of aortic diameters over time.

• Once the restorative capabilities are exhausted and the wall tension exceeds the tensile strength, structural failure and aortic rupture ensue. This occurs with increasing frequency at maximum aortic diameters >55 to 60 mm.

• The wall tension of the false lumen is not only determined by its thickness and geometry (or its clinical proxy, the maximum aortic diameter) but is also directly proportional to the hydrostatic pressure in the false lumen.

• The false lumen pressure is a function of true lumen pressure, modulated by blood flow alterations specific to morphological dissection features. Although the systolic pressure is always equal or higher in the true lumen, an uninhibited inflow into the false lumen (enabled by a large enough primary entry tear), combined with high outflow resistance from the false lumen, results in a reversal of the pressure gradients in diastole. Thus, the diastolic and mean pressures can be higher in the false lumen than in the true lumen.

• False lumen outflow resistance is related to the number of branch vessels draining the false lumen (more draining branches decrease resistance) and to the number, size, and location of communications between the true and false lumens across the dissection flap (fenestrations, re-entry tears).

• The dissection membrane also modulates the pressure gradients between true and false lumens. A mobile flap decreases and a stiff flap augments the pressure differences between the true lumen and the false lumen.

MANAGEMENT OF CHRONIC AORTIC DISSECTION

Watchful Waiting

The mainstay of management of chronic aortic dissection is (1) antihypertensive therapy, including anti-impulse therapy with β-blockade and afterload reduction; (2) lifelong surveillance with imaging; and (3) timely repair once the dissected aorta becomes significantly aneurysmal.^37,70

Anti-impulse therapy was pioneered by Wheat et al⁷¹ in 1965 to prevent dissection progression in the most acute phase. The effectiveness of combination β-blocker and afterload reducing drugs treatment is unquestioned in both the acute and chronic phases of aortic dissection.⁷² It is important to note that unopposed afterload reduction can lead to unwanted reflex tachycardia.

The rationale for the traditional “watchful waiting” strategy, conceived in the pre-endovascular era, is based on the need to balance the expected morbidity and mortality of preemptive open surgical repair against the risk of fatal aortic rupture. Currently accepted size thresholds were established from the 1997 landmark work of Cody and colleagues⁴⁶ from Yale University demonstrating rapidly increased risk of rupture once the aneurysmal aorta exceeds 6 cm (Supplemental Figure 2). The morbidity and mortality associated with contemporary surgical and endovascular repair are evolving; however, and the optimal treatment and timing for intervention in patients with chronic aortic dissection are becoming less certain.

Current guidelines, published between 2010 and 2017, consistently recommend intervention in patients with aneurysms >6.0 cm (Class I), with TEVAR to be considered in high-risk patients or nonsurgical candidates (Class IIa).^37,70,73 Open surgical repair is recommended in patients with connective tissue disease with an aneurysm diameter of 5.0 to 5.5 cm (Class I or IIa). TEVAR may be considered at that same 5.5-cm-diameter threshold as a Class IIb indication in more recent guidelines.^37,73 Of note, all pertinent guidelines acknowledge the poor quality of evidence for TEVAR recommendations in patients with chronic dissection and the lack of better evidence-based risk stratification and stress the need for individualized decision-making, ideally by a multidisciplinary team of experts.^{37,70,73–76}

The time intervals suggested for surveillance imaging in patients with chronic aortic dissection vary across guidelines and should be tailored for individual patients.^{37,70,73,75,76} Considering the natural course and reported growth rates of the dissected aorta,^46,51,77 several reasonable follow-up regimens have been proposed for different treatment groups, preferably beginning with a baseline predischarge scan. After surgical repair, patients could be followed up after 6 and 12 months and annually thereafter. If a moderate-sized aneurysm remains relatively stable over time, subsequent longer follow-up intervals of 1.5 to 3 years are reasonable.⁷⁰ In medically managed uncomplicated type B dissection, 1 additional earlier scan, for example, at 3 months, may reveal important changes occurring in the subacute phase, when the dissected aorta remains amenable to early TEVAR. Earlier follow-up imaging at 30 days is sometimes recommended to identify potential early complications, followed by routine 6-month and 12-month follow-up.^37,73,75,76 Continued annual imaging is suggested thereafter, even if the aorta stabilizes after TEVAR. If growth or other complications are detected, shorter follow-up intervals may be required in all treatment groups.⁷⁰ Eventually, most patients with chronic aortic dissection develop substantial aortic dilatation, especially in the proximal descending aorta, just beyond the left subclavian artery. These patients can be operated on safely with either open aortic replacement or endovascular therapy, depending on anatomy, age, and connective tissue disorder diagnosis.²³

Surgical Treatment of Chronic Aortic Dissection

Open surgical repair remains the dominant method of managing chronic aneurysmal thoracoabdominal aortic dissections given (1) the limitations of current endovascular techniques, including adequacy of landing zones, the noncompliant dissection flap, and the uncertain durability of TEVAR, and (2) the beneficial impact that advances in circulation management, cerebrospinal fluid drainage, neurocerebral monitoring, and perioperative management protocols have had on surgical outcomes.

Variability in the natural history of chronic descending aortic dissections demands careful monitoring by cardiovascular surgeons as part of a multidisciplinary team including vascular surgeons, cardiologists, and radiologists with interest and expertise in the management of aortic pathology. Acute ascending aortic dissections, despite proximal repair, often incur residual distal arch or descending and thoracoabdominal aortic dissections. Despite similarities to chronic type B aortic dissections, residual type A dissections often have different implications with respect to managing the aortic arch.

Indications for open operative intervention in uncomplicated chronic aortic dissections hinge primarily on aortic caliber, pathogenesis, and occasionally symptoms. Asymptomatic chronic thoracic and thoracoabdominal aortic dissections with a maximal orthogonal aortic diameter of 6.0 cm in patients without connective tissue disease warrant repair.^37,46,70 Rapid growth rate also may be an indication for repair. Patients with connective tissue diseases and those with a family history warrant more careful consideration and may warrant earlier intervention (diameter of 5.0–5.5 cm).^78,79 Griepp and colleagues⁸⁰ have used a formula that predicts risk according to aneurysm size, age, and the presence of comorbidities such as renal insufficiency and emphysema.⁵⁶ Operative morbidity risk with repair of aneurysmal chronic descending and thoracoabdominal aortic dissections is substantial and potentially includes respiratory insufficiency, renal failure, myocardial infarction, stroke, left vocal cord paralysis, wound infections, and paraplegia. The relative risk of complications imparted by chronic dissection in the context of thoracoabdominal aortic aneurysm replacement has not been distinctly determined by any of the large series. In the largest series (3309 cases) of thoracoabdominal aortic aneurysm repair by Coselli et al,⁸¹ 36% of cases were operated on for chronic aortic dissection: Operative mortality was 7.5%; dialysis dependence rate was 5.7%; and paraplegia rate was 2.9%. Smaller recent reports on outcomes in this specific patient population are similar with mortality rates ranging from 5.8% to 11%.^82–89 There remains no consensus among the large thoracic aortic centers around the world as to the optimal operative repair strategy for chronic aneurysmal thoracoabdominal aortic disease. Limiting the repair to the aneurysmal segment rather than the entire dissected aorta reduces the risk of complications and paraplegia, but several techniques have been successful in delivering acceptable morbidity and mortality for this difficult problem. The 2 centers with the lowest reported mortality for thoracoabdominal aortic aneurysm repair in the context of chronic dissection have used deep hypothermic circulatory arrest for their technique.^83,90

Endovascular Therapy

Open surgical repair may be too morbid for some patients with chronic dissection and is often postponed for too long.^88,91 TEVAR provides another mechanical treatment option that has been increasingly used because of the low acute risk compared with open repair and the potential to reduce aortic mortality and reoperations compared with medical therapy.^92,93 However, TEVAR faces anatomic and morphological limitations to successfully treating this disease, particularly in the chronic stage.^94–97

A TEVAR-based strategy seeks to seal intimal entry tears with device fabric and to expand the true lumen by outward radial force of the stents. In patients with chronic aortic dissection and aneurysm, covering entry tears is an important treatment objective to promote thrombosis and false lumen remodeling.⁹⁷ A patent false lumen has been associated with decreased survival and increased risk of reintervention.⁹⁸ TEVAR of chronic dissection may promote false lumen thrombosis at a variable rate over time. It has been shown, however, that thrombosis needs to be complete to reverse the natural history of disease.^96,97

In most patients with type B aortic dissection, the proximal entry tear can be covered with an endograft. Coverage of the left subclavian artery is often necessary to achieve an adequate proximal landing zone. Patients with dissections limited to the descending thoracic aorta (DeBakey type IIIA) often exhibit an adequate landing zone in the distal descending thoracic aorta.^41,99 Unfortunately, most dissections extend across the visceral branch origins. Re-entry tears and fenestrations of the dissection flap within the abdominal aorta frequently maintain false lumen perfusion and pressurization. Increased number and size of distal re-entry tears and false lumen origin of branch vessels are associated with persistent false lumen flow.^97,100

Several strategies can be used to facilitate a distal seal of an excluded false lumen aneurysm to prevent retrograde filling such as percutaneous “cheese-wire” flap fenestration, balloon septum rupture (Knickerbocker technique), laser aortic septotomy, or placement of coils of plugs (candy plug technique) for false lumen embolization (Supplemental Figure 3). Additional endovascular therapies may require the use of bare stents to expand the true lumen with or without ballooning or the use of investigational or homemade branched devices.^101–111

Role of Imaging for TEVAR Planning and Follow-Up in Chronic Dissection

Endovascular treatment planning is typically based on preoperative CT angiography (CTA). A fundamental requirement for successful TEVAR is an adequate device landing zone of healthy aorta of at least 15 mm in length and between 18 and 42 mm in diameter that will allow a parallel lie of the sealing endograft. Predischarge, assessment of adequate device position, successful coverage of communications and tears, and assessment of branch vessels that can contain re-entry tears is important. Surveillance imaging is critical to monitor aortic dimensions, persistent false lumen patency, or thrombosis and any new complications. Adequate imaging assessment after TEVAR requires a delayed-phase scan to identify small leaks that slowly fill yet pressurize the false lumen.

Considerations in Patients With Connective Tissue Disease

There is limited information about the best strategy to manage chronic dissections in patients with a genetic predisposition to aortic diseases. Although dissections are more frequent and occur at a younger age in these groups than in the general population, the overall prevalence is low, challenging the design of comparative trials. Most available data come from registries and retrospective studies that have not examined determinants for interventions.

In an analysis of registry data, patients with the Marfan syndrome were more likely to have an intervention than medical treatment when presenting with an acute type B dissection.¹¹² Although the reason for this was not specifically captured, a possible explanation is that patients with Marfan syndrome tend to have extensive aortic involvement,^112,113 which may influence the decision. There was no difference in organ malperfusion.

Although endovascular procedures can be lifesaving in the acute setting in these patients, open surgery is preferred for chronic dissections because of the concern for retrograde dissection proximally, early graft failure attributable to tensile forces of the graft in an abnormal aortic wall, and increased need for reinterventions after endovascular repair.¹¹⁴

Information on chronic dissections in Ehlers-Danlos syndrome (vascular type) is even scarcer. A multicenter cross-sectional retrospective study of 86 patients¹¹⁵ reported 17 with aortic pathology treated by medical management; thus, the authors recommended intervention only for distinct complications. Open surgery is preferred over an endovascular approach on the basis of the risk of erosion or perforation at the landing zones of the endograft resulting from the fragility of the aortic wall.^115–117

IMAGING TECHNIQUE

Computed Tomography

CTA is the most commonly used modality for imaging of patients with acute and chronic aortic dissection.^118–120 State-of-the-art CT scanners are widely available and capable of acquiring high-resolution volumetric data sets that can be reformatted into a wide range of 2-dimensional and 3D angiographic images.

A typical imaging protocol includes an optional noncontrast scan for identification of surgical grafts and high-attenuation foci that can mimic pseudoaneurysm, followed by a 50- to 150-mL contrast medium–enhanced angiographic acquisition. Additional delayed-phase scans are required in post-TEVAR patients to assess for endoleaks and can help distinguish slow flow–related unenhanced regions in the false lumen from thrombus. Scan times for each acquisition range between 5 and 20 seconds. Measurements of the ascending aorta are more reliable when electrocardiographic gating is used to suppress cardiac pulsation artifacts,¹²¹ although the scan time of many modern scanners is short enough to render electrocardiographic gating less necessary. If retrospective electrocardiographic gating is used, CT images can be reconstructed and viewed over the entire cardiac cycle, adding a fourth dimension to CTA.

The effective radiation dose from a single arterial phase CT scan is ≈8 mSv (range, 2–20 mSv) and from a 3-phase scan is ≈25 mSv (range, 3–60 mSv).^122,123 Even in the setting of lifelong surveillance with annual CT scans, the estimated excess risk of cancer attributable to CT radiation is low, between 0.43% and 0.65% for patients in their early 50s and even less in older individuals.^122,124 For comparison, the lifetime risk of cancer is almost 100 times greater at 44%.¹²⁵ The known benefits of a clinically indicated CT scan in the typical age group of patients with chronic dissection almost certainly exceed the uncertain and theoretical risks of imaging-related radiation exposure.

Magnetic Resonance Imaging

MRI is an accepted alternative to CTA for the assessment of patients with chronic aortic dissection.^126,127 Advantages of MRI include lack of ionizing radiation,⁷⁰ which may be a particularly important consideration in young individuals such as patients with Marfan syndrome in their 20s or 30s with near-normal life expectancy, especially before the aorta becomes dissected, substantially aneurysmal after dissection, or difficult to assess after serial interventions. Disadvantages of MRI compared with CT include a longer scan duration, lower spatial resolution, artifacts from implanted devices such as endografts or mechanical heart valves, and potential contraindications such as certain metallic implants.

A typical MRI protocol includes multiple sequences. Contrast-enhanced magnetic resonance angiography (MRA) with gadolinium-based contrast agents provides high-quality 3D data sets. Like CTA, MRA allows multiplanar evaluation to obtain orthogonal measurements. Recent advances in noncontrast MRA techniques have made this a viable alternative in patients with contraindications to gadolinium-based or iodinated contrast agents.¹²⁸ The addition of phase-contrast imaging allows the assessment of blood flow and false lumen thrombosis.¹²⁹

Four-dimensional flow MRI is an emerging technique that allows exquisite characterization of altered flow dynamics in both true and false lumens and quantification of parameters such as wall shear stress, which may have prognostic value.^130–133 Given the potential role of inflammation in the progression of aortic aneurysms, the recent integration of hybrid positron emission tomography/MRI systems may provide another avenue for simultaneous assessment of anatomic, physiological, and molecular imaging biomarkers.¹³⁴

Echocardiography

Transthoracic echocardiography (TTE) is commonly performed in patients with chronic aortic dissection with concomitant dilatation of the unrepaired aortic root. In select patients, TTE and transesophageal echocardiogram (TEE) also can be useful to evaluate the dissection flap,³⁷ to demonstrate communications between the true and false lumen, and to assess blood flow versus thrombus in the false lumen. TTE and TEE do not require contrast agents or ionizing radiation. The disadvantages of echocardiography are that not all aortic segments can be interrogated and TEE requires sedation.

TTE is noninvasive and allows evaluation of the aortic valve, root, and ascending aorta at each follow-up. In some patients, the arch can be visualized through the suprasternal view. The descending aorta is only occasionally visualized.¹³⁵

TEE allows visualization of the aortic arch but requires sedation and has a blind spot in the distal ascending aorta caused by the interposition of the right tracheobronchial angle.¹³⁵ For acute dissections, TEE has a high sensitivity (86%–100%) and specificity (90%–100%) compared with CTA.¹³⁶ In chronic dissections, 3D TEE and 2-dimensional TEE have been compared with CT and were equivalent in locating and measuring the entry tear size, and they were better at detecting secondary tears.¹³⁷

MEASUREMENT PRINCIPLES

The quintessential information sought from imaging studies in patients with chronic aortic dissections is aortic caliber. The maximum aortic diameter is strongly associated with the risk of aortic rupture and currently remains the predominant parameter for surgical decision-making.^{37,46,70,73–76} Measurements of adjacent segments as potential landing zones, separate diameter measurements for true and false lumen, and access vessel size are also important for determining anatomic eligibility and treatment planning for endovascular repair.¹³⁸

Despite its importance and the large number of measurements obtained every day in hundreds of patients with chronic aortic dissection, few specifics on how to perform and compare measurements of the dissected aorta can be found in the literature.^119,120

Definition of Aortic Diameters

The aorta is a curved, tubular structure, often with a non-circular cross section, tapering from proximal to distal, while giving off multiple branch vessels. Aortic diameters refer to the cross-sectional distances obtained in a plane that is orthogonal to the aortic flow channel, often represented as its medial axis¹³⁹ or aortic centerline. On a given orthogonal cross section, aortic diameters are expressed as the diameters of the long (major) axis and the short (minor) axis, respectively (Figure 5). The maximum aortic diameter is the largest of all major axis diameters on any of the aortic cross sections along the aorta.

Figure 5. Definition and measuring of the maximum aortic diameter.

Maximum diameter is measured on double-oblique reformations oriented orthogonal to the aortic flow channel. Note the remarkable difference between the long-axis diameters between the transverse images (47 mm vs 54 mm). Major and minor axis diameters include the luminal thrombus if present. Ann indicates aortic annulus; bif, bifurcation; Darc, distal arch; Ddes, distal descending aorta; MidAsc, middle ascending aorta; Parc, proximal arch; Pdes, proximal descending aorta; SOV, sinus of Valsalva; and STJ, sinotubular junction.

In addition to the maximum aortic diameter, aortic measurements typically include diameters at predefined landmarks along the aorta such as the aortic root at the annulus, at the sinuses of Valsalva, and sinotubular junction, as well as select locations in the ascending aorta, in the aortic arch, and along the descending thoracoabdominal aorta.^140,141

Measurement Technique

There is unanimous agreement that aortic diameters should be measured perpendicular to the aortic long axis.^142,143 CTA and MRA data are 3D volumes consisting of stacks of narrow-spaced 2-dimensional slices. Obtaining the appropriate measurement plane in CTA and MRA therefore requires reformatting the data into so-called double-oblique images or multiplanar reformations (MPRs).

MPRs can be obtained either manually or with the assistance of (semi)automatically generated aortic centerlines. With the centerline method, perpendicular MPRs can easily be created, and distance measurements can be obtained along the aortic axis. The aorta can also be unfolded along its centerline as a straightened- or a curved-planar reformation, which also facilitates length measurements, in addition to diameters.¹⁴⁴

In patients with chronic dissection, selecting the best MPR orientation becomes more difficult as the false lumen becomes larger and approaches the surgical size threshold. Asymmetry and stark differences in contrast enhancement between the true and false lumen, with or without the presence of false lumen mural thrombus, often preclude the use of the centerline method at this stage, whereas increasing tortuosity and kinking caused by aortic elongation pose a formidable challenge to manually selecting the best MPR orientation, even for experienced users. Selection of MPR location and orientation is likely the largest source of measurement uncertainty and discrepancies in patients with chronic dissection and false lumen aneurysms.

Aortic Measurements on Transverse Images

Axial images are commonly reviewed to obtain a first impression and approximate size of a patient’s aorta. It cannot be overstated, however, that transverse (axial) images are unsuitable to accurately determine the maximum aortic diameter.¹⁴⁵ Incorrectly using the largest dimension of an oblong aortic cross section on axial images has been shown to overestimate vessel size by up to 6 to 7 mm on both CT and MRA.^146,147 If an aortic diameter must be measured on axial images (eg, in an emergency or without CT data or 3D viewing capability), the short-axis diameter is a closer approximation of the true diameter and should be recorded.^148,149 However, this approach may underestimate the true maximum aortic diameter in patients with an ellipse-shaped aneurysm.^150,151 Viewing of coronal or sagittal images mitigates but does not eliminate this problem.

Reproducibility and Discrepancies

Although most of the literature and guidelines on imaging-based measurements of the aorta focus on the aortic root and thoracic or abdominal aortic aneurysms, several technical limitations and uncertainties also apply to chronic aortic dissection.

Cardiac Phase and Electrocardiographic Gating

Although echocardiographic diameter measurements are typically obtained in middiastole,¹⁴³ there are no consistent recommendations for CT and MRI across multimodality guidelines with respect to the most appropriate cardiac phase for aortic assessment.¹⁴³ Both the 2010 American College of Cardiology Foundation/American Heart Association guidelines and the 2020 International Society for Magnetic Resonance in Medicine recommendations suggest the use of electrocardiographic gating triggered to end diastole for the assessment of the aortic root.^70,140

In chronic dissection, this is most relevant for patients with unrepaired root aneurysms. Aortic pulsation is typically small (1–2 mm) in patients with noncompliant aneurysmal aortas. The main reason for electrocardiographic gating is not to capture maximum dilatation (which would be during the short systolic phase) but to suppress the more pronounced translational motion of the aortic root and proximal ascending aorta attributable to ventricular contraction. Thus, diastolic reconstructions are favored for ECG-gated CT as well. Gating is a prerequisite for accurate aortic root and proximal ascending measurements but is not critical for the distal aorta.

Inclusion or Exclusion of the Aortic Wall

There is currently no clear consensus on whether the aortic wall should be included in aortic diameter measurements, although the difference may be several millimeters, particularly if there is atherosclerotic disease or mural thrombus.^142,152

For TTE measurements of the aortic root, the 2010 American College of Cardiology Foundation/American Heart Association guidelines recommend using the internal diameter.⁷⁰ The American Society of Echocardiography recommends measurement of the aortic root from the leading edge of the anterior wall to the leading edge of the posterior wall (leading edge to leading edge) on a parasternal long-axis view, thus including half of the aortic wall.¹⁵³ Most of the normative data on aortic size in the literature were obtained with this technique, and this convention is used clinically in many centers.^154–156 A 2015 multisociety expert consensus statement recommends maintaining the use of the leading edge–to–leading edge technique because of the greater reproducibility of this approach.¹⁴³

For CT and MRI, there also is no clear consensus.¹⁵⁷ The 2010 American College of Cardiology Foundation/American Heart Association guidelines recommend measurement of aortic diameters between external surfaces on CT and MRI (outer wall to outer wall) given that lumen size may not reflect the true size of the aorta in the presence of intra-aortic thrombus, atheroma, aortic wall inflammation, or dissection.^70,158 The outer wall technique also can be applied to noncontrast CT. In clinical practice, however, and in most research studies, the internal cross-sectional diameter is reported.^152,159 The 2020 International Society for Magnetic Resonance in Medicine guidelines also recommend inner lumen–to–inner lumen measurements of the aorta on MRI.¹⁴⁰

Patients with aortic dissection have 2 flow channels. The outer wall of the true lumen usually has normal wall thickness (≈1 mm), which is simply too thin to be resolved as a separate structure even on a high-quality CTA (≈1-mm³ voxel dimensions). The distinction between inner and outer contours is thus moot. The outer wall of the false lumen is initially even thinner and cannot be resolved either until it becomes >1 mm thick. The high-contrast interface between lumen and surrounding tissue is therefore a reasonable representation of diameter and can be detected automatically. Once mural thrombus occurs and obscures the interface between the (now thrombosed) lumen and the inner border of the aortic wall, it becomes necessary to use the outer aortic contour as a landmark. The resulting “error” of including 1 of the 2 walls in the total measurement is probably small, 1 to 2 mm.

A final example of inconsistency with respect to inclusion versus exclusion of the aortic wall can be found in TEVAR device sizing guidelines. Cook Medical (Bloomington, IN) bases its recommendations for the Zenith Alpha Thoracic Endovascular Graft on outer wall–to–outer wall dimensions,¹⁶⁰ whereas W.L. Gore (Flagstaff, AZ) recommends sizing of its Conformable TAG Thoracic Endoprosthesis on the flow lumen (including thrombus), not the vessel wall.¹⁶¹

Differences Within and Between Imaging Modalities

Most studies comparing 2-dimensional TTE with CT and MRI found that TTE significantly underestimates aortic root diameters compared with CT^162–164 and MRI,^163–166 likely related to off-axis image acquisition and root asymmetry. On the other hand, 3D TTE correlates well with CT.¹⁶⁵ Several studies also have demonstrated that abdominal ultrasound systematically underestimates the size of abdominal aortic aneurysms by ≈1 to 3 mm compared with CT.^167–171

In the nondissected thoracic and abdominal aorta, the agreement between diameter measurements obtained with CTA and MRA is good when the same analysis technique is used,^168–171 and interobserver agreement is excellent.^163,164,172 No such data are available for chronically dissected distal aortic segments, however. It is reasonable to assume that agreement is less, perhaps more similar to the interobserver and intraobserver variability of CT for abdominal aortic aneurysms, reported at ≈5 and 3 mm, respectively.^168–170

Implications for Serial Measurements

There is broad consensus that an identical imaging technique be used for serial aortic imaging.^37,70,143 Given the degree of variability in measurements, a recent expert consensus statement recommended caution in interpreting apparent changes in aortic diameter ≤3 mm by ECG-gated CT and ≤5 mm by nongated CT.¹⁴³ Aortic growth estimates based on small apparent diameter increases over short time intervals (≤6 months) also should be treated with caution given the potential implication for surgical repair.

This represents a considerable degree of uncertainty given the clinical implications for patients with chronic aortic dissection. It is currently unknown whether including details of the measurement technique (eg, gating, inclusion of wall, orientation of measurement plane), as proposed by Freeman et al,¹⁵² in the radiology report improves interpretation of imaging results by the treating physician. Selecting identical anatomy and measurement planes may be the most important steps to reduce overall measurement discrepancies in patients with chronic dissections and aneurysms; however, this has not been formally studied.

False Lumen Measurements

Separate measurements of true and false lumen diameter, area, or outer aortic wall circumference have been used in research but are not routinely obtained in patients with aortic dissection.^59,119 Diameter and area measurements (but not circumference) are affected by the deformation of the dissection flap over the cardiac cycle.^30,59 A proximal descending aorta false lumen diameter >2.2 cm has been associated with late complications in patients with uncomplicated type B aortic dissections.¹⁷³ Volumetric measurement techniques may be more reliable for documenting small changes over time and have been used to define positive and negative remodeling in device trials.^96,174

SURVEILLANCE IMAGING

Providing technically accurate aortic measurements in patients with chronic aortic dissections is a prerequisite but is not nearly enough for adequate surveillance imaging, for which a large number of patients with numerous follow-up studies are routinely encountered.

Comparisons With Priors, Including Older and Outside Imaging

In patients with chronic aortic dissection, it is imperative that each imaging study is carefully compared with priors, including studies done at other facilities.^175,176 Ideally, all prior studies should be reviewed before a therapeutic decision is made.³⁷ At a minimum, the most recent prior study and the earliest (baseline) study should be compared with the current study.

Serial Measurements and Comparisons

Serial measurements should always be obtained at the same anatomic location and identical orientation along the aorta, particularly when the diameters approach the intervention threshold. This requires direct side-by-side comparisons of original image data sets, typically done on a 3D workstation, and can be a time-consuming task in patients with elongated and tortuous dissected aortas (Figure 6). It cannot be overemphasized that comparing independently reported measurements without reviewing the actual images is highly unreliable and can lead to spurious aneurysm growth or missing significant enlargement, with potentially serious consequences.

Figure 6. Serial imaging: spatial alignment of orthogonal measurement planes.

Baseline (left) and 3-month, 6-month, and 12-month (right) follow-up computed tomography scan of a patients with type A dissection treated with ascending aortic repair. Accurate growth measurement requires double-oblique images to be obtained at the same location and orientation along the aorta. A indicates anterior; F, feet; H, head; L, left; P, posterior; and R, right.

None of the above requirements are easily accommodated within a busy imaging service. First, it is necessary to clearly identify the patients for whom elaborate measurements are requested. The full gamut of necessary measurements and documentation is often too time consuming to be routinely carried out by the interpreting physician, notably if similar measurements also need to be performed on prior images at the time of interpretation. Prior images are not always easily available, and often no mechanism to review, analyze and report outside imaging studies exists.

These predicaments can be overcome by building a dedicated surveillance program made up of agreed-on measurement, comparison, and reporting protocols at its core. Specific recommendations for structuring and implementing a surveillance program are currently not available, and although there is little doubt that better standardization has benefits, we do not have any evidence that implementation of a surveillance program improves outcomes. The initial investment of time, effort, and resources to build the logistical framework is considerable and may be warranted only for aortic centers specialized in the lifelong care for patients with chronic dissection and other aortic diseases.

Imaging Surveillance Program

At the core of any dedicated aortic imaging surveillance program is the standardization of the measurement protocol and the development of a mechanism how to perform, store, and retrieve comparison measurements (including for outside studies). This requires that a patient’s measurements are obtained at 1 point in time and are available at all subsequent visits. This way, a patient’s surveillance record only needs to be updated with the latest, carefully obtained side-by-side measurements, except if a patient enters the system at a later stage or at the time of program initiation. Reporting standards also need to be agreed on, including how the pertinent information can be incorporated into or added to the radiological report and the patient’s medical record. Additional logistics related to prioritizing measurements for clinic visits, ordering and billing of measurements on outside studies, and quality assurance need to be developed in close collaboration between radiologists, cardiothoracic and vascular surgeons, cardiologists, and vascular specialists and their respective teams. Last, close interaction between clinical and postprocessing teams to reconcile any discrepancies and to improve the system is important for maintaining a vital service.

The execution of the core tasks is best implemented in a 3D laboratory or equivalent where trained technologists perform measurements with consistent quality. A sample measurement protocol is provided in Supplemental Figure 4. Most large institutions housing an aortic center most likely also have a 3D laboratory or at least the functionality of such a laboratory in their imaging departments.

Reporting and Visualization

The key information, that is, the validated diameter measurements compared with corresponding prior measurements, is best reported in a table rather than in text format augmented by a graphical display that better visualizes the information for physicians, staff, and patients. A simple example is shown in Figure 7, in which aortic diameter measurements at different time points are listed in a table (a simple spreadsheet suffices) and plotted graphically.

Figure 7. Surveillance imaging measurement report.

A, For each baseline and follow-up imaging study, orthogonal aortic diameter measurements are entered into a spreadsheet or database. Note that in this patient the maximum diameter is approximately halfway between 2 standard aortic landmarks, the proximal (Pdes) and distal (Ddes) descending thoracic aorta. B, Diameters (y axis) are also plotted graphically against anatomic landmarks along the aorta (x axis). Time is color coded, with the baseline study plotted in green, the second-to-last study in blue, and the latest study in red.

An infinite range of alternative methods can be used to visualize the relevant information, including interactive tools switching between simple diameter plots and displays emphasizing the time axis.¹⁷⁷ Several symbols or cartoon representations for differently treatment segments (surgical grafts, endografts) also can be displayed.¹⁷⁸ Currently, these techniques are custom made and developed out of necessity at academic institutions, but similar to other applications such as planning for abdominal endovascular aortic repair or transcatheter aortic valve replacement, commercial tools are likely to become available eventually.

RISK STRATIFICATION AND NEW DEVELOPMENTS

After decades of management choices limited to watchful waiting followed by elective open repair in the chronic phase, new treatment options during the acute or subacute phase have become available that may reduce the risk of late complications in patients with aortic dissection.^95,179,180 Patient selection and choice of the best therapy remain a challenge, however. For instance, identifying patients who are most likely to benefit from preventive TEVAR in type B dissection requires the ability to predict the time course of the false lumen dilatation. Likewise, considering alternative treatment options requires the ability to forecast the effect of each intervention on the fate of a patient’s aorta.

Several attempts to improve risk stratification in patients with aortic dissection have been made over the past years. A major focus of interest and ongoing matter of debate is risk stratification of patients with uncomplicated type B dissection. Initial strides toward predicting and simulating the effects of interventions in patients with aortic dissection have also been reported.

Predicting Late Adverse Events in Uncomplicated Type B Dissections

Patients with initially uncomplicated type B aortic dissection have excellent early survival but a 20% to 50% risk of late adverse events.^55,56,181 Although early TEVAR may prevent or at least delay late complications, only a subset of high-risk patients are likely to benefit from the procedure, which is not without its own risks.^182–184 Several studies have sought to identify morphological predictors of aneurysm growth and late adverse events to better risk-stratify these patients.

In a systematic review of 51 original studies, Spinelli and colleagues⁵⁹ analyzed the consistency of reported morphological predictors of adverse events in patients with uncomplicated type B dissections. Of the wide range of published morphological predictors, only 2 had a consistent effect across the literature: (1) Aortic diameter at presentation was consistently associated with adverse events, and (2) complete false lumen thrombosis always had a protective role. Inconsistent and even conflicting results were reported for other investigated predictors, including false lumen diameter and length; false lumen location (lesser or greater curvature of the aorta); number, size, and location of the primary entry tear; branch vessel involvement; and partial false lumen thrombosis.

The reported inconsistency of morphological features to predict adverse events might be explained by the fact that candidate predictors were analyzed as isolated, unconnected features, not in the context of our evolving understanding of the pathophysiology of aortic dissection in which features are hemodynamically linked. For instance, the effect of the size of the primary entry tear (allowing false lumen inflow) cannot be interpreted meaningfully without knowledge of the false lumen drainage (outflow resistance). Only 1 small study of 83 patients with uncomplicated type B dissection combined 4 independent morphological and 1 clinical predictor into a single prediction model for late adverse events, and 1 of the independent predictors was false lumen drainage (outflow).³⁰ External validation and more comprehensive analyses of aggregated dissection features in larger cohorts are needed. A prerequisite is the development of automated tools to extract the potentially important features because manual segmentation is prohibitively time consuming and unreliable.¹⁸⁵

Last, it is important to recognize that some of the morphological predictors such as maximum diameter, the relative proportion of the outer aortic wall circumference encompassing the false lumen, and the occurrence of partial false lumen thrombus change over time. Those changes over time (eg, growth) on serial CT imaging have predictive power of their own.¹⁸⁶ Hence, an ideal prediction model for patients with uncomplicated type B dissection not only would predict the risk at baseline (eg, at time of hospital discharge) but also would allow the risk estimate to be updated with new clinical and morphological information at every follow-up encounter. Currently, such models do not exist.

Predicting the Fate of the False Lumen After Surgical or Endovascular Intervention

In experienced hands, surgical repair of the ascending aorta can be safely extended into the aortic arch or beyond with placement of a conventional or “frozen” elephant trunk into the descending thoracic aorta.^179,180 Hybrid and endovascular treatment options are also evolving. Ascending repair can be combined with antegrade placement of a descending aortic endograft¹⁸⁷ or with a dedicated partially uncovered aortic arch hybrid graft (Ascyrus Medical Dissection Stent) at the time of operation.¹⁸⁸ Covered endografts also can be combined with uncovered devices, for example, in the provisional extension to induce complete attachment approach,¹⁸⁹ in acute type B aortic dissection.

Although the long-term effect of these interventions will not be known for many years, imaging and simulations can help anticipate early markers of success, notably the induction of false lumen thrombosis. Computer simulations developed at the Imperial College London have shown that the location of thrombus formation can be predicted from patient-specific dissection morphology before and after placement of, for example, an endograft.^190,191

New Developments: Machine Learning

The potential of using the massive amounts of accumulated radiological data to predict rupture and to determine the best timing for intervention was recognized many years ago.⁷⁰ Except for maximum aortic diameter, however, the importance of many candidate morphological features remains uncertain, and manual annotation (segmentation) of imaging features in patients with aortic dissection is difficult and prohibitively time consuming for clinical application.

New machine-learning algorithms may overcome these limitations. Several dissection features such as the true and false lumen and diameter measurements have been successfully identified with deep-learning algorithms.^185,192 Aortic landmarks have been detected with a range of techniques, including deep-reinforcement learning.¹⁹³ Random forest classification methods have been used to investigate the risk of developing type A dissection according to CT and clinical data.¹⁹⁴ Aneurysm growth also has been predicted with encouraging accuracy in abdominal aortic aneurysms with a set of benchmark learning techniques.¹⁹⁵

Although immensely promising, these initial results from rapidly evolving machine-learning applications for segmentation, classification, and risk prediction in aortic dissection are still a far cry from clinical-grade, robust applications. Validating these new tools with large enough well-curated and expert-annotated imaging databases with clinical outcomes is still required but difficult to accomplish in patients with a relatively rare, anatomically complex chronic disease.

SUMMARY

Chronic aortic dissection is an underrecognized cardiovascular disease and growing health care burden. Imaging and image processing will continue to play a pivotal role in surveillance and clinical decision-making in these patients, as long as accurate and standardized measurements of aortic dimensions over time are provided.

There is a great need for additional research into the epidemiology and natural history, pathophysiology, management, and prediction of outcomes in patients with chronic aortic dissection to support the ongoing paradigm shift from the prevailing watch-and-wait treatment strategy to an individualized selection of traditional or new surgical, endovascular, or hybrid interventions in the future. Important questions identified by the writing group, intended to serve as an impetus for additional research, are listed in the Table. Given the uncommon nature, unspecified economic impact, and long follow-up needed to study chronic aortic dissection, funding for research is challenging, and significant advances will undoubtedly require interdisciplinary collaboration across a large network of research and clinical centers. High-resolution imaging data, combined with computer simulations, biomechanical modeling, and machine learning, may help answer many of the remaining questions and ultimately improve the lives of patients with chronic aortic dissection.

Table.

Critical Knowledge Gaps and Research Needs in Imaging and Surveillance of Chronic Aortic Dissection

Epidemiology and economic impact

The actual prevalence of patients living with chronic aortic dissection is currently unknown.

The cumulative lifetime expenses for long-term specialty care and interventions in patients with chronic dissection have not been adequately appraised.

Natural history and pathophysiology

The natural history and the effect of evolving therapies on outcomes in patients with chronic dissection are incompletely established. Prospective, multicenter registries with concurrent recording of clinical and imaging data are needed.

Sex-and racial differences in false lumen growth over time and in long-term outcomes and prognosis have not been investigated in patients with chronic dissection, despite known differences in incidence age, clinical presentation, and short-term outcomes in women.

The hemodynamics and mechanobiology of the dissected aortic wall and their interplay with progression and complications are incompletely understood.

The biological, cellular, and molecular processes involved in the repair, progression, and remodeling of affected arteries need to be further investigated.

Clinical management

New tools and methods for modeling treatment effects, eg, how different surgical or endovascular techniques affect blood flow, false lumen pressure, and the chance of false lumen thrombus formation, are needed.

Adequate data to establish evidence-based standardization and personalization of surveillance intervals are currently lacking.

Imaging techniques and measurements

Better tools to facilitate the current manual and semiautomated diameter measurements, including comparison measurements and documentation, are urgently needed.

New machine-learning tools for fully automated or semiautomated extraction of prognostically relevant features beyond diameters are required.

Risk prediction and modeling

Validated prediction models for risk stratification and guidance of treatment at the time of the acute event are highly desirable. Ideally, patient risk can be updated at each follow-up interval.

The best combination of morphological, functional, and clinical parameters for predicting adverse events has not yet been determined.

Projecting geometric changes over time is necessary to foresee potential changes in anatomic suitability for endografting over time.

Supplementary Material

Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/HCI.0000000000000075.