Endovascular Repair of Descending Thoracic Aorta ... - TeraRecon

Endovascular Repair of Descending Thoracic Aorta AneurysmRoy K. Greenberg, MDDirector of Endovascular ResearchDepartments of Vascular Surgery,CardiothoracicSurgery and Biomedical EngineeringKate Pfaff, BSCorresponding author:Roy K. Greenberg, MDDepartment of Vascular SurgeryThe Cleveland Clinic9500 Euclid Ave.Cleveland, Ohio 44195Email: greenbr@ccf.orgProgram Manager, Biomedical EngineeringCleveland Clinic, Cleveland, OhioIntroductionAdvances in image acquisition and reconstructiontechnology have paralleled the developmentof endovascular devices. Of particularinterest is the application of endovascular grafts tothe treatment of aortic aneurysms. Prior publicationshave defined the feasibility of the use of customizeddevices to treat aneurysms abutting or involving thevisceral vessels. To discuss this technology further,we will describe the use of sophisticated imagingtechniques to plan procedures to treat aneurysms ofthe descending thoracic aorta. Although this regiondoes not have branch vessels of large diameter, theinherent tortuosity, larger aortic diameter, and variedetiology of diseases underscores the need for comprehensiveevaluation of the aortic pathology combinedwith careful procedure planning and device sizing.The follow-up of patients treated with thoracic endovasculargrafts is also critical and requires preciseimage assessment using similar techniques.Classification of Thoracic Aortic DiseasesThe descending thoracic aorta (DTA) extends fromthe left subclavian artery to the level of the aortichiatus in the diaphragm, although many consider theportion of the abdominal aorta proximal to the celiacartery origin also to be part of the descendingthoracic aorta. Aneurysms are defined as an aortictransverse diameter exceeding twice the normal surroundingaortic diameter. Although aneurysms aredefined on the basis of diameter measurements, thereis considerable lengthening of the diseased segmentas well. Most aneurysms are considered non-specificin nature, implying that we do not fully understandthe pathogenesis of this disease. Several theories exist,and experts cite many causative factors includinggenetic predisposition, atherosclerosis, hypertension,remote infections, and smoking. However, notall aneurysms are non-specific in nature. Aneurysmscan be saccular (Figure 1) or fusiform (Figure 2) innature, the latter being more diffuse, frequently affectingthe entire DTA.Endovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS123

Figure 1: Saccular aneurysm of the thoracic aortaFigure 2: Fusiform aneurysm of the thoracic aortaDissections are the second most common cause ofthoracic aortic dilation (Figure 3). These occur whenthere is a tear in the intima of the aorta extendinginto the media (typically just distal to the left subclavianartery). This split in the aortic wall tends topropagate distally, frequently involving the abdominalaorta, the visceral and iliac vessels. The observedflap within an aortic lumen is characteristic of a dissection,which is managed differently than a nonspecificaneurysm (Figure 3).Other less common conditions exist that are frequentlyconfused with aortic dissections. These includeintramural hematomas and ruptured aortic plaques,but are pathologically distinct entities. Additionalpathologies that afflict the thoracic aorta must alsobe considered, including inflammatory diseases, autoimmunediseases, sarcoma and trauma. Therefore,it is important to define the etiology of the aorticdisease prior to treatment. This is best accomplishedusing a variety of image reconstruction techniquesthat are common to the Aquarius Workstation. Forthe purpose of this article, we will limit most of thediscussion to non-specific aneurysms.Figure 3: Green arrow depicts Type B thoracic aortadissection, beginning distal to the left subclavian arteryand continuing to abdominal aorta.124 SECTION 3 Surgical and Endovascular Procedure Planning

AnatomyThe descending thoracic aorta is bounded on oneside by the brachiocephalic vessels (innominate, leftcarotid and left subclavian arteries) (Figure 4), andthe abdominal visceral segment distally. Most aneurysmsof the DTA involve the proximal segment,near the brachiocephalic vessels. This region is tortuousand has a diameter typically between 18 and30 mm. Anatomic anomalies exist, such as aberrantright subclavian arteries, vertebral arteries arising directlyfrom the aorta, and more complex situationssuch as nutcracker aortas (where the aorta is splitand traverses anterior and posterior to the esophagus)that are beyond the scope of this summary, butmust be defined prior to the endovascular treatmentof aneurysmal disease.Endovascular devices should be fixated and sealedwithin healthy aorta. Thus, the importance of assessingthe regions immediately proximal and distalto the aneurysmal segment (the neck) is critical.Healthy aorta is typically cylindrical, free of debrisalong the wall, and not dilated. Endovascular devicestypically require 2-3 cm of non-aneurysmal proximaland distal aortic neck. The diameter of the nativeaorta is measured such that the implant devicediameter can be calculated by oversizing the aortaby 15-20%. Determining the length of the aorta tobe covered using the center line of flow is necessaryto select the length and number of components to beimplanted. Most devices are have a maximum lengthof about 20 centimeters; thus if the aneurysm properis longer than 15 centimeters, consideration must begiven to using multiple devices. Of particular interestis the morphology of the sealing and fixation zones.This has implications as to how the proximal and distalstents will reside along the wall. The overall tortuosityof the aorta will determine whether a deliverysystem can be readily advanced (extreme tortuositywill preclude most device deliveries), and the morphologyof the iliac access vessels will predict wheredevices may be introduced into the arterial system(femoral versus iliac approach).Figure 4: Vessels of Aortic ArchCAVEAT: Preoperative CT data acquisition must includecomprehensive imaging of the thoracic aorta,brachiocephalic vessels, and abdominopelvic vasculature.Precontrast images should be obtained tohighlight regions of calcification. High-resolutionstudies are optimal with properly timed contrastboluses allowing opacification of the aorta and theproximal portion of all branches. Data should be reconstructedwith a smoothing algorithm.Endovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS125

Figure 5: Evaluating vasculature by slicing volumetric datathrough different planes on the Aquarius WorkstationConsiderations to Planning and SizingThe complexity of defining limits of healthy tissue inthe thoracic aorta is similar to challenges in the abdominalaortic neck with additional considerations.In a population where prior aortic aneurysm repairis not uncommon, concerns of postoperative neuromusculardeficit are very real. Aortic tortuosity precludesthe use of thin slice axial CT images for theassessment of the region to be treated. Cross sectionalimages obtained from high-resolution spiral CTscans and reconstructed on a 3D workstation allowclinicians to evaluate these images using MPRs orthin MIPs. “Slicing through” data in axial, coronaland sagittal planes demonstrates the relationship betweenvascular and other structures, improving theevaluation of a patient’s disease process compared toevaluating axial images alone (Figure 5).126 SECTION 3 Surgical and Endovascular Procedure Planning

Figure 6: 3D reconstruction demonstrates conicalneck distal to curvature in aorta; therefore, the entireproximal fixation (2 – 3cm) must be planned to lieproximal to green arrow.Figure 7A: Grayscale volume-rendered image showingdistal conical neck. The distal fixation in this patientshould be planned to lie distal to the green arrow, therebyinvolving the visceral arteries in the repair.Given that the aorta should fundamentally decreasein diameter as blood travels distally from the leftventricle, unhealthy aorta can be simply defined asany region of increased diameter. Regions that appearconical on a straightened view are prototypicalof aortic pathology that is frequently missed whentwo-dimensional studies are used in isolation to assessthe disease (Figures 6 and 7).Figure 7B: MIP post stentgraft. Red arrow shows distalleak. Yellow arrow identifies component of stentgraftabutting the celiac artery.Figure 7C: MIP after placement of a distal extension witha scallop to maintain distal patency of the celiac artery(yellow arrow).Endovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS127

Figure 8 Figure 9 Figure 10Figures 8-10: Using the center line of flow (CLF) tool to create a straightened image of the aortaThus, the limits of coverage must be defined usingCLF (center line of flow) techniques, during whichaccurate diameter measurements of the regions canbe obtained. Using the vessel analysis tools on the3D Aquarius Workstation, a center line of flow iscreated by placing “seed points” at the proximal anddistal areas of interest (Figure 8). The informationthen generates a straightened image of the aorta (Figures9 and 10). This vessel analysis tool additionallyenables the physician the option of manipulating thecenter line of flow in order to simulate the path towhich the stentgraft may conform. Orthogonal slicesthrough the center line of flow provide measurementsthat are then used for endograft sizing (Figure 11).Figure 11: Measurements on the straightened aorta for endograft sizing128 SECTION 3 Surgical and Endovascular Procedure Planning

Tapered AortasUnder normal circumstances, the diameters of theproximal and distal DTA are not markedly different.However, it is not uncommon to have marked differencesin the setting of aortic pathologies. In this situation,when the proximal and distal DTA diametersdiffer by more than 5 mm, using a tapered deviceshould be considered. Such a device is designed tobe 15-20% oversized at the region of the proximalseal and distal seal, rather than compromising the designon one or the other seals. This occurs frequentlywhen a patient has had prior surgery that involves theplacement of a graft within the proximal (such as anelephant trunk graft) or distal DTA. Currently, thereare no commercially available tapered devices totreat thoracic aortic pathology in the United States;however, both the Zenith (Cook, Inc.) and Talent(Medtronic) are accumulating data on such designs.The Proximal Landing ZoneIt is critical that thoracic devices remain fixed in positionand continue to exclude the aneurysm for theduration of the patient’s life. Thus, the proximal sealingzone within the aorta must be carefully chosen.This decision is based primarily on aortic diameter,angulation, tortuosity, calcification, and branches.Aneurysms involving the proximal DTA can be particularlychallenging with respect to the proximallanding zone. The intended location of the sealingis most frequently near the left subclavian artery,which is more properly considered the distal aorticarch. This region has inherent tortuosity, forcing thestent to lie upon a lesser (inferior) arch curvature, andappose the greater curvature (superior aspect) of thearch along which the origins of the brachiocephalicvessels arise. This is a treacherous region for a stentwith any rigidity. In fact, almost all of the deviceswhen deployed will tend to project along the lessercurvature into the aortic lumen. Although there arepotential device designs that will prevent this, themost readily available solution for the interventionalistis to deploy the device in a non-tortuous region.That means deploying distal to the curvature if theproximal DTA is healthy. However, if the proximalDTA is diseased, then a deployment that will coverthe origin of the subclavian artery must be considered.This decision should be made during the designand planning phase of the procedure. Regardless ofthe landing zone location, the aortic diameter (outerwall to outer wall) should be assessed orthogonal tothe center line of flow, and the device oversized by15-20%. The tortuosity of the arch and status of theproximal DTA should be accurately assessed fromthe preoperative images to establish an operativeplan. Failure to do so will relegate the interventionalistto undesirable outcomes including more complicatedprocedures, the frequent use of proximalextensions, excessive manipulation or ballooning inthe arch, and longer procedure times.The Distal Landing ZoneIn most circumstances, the distal landing zone issimpler to deal with than the proximal landing zone.However, it is important not to underestimate thepotential for tortuosity within the distal aorta, particularlyas the aorta traverses the left chest to resideupon the anterior aspect of the spine in the regionof the diaphragmatic hiatus. Many of the same concernsexist for this region. The operator must delineatethe distal margin of the aneurysmal disease, andmeasure the length and diameter of the distal fixationand sealing zone. The distal device diameter is oversizedby 15-20% in a manner similar to the proximaldevice diameter calculation.Length of Aortic CoverageThe distance between the proximal and distal landingzones delineates the required length of aorticcoverage. Although it may be appealing to simplycover the entire DTA, one must be cautious of therisk of paraplegia, which has been linked to the extentof aortic tissue lined with the endovascular graft.In general, if the amount of coverage is 12 cm orless, a single piece design can be considered. Theactual length is a simple calculation based upon theCLF, and most devices have prefabricated deviceswith lengths in increments of 2-5 cm that will suffice.However, most commonly the diseased segmentof aorta is in excess of 12 cm and will require twoor more pieces to achieve an optimal result. In thesecircumstances, there will be an overlap region. Theimportance of a proper design for this region cannotbe overstated. Morphologic changes of the DTAand the endograft over time will result in componentseparation and sac repressurization if an adequateoverlap between components is not planned.Endovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS129

The overlap region between the devices should beplanned such that it resides in a relatively straightsegment if possible and within the aneurysm proper.A simple way to design devices requiring two componentsis to plan the first (proximal) piece so that itextends from the origin of the proximal landing zoneto just above the distal margin of the aneurysm. Thesecond (distal) component extends from the origin ofthe aneurysm through the distal landing zone. Thus,the entire aneurysm will then represent the region ofoverlap. Yet, given the sometimes excessive length ofthe DTA, a third piece may be required. This pieceshould be used to supplement the overlap zone ratherthan constitute the proximal or distal landing zone(Figure 12).Figure 13: 3D reconstruction of coarctation patient. Whitearrow is area of aortic stricture. The red arrow showssignificant collateral flow.Figure 12: Three-piece stentgraft design (red arrowsshow areas of overlap)Miscellaneous PathologiesAneurysms represent the most common manifestationof DTA disease; however, proper assessment ofimaging data will allow for the recognition of otherpathology. Aortic coarctation (Figures 13 and 14) anddissection are treated using similar techniques withrespect to image manipulation, planning and sizing.Figure 14: 3D reconstruction of same patient postangioplastyand endovascular stentgraft placement distalto the subclavian artery.130 SECTION 3 Surgical and Endovascular Procedure Planning

Follow-up StudiesThe methods for analyzing migration in a patientwith an abdominal aortic device are insufficient whenapplied to a thoracic aortic device. A 3D scan isneeded to make an accurate observation of devicestability and device migration within the thoracicaorta since a significant portion of the aorta is orientedparallel to axial images as indicated by the redarrow (Figure 15).reference vessel (the celiac artery) and the distal aspectof the stentgraft is determined at various timepoints. Movement of the device in relation to thesereference vessels warrants further investigation. Inthis case, and in cases of aortic lengthening, attentionto the three-dimensional reconstructed image isrequired. Calcification patterns within the aortic wallthat exist in close proximity to the implanted deviceare used as reference points. If the device positionhas changed relative to nearby aortic wall patterns,migration must have occurred. However, all of thismust be considered in the context of the image resolutionat each stage. Not infrequently, careful imageassessment of the aortic wall will yield concrete evidencethat migration has occurred, as there exists animprint of the initial device position in addition to atraceable path of slippage.Figure 15: Volumetric image of the thoracic aorta. Redarrow indicates portion oriented parallel to axial images.The aorta will frequently remodel during the life ofthe patient following the placement of an endovasculargraft within the DTA. Consequently, the assessmentof device stability is complicated followingendovascular repair. Historically, axial image tablepositions were used to determine migration of stentgraftswithin the aorta. However, a detailed analysisrecently published by O’Neill et al. illustratesthe complexity of the problem of follow-up andclearly mandates careful image manipulation using3D tools. In his study, the CLF distance measurementbetween the left common carotid and the celiacartery was used to assess aortic lengthening. In theabsence of aortic lengthening, the distance betweena reference vessel (such as the left common carotidorigin) and the proximal aspect of the graft is establishedshortly after implantation and during each follow-upvisit. Similarly, the distance between anotherEndovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS131

Figure 16A: The blue line indicates area of overlapin a stentgraft with suspected barb fracture in twolocations. Closer inspection using this technologywith stent specific template clearly identifiesa fracture of both a “z” stent (red arrow) and alongitudinal nitenol wire (red circle).Figure 16B: Enlarging aneurysm of patient with stentgraftfracture.Device IntegrityIn addition to the relationship between the deviceand the aortic wall, the integrity of the device, anddevice intercomponent relationships can be readilyestablished using specific reconstruction techniques.Data reconstruction using an edge detection algorithm,rather than a smoothing algorithm, illustratesthe ability to detect fractures of the metallic devicecomponents (Figure 16 A, B).132 SECTION 3 Surgical and Endovascular Procedure Planning

Figure 17: Fixed surgical landmarks are surgical clips (red arrows) and pacing wire (denoted by white arrows)sutured to the distal end of previously placed elephant trunk. Figure on left is at time zero. Comparisons are made atfollow-up with respect to migration. This example is evidence that the same amount of overlap of stentgraft is in theelephant trunk at one year.The degree of overlap between components can be accuratelymeasured, and the relationship between thedevice and fixed surgical landmarks can also help toassess device movement during follow-up (Figure 17).Endoleak is also assessed during follow-up using thevascular side-by-side module of the Aquarius Workstation.A non-contrast and contrast CT scan arecompared to determine if a true endoleak is present,and its source (Figure 18).Figure 18: Endoleak confirmed comparing non-contrast scan on left with contrastscan on right at same table position.Endovascular Repair of Descending Thoracic Aorta Aneurysm Roy K. Greenberg, MD and Kate Pfaff, BS133

ConclusionEndovascular technology is expected to benefit thetreatment of aneurysmal disease in the chest morethan in the abdomen because of the higher morbidityof open thoracic aortic procedures. However, theassessment of long-term durability of the endovasculardescending thoracic aorta repair will requirecompliancy in patient follow-up and sophisticatedimaging techniques.References1. Greenberg. RL Endovascular aneurysm repair using branchedor fenestrated devices. Clinical Case Studies in the Third-Dimension, 2005, TeraRecon, pp. 205-2142. Svensson LG, Crawford ES. Cardiovascular and Vascular Diseaseof the Aorta. 1997 W.B. Saunders Company, p.29.3. Greenberg, RK. Abdominal aortic endografting: Fixation andsealing. J of American College of Surgeons, Vol 194 Jan. 20024. Kasirajan, K. Milner R, Chaikof E. Late complications of thoracicendografts. J Vasc Surgery 2006; 43(A) 94A-99A5. Greenberg, RK, O’Neill, S, Walker, E et al. Endovascular repairof thoracic aortic lesions with the Zenith TX1 and TX2 thoracicgrafts: Intermediate-term results. J. Vasc Surgery, April 2005; Vol41. Issue 4 p. 589-5966. Gravereaux EC, Faries PL, Burks JA, et al. Risk of spinal cordischemia after endograft repair of thoracic aortic aneurysm. JVasc Surg 2001; 34; 997-10037. Ishimary S, Endografting of the aortic arch, J Endovasc Therapy2004: 118. O’Neill S, Greenberg RK, Resch T, et al. An evaluation of centerline of flow measurement techniques to assess migration afterthoracic endovascular aneurysm repair. J Vasc Surg. 2006 Jun;43(6) 1103-10.134 SECTION 3 Surgical and Endovascular Procedure Planning