- Email: [email protected]
Heartbeat-related displacement of the thoracic aorta in patients with chronic aortic dissection type B: Quantification by dynamic CTA
European Journal of Radiology 72 (2009) 483–488
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Heartbeat-related displacement of the thoracic aorta in patients with chronic aortic dissection type B: Quantification by dynamic CTA Tim F. Weber a,∗ , Maria-Katharina Ganten b,1 , Dittmar Böckler c,2 , Philipp Geisbüsch c,3 , Hans-Ulrich Kauczor a,4 , Hendrik von Tengg-Kobligk b,5 a
University of Heidelberg, Department of Diagnostic and Interventional Radiology, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany German Cancer Research Center, Department of Radiology, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany c University of Heidelberg, Department of Vascular and Endovascular Surgery, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany b
a r t i c l e
i n f o
Article history: Received 25 May 2008 Received in revised form 22 July 2008 Accepted 31 July 2008 Keywords: Computed tomography angiography Aorta Aortic dissection Endovascular aortic repair
a b s t r a c t Purpose: The purpose of this study was to characterize the heartbeat-related displacement of the thoracic aorta in patients with chronic aortic dissection type B (CADB). Materials and methods: Electrocardiogram-gated computed tomography angiography was performed during inspiratory breath-hold in 11 patients with CADB: Collimation 16 mm × 1 mm, pitch 0.2, slice thickness 1 mm, reconstruction increment 0.8 mm. Multiplanar reformations were taken for 20 equidistant time instances through both ascending (AAo) and descending aorta (true lumen, DAoT; false lumen, DAoF) and the vertex of the aortic arch (VA). In-plane vessel displacement was determined by region of interest analysis. Results: Mean displacement was 5.2 ± 1.7 mm (AAo), 1.6 ± 1.0 mm (VA), 0.9 ± 0.4 mm (DAoT), and 1.1 ± 0.4 mm (DAoF). This indicated a significant reduction of displacement from AAo to VA and DAoT (p < 0.05). The direction of displacement was anterior for AAo and cranial for VA. Conclusion: In CADB, the thoracic aorta undergoes a heartbeat-related displacement that exhibits an unbalanced distribution of magnitude and direction along the thoracic vessel course. Since consecutive traction forces on the aortic wall have to be assumed, these observations may have implications on pathogenesis of and treatment strategies for CADB. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction The development of aortic dissections is associated with arterial hypertension and degenerative or hereditary vascular disease [1,2]. Besides a loss of vessel wall elasticity that accompanies these conditions, aortic motion has been identified as a potential trigger for the onset of dissections within the ascending aorta (type
∗ Corresponding author. Tel.: +49 6221 38438; fax: +49 6221 565730. E-mail addresses: [email protected] (T.F. Weber), [email protected] (M.-K. Ganten), [email protected] (D. Böckler), [email protected] (P. Geisbüsch), [email protected] (H.-U. Kauczor), [email protected] (H. von Tengg-Kobligk). 1 Tel.: +49 6221 422493, fax: +49 6221 422462. 2 Tel.: +49 6221 566249; fax: +49 6221 565423. 3 Tel.: +49 6221 566249; fax: +49 6221 565423. 4 Tel.: +49 6221 56410; fax: +49 6221 565730. 5 Tel.: +49 6221 422492; fax: +49 6221 422462. 0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.07.045
A, Stanford classification) [3,4]. Both of the phenomena – aortic motion and reduction of elasticity – lead to an increase of vessel wall shear stress, which is in turn considered to be the underlying biomechanical factor for intimal detachment and chronic aortic expansion [4,5]. Expanding chronic aortic dissections of the descending aorta (type B, Stanford classification, CADB) are increasingly treated with thoracic endovascular aortic repair (TEVAR) [6]. Although this minimal-invasive approach reduces periprocedural mortality in selected patients, there are several procedure-related challenges that correlate with both the individual’s aortic anatomy and the geometrical characteristics of available endografts [7,8]. TEVAR of CADB is hampered by the tortuosity of aortic arch and descending aorta, which may circumvent stable attachment of insufficiently flexible endografts at the proximal landing zone. The implanted endograft fabric is constantly exposed to significant radial forces that result from a heartbeat-associated cyclic distension of the aortic cross-section being observed in CADB as well [9]. In addition, continuous motion of the thoracic aorta following from cardiac pul-
484
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
sation is supposed to increase structural demands on design and material properties of thoracic aortic prostheses [8]. It has been demonstrated that the aortic root may undergo systolic displacement directed towards the left ventricular apex [10,11]. Recently, corresponding observations of aortic motion have been extended to the ascending thoracic aorta in patients with abdominal aortic aneurysms [12]. Aim of this study was to elucidate both the magnitude and direction of the heartbeat-related displacement of the ascending and descending thoracic aorta as well as of the vertex of the aortic arch in patients with CADB by using electrocardiogram-gated dynamic computed tomography angiography (CTA). 2. Materials and methods 2.1. Study patients From the database of a prospective study a subpopulation of 11 consecutive patients with CADB (9 men, 2 women; mean age ± standard deviation, 62.7 ± 12.0 years, mean age of dissection ± standard deviation, 22.9 ± 21.0 months) were enrolled in this evaluation [13]. For inclusion into this study, candidates had to match the following criteria: existence of chronic aortic dissection of the descending aorta with an intimal detachment arising immediately distally to the origin of the left subclavian artery and exclusively treated with antihypertensive pharmacotherapy. Patients with aortic dissection arising at other localizations – including (retrograde) dissections of the ascending aorta and dissections of the descending aorta located more distally than described above – as well as patients with preceded surgical or endovascular repair of the aorta were primarily excluded. Approval of the local ethics committee and written informed consent were granted before CTA exams. 2.2. Image acquisition CTA of the aorta was performed with a 16-row multislice scanner (Aquilion 16® ; Toshiba Medical Systems, Otawara, Japan). As the CT scan served as a regular follow-up examination, a dose optimized biphasic acquisition protocol was designed: A time-resolved dynamic data set of the thoracic aorta provided for scientific and clinical assessment, and static cross-sections covering abdominal aorta and iliofemoral branches completed diagnostic work-up [13–15]. In the first step the thoracic aorta was scanned during inspiratory breath-hold with the following parameters: collimation 16 mm × 1 mm, tube rotation time 0.4 s, pitch 0.2, tube voltage 120 kV, tube current 300 mA, field of view 320 mm. An electrocardiogram (ECG) was simultaneously recorded for retrospective ECG-gating during this period of acquisition. In the second step the abdominal aorta was examined conventionally after switching off the ECG-gating and setting the scan parameters as follows: collimation 16 mm × 1 mm, tube rotation time 0.5 s, pitch 0.94, tube voltage 120 kV, tube current 170 mA, field of view 320 mm. Images generated in the second step were not part of scientific evaluation. For enhanced vessel contrast each patient received 130 ml of a non-ionic iodinated contrast medium with an iodine content of 300 mg/ml at an injection rate of 4 ml/s (Iomeprol, Iomeron 300® ; Bracco International, Milan, Italy) through an 18-gauge catheter positioned in a right-sided antecubital vein. Bolus timing was achieved using an automated triggering technique with a threshold of 110 HU within a region of interest (ROI) placed in the ascending aorta.
2.3. Image reconstruction and post-processing Retrospective ECG-gating enabled multiphasic image reconstruction of the thoracic aorta at 20 equidistant time points of the R-R interval. Transverse source data were reconstructed with a section thickness of 1 mm and an overlapping increment of 0.8 mm. The 20 data sets of each patient were transferred to a postprocessing workstation equipped with dedicated medical imaging software that permitted editing of a fourdimensional scan volume (Aquarius Workstation® ; TeraRecon, San Mateo, USA). To assess aortic in-plane displacement multiplanar reformations (MPR) were taken in strictly transverse orientation through the ascending (half way between the aortic annulus and the origin of the brachiocephalic trunk, segment AAo) and descending aorta (10 cm distal from the origin of the left subclavian artery, segment DAo) as well as in strictly coronary orientation through the vertex of the aortic arch (segment VA) (Fig. 1). As segment DAo was located at the dissected aorta, separate evaluations were performed for true (DAoT) and false channels (DAoF). The vertex of the aortic arch was defined as the most cranial location of the curvature and was situated either proximal (n = 6) or distal (n = 5) from the origin of the left subclavian artery. The resulting MPR (n = 20 per location) represented both temporal and spatial positions of the aorta and were exported as primary DICOM data for further quantitative image evaluation. 2.4. Image evaluation MPR image series were loaded into a software application that provided quantitative visualization and analysis of medical images (MIPAV® ; Center for Information Technology, National Institutes of Health, Bethesda, USA). Vessel analysis was carried out in a consensus reading of two radiologists with 5 years and 2 years of experience in computed tomography, respectively. Using a semiautomatic segmentation algorithm and based upon contrast differences to the lumen boundary, a ROI was generated over the aortic contour on each time instance. Propagated and adapted to each time instance the software calculated the centroid of each ROI based upon its geometry. From the temporal alteration of the centroid coordinates total extent, temporal development and spatial direction of in-plane vessel displacement were determined relative to the first reconstruction interval (Fig. 2). Mean displacement, standard deviation and range were calculated from the maximum segmental in-plane displacement of each time series. For assessing the temporal development of aortic displacement mean displacement of each time instance was plotted against the reconstruction interval. 2.5. Statistical analysis For testing for significance of differences between segmental distributions of aortic displacement the Wilcoxon Signed Ranks test was used. A p-value smaller 0.05 was considered to indicate statistical significance. 2.6. Study limitations To provide a homogeneous study population and to facilitate comparable measurements, all individuals to be included had to feature similar disease morphologies, namely a CADB with an intimal flap originating immediately distally to the outlet of the left subclavian artery. Because of this, only a limited number of patients could be included into this study, and statistics had to be calculated from small case numbers.
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
485
Fig. 2. Quantification of the magnitude of in-plane displacement of the aortic cross-sectional centroid. Segmentation of the aortic lumen (segment AAo) provided coordinates of the luminal centroid at each time instance. The distance between those were calculated from the coordinates’ differences using the theorem of Pythagoras.
The magnitude of aortic displacement may depend on blood pressure and cardiac stroke volume. Although these patients were on best medical treatment including blood pressure control, it is possible that the mean values of aortic displacement documented here may be influenced by arterial hypertension potentially existent in some individuals. The presented evaluation procedure is able to reveal both magnitude and orientation of in-plane displacement of reformatted aortic cross-sections. However, the effect of a concomitant displacement directed through-plane and its influence on our measurements had to be disregarded. Although mere visualization of fourdimensional data sets is possible with available imaging software, quantifying aortic motion in all spatial directions simultaneously requires computational algorithms that enable not only segmentation of the aortic lumen but also identification and particularly tracking of specific anatomic landmarks. 3. Results Mean displacement (±standard deviation) of AAo was 5.2 ± 1.7 mm (range, 3.2–7.5 mm) and of VA 1.6 ± 1.0 mm (0.4–4.0 mm). While displacement of AAo was directed anteriorly, VA was displaced in cranial orientation (Figs. 2–4). The true channel DAoT was displaced by 0.9 ± 0.4 mm (0.4–1.5 mm) and the false channel DAoF by 1.1 ± 0.4 mm (0.6–1.9 mm). DAoT and DAoF did not reveal a significant direction of displacement. At all segments a phase of fast systolic displacement was followed by slow diastolic regression to the starting point (Fig. 5). Displacement was significantly greater at AAo compared to VA and both DAoT and DAoF (p < 0.005 each). VA was displaced significantly greater than DAoT (p = 0.023). No differences were found between VA and DAoF (p = 0.095) and between DAoT and DAoF (p = 0.250) (Fig. 6). Fig. 1. Representative thoracic aortic cross-sections adjusted by single-oblique MPR. To assure analysis of precise orientation of aortic displacement along (segment VA, B) or perpendicular to (segments AAo and DAo, A and C) the patient’s longitudinal axis, each target reformation was adjusted by single-oblique MPR keeping the imaging plane of the target orientation untouched.
4. Discussion This study reports on magnitude and direction of the heartbeatrelated displacement of the thoracic aorta in patients with CADB. Using a previously developed and validated ECG-gated CTA
486
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
Fig. 3. Visualization of the orientation of aortic motion at the mid-ascending aorta and the vertex of the aortic arch. While the ascending aorta underwent displacement predominantly directed anteriorly (A), the vertex of the aortic arch exhibited a slight cranial displacement (B).
protocol, a dynamic data set was acquired that provided multidimensional information on the spatial position of the thoracic aorta [13]. The cyclic displacement of the thoracic aorta is of clinical interest, because on the one hand pathogenesis of ascending aortic dissections has been previously associated with aortic motion and on the other hand CADB and other diseases of the thoracic aorta are increasingly referred to endovascular aortic repair [4,8]. TEVAR of CADB aims at sealing entry tears to initiate false lumen thrombosis and resolution of true lumen collapse [16,17]. The implanted endograft is thereafter exposed to continuous physical strains that result from pulsatile blood flow and aortic motion. For this reason, life-long surveillance is necessary after performing endovascular procedures to timely detect endograft failure, e.g., migration of the prosthesis or wire fracture. In CADB, endografts are generally deployed at a proximal landing zone within the distal aortic arch and at a distal landing zone within
Fig. 4. Representative image series displaying the displacement of the vertex of the aortic arch. Compared to the beginning of the heart cycle the aortic arch exhibited at end-systole a displacement directed cranially.
the true lumen of the descending aorta. Moreover, hybrid procedures combining endovascular and open surgical access enable endograft implantation within the proximal aortic arch or even within the ascending aorta [18,19]. For these reasons, our evaluation of aortic displacement comprised potential proximal (segments AAo and VA) as well as distal (descending aorta, true lumen, DAoT) landing zones. Image evaluation was performed by semiautomatic segmentation of the aortic lumen and calculation of the luminal centroid. Although the results of segmentation algorithms vary with the threshold selection between aortic lumen and wall, Fig. 7 demonstrates that the calculation of the geometrical centroid of the aortic cross-section remains rather unaffected from the size of the actually isolated aortic contour and, therefore, can be performed widely observer-independently. Several studies have examined the vertical motion of the aortic annulus. The contraction of the myocardium causes a shortening of the long cardiac axis that induces downward movement of the aortic root during systole and a return to the initial position during diastole [4,11]. Quantification of this motion pattern has shown a craniocaudal displacement of the aortic root of up to 14 mm [4,20]. Pathological conditions that are associated with an alteration of the cardiac stroke volume lead to an increase (aortic regurgitation) or decrease (left ventricular hypokinesis) of vertical displacement [4,20,21]. Data obtained in this investigation add to the threedimensional understanding of aortic motion associated with cardiac pulsatility. Our evaluation of ascending aortic displacement in patients with CADB generally confirms the results described by other groups in regards of the magnitude of aortic motion, but supplements the representation of its spatial orientation [12]. In addition to the vertical traction of the aortic root, a significant horizontal displacement of the ascending aorta was found in our patient group with an average of approximately 5 mm. The mid-ascending aortic displacement exhibits an ovoid motion pattern that is characterized by a fast and early systolic excursion to anterior and a slow diastolic regression to the posterior starting point (Figs. 3A and 4). In contrast to the ascending aorta, the dissected descending segments present a significantly lesser magnitude of displacement and appear rather fixed at their paravertebral position. This circumstance results in a variable distance between both locations during
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
487
Fig. 5. Time courses of segmental aortic motion. Despite of the different magnitudes of displacement a biphasic motion pattern including systolic displacement and diastolic regression was observed at the ascending aorta (a), the vertex of the aortic arch (b) and the dissected descending aorta (c). The difference between the ascending and descending aorta resulted in a distance variability between these segments (d).
the heart cycle, and, thus, concomitant cyclic strains on pivotal sites at the aortic arch may be suspected. Downward movement of the aortic root alone was shown to cause an increase in wall shear stress at the mid-ascending aorta and, therefore, was discussed as representing a potential manifestation factor for type A dissections. By adding another anterior–posterior dimension of displacement,
Fig. 6. Segmental distribution of aortic motion along thoracic vessel course. There was significant reduction of aortic motion from ascending aorta to aortic arch and dissected descending aorta.
local wall shear stress may increase even more at those critical locations and may favor – combined with, e.g., elevated blood pressure values – the formation of aortic dissection [4,10]. Results published by van Prehn et al. describe that aortic motion decreases along the course of the ascending aorta [12]. Our analysis of the aortic vertex confirms that this pattern is being carried forward distally through the aortic arch. Surprisingly, in our patient group the aortic arch seems not to be dislocated caudally like the aortic root but rather cranially by approximately 2 mm (Figs. 4 and 5B). After ruling out that this observation was an artifact from intermittent inclusion of supraaortic origins into the segmented vessel contour, we assume that it could be an effect from the stroke volume being driven through the aortic curvature resulting in considerable centrifugal forces. According to biomechanical investigations the wall stress within the aortic arch consistently reaches maximum values at its top [3,5,10]. As the distribution of the stress peaks was implicated with the onset and morphology of aortic dissection, it may be hypothesized that the craniocaudal displacement of the thoracic aortic vertex plays an additional role in the development of aortic dissection type B in loco typico. Although TEVAR is associated with favorable short- and midterm outcomes, endograft design and manufacture have not kept pace with the rapidly growing clinical ambition [6,8,22,23]. Major challenges concerning endovascular procedures are the relative rigidity of thoracic endografts and their failure to conform to the anatomy of the aortic arch. However, non-conformity of endografts
488
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
Fig. 7. Threshold independent identification of aortic luminal center. Differing threshold selections for segmentation of aortic cross-section do not necessarily result in deviating calculations of aortic luminal center.
may lead to device instability and endograft migration [24,25]. As this study has demonstrated that aortic displacement in patients with CADB may be distributed inhomogeneously over the thoracic aorta, magnitude and direction of aortic motion may influence design and durability requirements of dedicated thoracic endografts. The significantly lesser extent of aortic displacement at the distal landing zone (segment DAoT) compared to the proximal landing zones (VA and AAo) suggests special strains on endografts that are implanted within these segments. 5. Conclusions In addition to the downward movement of the aortic annulus during the cardiac cycle, the ascending aorta undergoes a significant displacement directed to the anterior part of the thorax. Since true and false channel at the descending aorta remain in a rather stable spatial position and the vertex of the aortic arch tends to show a cranial displacement, substantial traction forces acting on the aortic arch have to be assumed. These observations may contribute to the understanding of the onset of aortic dissections of type B at their preferred localization at the pivotal aortic arch and may be meaningful for the ongoing development and design of thoracic endografts. References [1] Larson EW, Edwards WD. Risk factors for aortic dissection: a necropsy study of 161 cases. Am J Cardiol 1984;53(6):849–55. [2] Januzzi JL, Isselbacher EM, Fattori R, et al. Characterizing the young patient with aortic dissection: results from the International Registry of Aortic Dissection (IRAD). J Am Coll Cardiol 2004;43(4):665–9. [3] Gao F, Watanabe M, Matsuzawa T. Stress analysis in a layered aortic arch model under pulsatile blood flow. Biomed Eng Online 2006:525. [4] Beller CJ, Labrosse MR, Thubrikar MJ, Robicsek F. Finite element modeling of the thoracic aorta: including aortic root motion to evaluate the risk of aortic dissection. J Med Eng Technol 2008;32(2):167–70. [5] Koullias G, Modak R, Tranquilli M, et al. Mechanical deterioration underlies malignant behavior of aneurysmal human ascending aorta. J Thorac Cardiovasc Surg 2005;130(3):677–83. [6] Lopera J, Patino JH, Urbina C, et al. Endovascular treatment of complicated type-B aortic dissection with stent-grafts: midterm results. J Vasc Interv Radiol 2003;14(2 Pt 1):195–203. [7] Böckler D, Schumacher H, Ganten M, et al. Complications after endovascular repair of acute symptomatic and chronic expanding Stanford type B aortic dissections. J Thorac Cardiovasc Surg 2006;132(2):361–8.
[8] Nienaber CA, Kische S, Ince H. Thoracic aortic stent-graft devices: problems, failure modes, and applicability. Semin Vasc Surg 2007;20(2):81–9. [9] Weber TF, Ganten M, Böckler D, et al. Assessment of thoracic aortic conformational changes by fourdimensional computed tomography angiography in patients with chronic aortic dissection type b. Eur Radiol 2008 Jul 22 [Epub ahead of print]. [10] Beller CJ, Labrosse MR, Thubrikar MJ, Robicsek F. Role of aortic root motion in the pathogenesis of aortic dissection. Circulation 2004;109(6):763–9. [11] Emilsson K, Egerlid R, Nygren BM. Comparison between aortic annulus motion and mitral annulus motion obtained using echocardiography. Clin Physiol Funct Imaging 2006;26(5):257–62. [12] van Prehn J, Vincken KL, Muhs BE, et al. Toward endografting of the ascending aorta: insight into dynamics using dynamic cine-CTA. J Endovasc Ther 2007;14(4):551–60. [13] Ganten M, Weber TF, von Tengg-Kobligk H, et al. Motion characterization of aortic wall and intimal flap by ECG-gated CT in patients with chronic B-dissection. Eur J Radiol 2008 Aug 2 [Epub ahead of print]. [14] Ganten M, Boese JM, Leitermann D, Semmler W. Quantification of aortic elasticity: development and experimental validation of a method using computed tomography. Eur Radiol 2005;15(12):2506–12. [15] Ganten M, Krautter U, Hosch W, et al. Age related changes of human aortic distensibility: evaluation with ECG-gated CT. Eur Radiol 2007;17(3):701–8. [16] Chung JW, Elkins C, Sakai T, et al. True-lumen collapse in aortic dissection: part I. Evaluation of causative factors in phantoms with pulsatile flow. Radiology 2000;214(1):87–98. [17] Chung JW, Elkins C, Sakai T, et al. True-lumen collapse in aortic dissection: part II. Evaluation of treatment methods in phantoms with pulsatile flow. Radiology 2000;214(1):99–106. [18] Schumacher H, Böckler D, Bardenheuer H, Hansmann J, Allenberg JR. Endovascular aortic arch reconstruction with supra-aortic transposition for symptomatic contained rupture and dissection: early experience in 8 high-risk patients. J Endovasc Ther 2003;10(6):1066–74. [19] Schumacher H, Von Tengg-Kobligk H, Ostovic M, et al. Hybrid aortic procedures for endoluminal arch replacement in thoracic aneurysms and type B dissections. J Cardiovasc Surg (Torino) 2006;47(5):509–17. [20] Kozerke S, Scheidegger MB, Pedersen EM, Boesiger P. Heart motion adapted cine phase-contrast flow measurements through the aortic valve. Magn Reson Med 1999;42(5):970–8. [21] Pratt RC, Parisi AF, Harrington JJ, Sasahara AA. The influence of left ventricular stroke volume on aortic root motion: an echocardiographic study. Circulation 1976;53(6):947–53. [22] Krohg-Sorensen K, Hafsahl G, Fosse E, Geiran OR. Acceptable short-term results after endovascular repair of diseases of the thoracic aorta in high risk patients. Eur J Cardiothorac Surg 2003;24(3):379–87. [23] Stone DH, Brewster DC, Kwolek CJ, et al. Stent-graft versus open-surgical repair of the thoracic aorta: mid-term results. J Vasc Surg 2006;44(6): 1188–97. [24] Böckler D, von Tengg-Kobligk H, Schumacher H, et al. Late surgical conversion after thoracic endograft failure due to fracture of the longitudinal support wire. J Endovasc Ther 2005;12(1):98–102. [25] Böckler D, Allenberg JR, Kauczor HU, von Tengg-Kobligk H. Images in vascular medicine. Postprocessing ‘fly-through’ of multislice computed tomography after thoracic endografting. Vasc Med 2005;10(1):61–2.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Heartbeat-related displacement of the thoracic aorta in patients with chronic aortic dissection type B: Quantification by dynamic CTA Tim F. Weber a,∗ , Maria-Katharina Ganten b,1 , Dittmar Böckler c,2 , Philipp Geisbüsch c,3 , Hans-Ulrich Kauczor a,4 , Hendrik von Tengg-Kobligk b,5 a
University of Heidelberg, Department of Diagnostic and Interventional Radiology, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany German Cancer Research Center, Department of Radiology, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany c University of Heidelberg, Department of Vascular and Endovascular Surgery, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany b
a r t i c l e
i n f o
Article history: Received 25 May 2008 Received in revised form 22 July 2008 Accepted 31 July 2008 Keywords: Computed tomography angiography Aorta Aortic dissection Endovascular aortic repair
a b s t r a c t Purpose: The purpose of this study was to characterize the heartbeat-related displacement of the thoracic aorta in patients with chronic aortic dissection type B (CADB). Materials and methods: Electrocardiogram-gated computed tomography angiography was performed during inspiratory breath-hold in 11 patients with CADB: Collimation 16 mm × 1 mm, pitch 0.2, slice thickness 1 mm, reconstruction increment 0.8 mm. Multiplanar reformations were taken for 20 equidistant time instances through both ascending (AAo) and descending aorta (true lumen, DAoT; false lumen, DAoF) and the vertex of the aortic arch (VA). In-plane vessel displacement was determined by region of interest analysis. Results: Mean displacement was 5.2 ± 1.7 mm (AAo), 1.6 ± 1.0 mm (VA), 0.9 ± 0.4 mm (DAoT), and 1.1 ± 0.4 mm (DAoF). This indicated a significant reduction of displacement from AAo to VA and DAoT (p < 0.05). The direction of displacement was anterior for AAo and cranial for VA. Conclusion: In CADB, the thoracic aorta undergoes a heartbeat-related displacement that exhibits an unbalanced distribution of magnitude and direction along the thoracic vessel course. Since consecutive traction forces on the aortic wall have to be assumed, these observations may have implications on pathogenesis of and treatment strategies for CADB. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction The development of aortic dissections is associated with arterial hypertension and degenerative or hereditary vascular disease [1,2]. Besides a loss of vessel wall elasticity that accompanies these conditions, aortic motion has been identified as a potential trigger for the onset of dissections within the ascending aorta (type
∗ Corresponding author. Tel.: +49 6221 38438; fax: +49 6221 565730. E-mail addresses: [email protected] (T.F. Weber), [email protected] (M.-K. Ganten), [email protected] (D. Böckler), [email protected] (P. Geisbüsch), [email protected] (H.-U. Kauczor), [email protected] (H. von Tengg-Kobligk). 1 Tel.: +49 6221 422493, fax: +49 6221 422462. 2 Tel.: +49 6221 566249; fax: +49 6221 565423. 3 Tel.: +49 6221 566249; fax: +49 6221 565423. 4 Tel.: +49 6221 56410; fax: +49 6221 565730. 5 Tel.: +49 6221 422492; fax: +49 6221 422462. 0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.07.045
A, Stanford classification) [3,4]. Both of the phenomena – aortic motion and reduction of elasticity – lead to an increase of vessel wall shear stress, which is in turn considered to be the underlying biomechanical factor for intimal detachment and chronic aortic expansion [4,5]. Expanding chronic aortic dissections of the descending aorta (type B, Stanford classification, CADB) are increasingly treated with thoracic endovascular aortic repair (TEVAR) [6]. Although this minimal-invasive approach reduces periprocedural mortality in selected patients, there are several procedure-related challenges that correlate with both the individual’s aortic anatomy and the geometrical characteristics of available endografts [7,8]. TEVAR of CADB is hampered by the tortuosity of aortic arch and descending aorta, which may circumvent stable attachment of insufficiently flexible endografts at the proximal landing zone. The implanted endograft fabric is constantly exposed to significant radial forces that result from a heartbeat-associated cyclic distension of the aortic cross-section being observed in CADB as well [9]. In addition, continuous motion of the thoracic aorta following from cardiac pul-
484
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
sation is supposed to increase structural demands on design and material properties of thoracic aortic prostheses [8]. It has been demonstrated that the aortic root may undergo systolic displacement directed towards the left ventricular apex [10,11]. Recently, corresponding observations of aortic motion have been extended to the ascending thoracic aorta in patients with abdominal aortic aneurysms [12]. Aim of this study was to elucidate both the magnitude and direction of the heartbeat-related displacement of the ascending and descending thoracic aorta as well as of the vertex of the aortic arch in patients with CADB by using electrocardiogram-gated dynamic computed tomography angiography (CTA). 2. Materials and methods 2.1. Study patients From the database of a prospective study a subpopulation of 11 consecutive patients with CADB (9 men, 2 women; mean age ± standard deviation, 62.7 ± 12.0 years, mean age of dissection ± standard deviation, 22.9 ± 21.0 months) were enrolled in this evaluation [13]. For inclusion into this study, candidates had to match the following criteria: existence of chronic aortic dissection of the descending aorta with an intimal detachment arising immediately distally to the origin of the left subclavian artery and exclusively treated with antihypertensive pharmacotherapy. Patients with aortic dissection arising at other localizations – including (retrograde) dissections of the ascending aorta and dissections of the descending aorta located more distally than described above – as well as patients with preceded surgical or endovascular repair of the aorta were primarily excluded. Approval of the local ethics committee and written informed consent were granted before CTA exams. 2.2. Image acquisition CTA of the aorta was performed with a 16-row multislice scanner (Aquilion 16® ; Toshiba Medical Systems, Otawara, Japan). As the CT scan served as a regular follow-up examination, a dose optimized biphasic acquisition protocol was designed: A time-resolved dynamic data set of the thoracic aorta provided for scientific and clinical assessment, and static cross-sections covering abdominal aorta and iliofemoral branches completed diagnostic work-up [13–15]. In the first step the thoracic aorta was scanned during inspiratory breath-hold with the following parameters: collimation 16 mm × 1 mm, tube rotation time 0.4 s, pitch 0.2, tube voltage 120 kV, tube current 300 mA, field of view 320 mm. An electrocardiogram (ECG) was simultaneously recorded for retrospective ECG-gating during this period of acquisition. In the second step the abdominal aorta was examined conventionally after switching off the ECG-gating and setting the scan parameters as follows: collimation 16 mm × 1 mm, tube rotation time 0.5 s, pitch 0.94, tube voltage 120 kV, tube current 170 mA, field of view 320 mm. Images generated in the second step were not part of scientific evaluation. For enhanced vessel contrast each patient received 130 ml of a non-ionic iodinated contrast medium with an iodine content of 300 mg/ml at an injection rate of 4 ml/s (Iomeprol, Iomeron 300® ; Bracco International, Milan, Italy) through an 18-gauge catheter positioned in a right-sided antecubital vein. Bolus timing was achieved using an automated triggering technique with a threshold of 110 HU within a region of interest (ROI) placed in the ascending aorta.
2.3. Image reconstruction and post-processing Retrospective ECG-gating enabled multiphasic image reconstruction of the thoracic aorta at 20 equidistant time points of the R-R interval. Transverse source data were reconstructed with a section thickness of 1 mm and an overlapping increment of 0.8 mm. The 20 data sets of each patient were transferred to a postprocessing workstation equipped with dedicated medical imaging software that permitted editing of a fourdimensional scan volume (Aquarius Workstation® ; TeraRecon, San Mateo, USA). To assess aortic in-plane displacement multiplanar reformations (MPR) were taken in strictly transverse orientation through the ascending (half way between the aortic annulus and the origin of the brachiocephalic trunk, segment AAo) and descending aorta (10 cm distal from the origin of the left subclavian artery, segment DAo) as well as in strictly coronary orientation through the vertex of the aortic arch (segment VA) (Fig. 1). As segment DAo was located at the dissected aorta, separate evaluations were performed for true (DAoT) and false channels (DAoF). The vertex of the aortic arch was defined as the most cranial location of the curvature and was situated either proximal (n = 6) or distal (n = 5) from the origin of the left subclavian artery. The resulting MPR (n = 20 per location) represented both temporal and spatial positions of the aorta and were exported as primary DICOM data for further quantitative image evaluation. 2.4. Image evaluation MPR image series were loaded into a software application that provided quantitative visualization and analysis of medical images (MIPAV® ; Center for Information Technology, National Institutes of Health, Bethesda, USA). Vessel analysis was carried out in a consensus reading of two radiologists with 5 years and 2 years of experience in computed tomography, respectively. Using a semiautomatic segmentation algorithm and based upon contrast differences to the lumen boundary, a ROI was generated over the aortic contour on each time instance. Propagated and adapted to each time instance the software calculated the centroid of each ROI based upon its geometry. From the temporal alteration of the centroid coordinates total extent, temporal development and spatial direction of in-plane vessel displacement were determined relative to the first reconstruction interval (Fig. 2). Mean displacement, standard deviation and range were calculated from the maximum segmental in-plane displacement of each time series. For assessing the temporal development of aortic displacement mean displacement of each time instance was plotted against the reconstruction interval. 2.5. Statistical analysis For testing for significance of differences between segmental distributions of aortic displacement the Wilcoxon Signed Ranks test was used. A p-value smaller 0.05 was considered to indicate statistical significance. 2.6. Study limitations To provide a homogeneous study population and to facilitate comparable measurements, all individuals to be included had to feature similar disease morphologies, namely a CADB with an intimal flap originating immediately distally to the outlet of the left subclavian artery. Because of this, only a limited number of patients could be included into this study, and statistics had to be calculated from small case numbers.
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
485
Fig. 2. Quantification of the magnitude of in-plane displacement of the aortic cross-sectional centroid. Segmentation of the aortic lumen (segment AAo) provided coordinates of the luminal centroid at each time instance. The distance between those were calculated from the coordinates’ differences using the theorem of Pythagoras.
The magnitude of aortic displacement may depend on blood pressure and cardiac stroke volume. Although these patients were on best medical treatment including blood pressure control, it is possible that the mean values of aortic displacement documented here may be influenced by arterial hypertension potentially existent in some individuals. The presented evaluation procedure is able to reveal both magnitude and orientation of in-plane displacement of reformatted aortic cross-sections. However, the effect of a concomitant displacement directed through-plane and its influence on our measurements had to be disregarded. Although mere visualization of fourdimensional data sets is possible with available imaging software, quantifying aortic motion in all spatial directions simultaneously requires computational algorithms that enable not only segmentation of the aortic lumen but also identification and particularly tracking of specific anatomic landmarks. 3. Results Mean displacement (±standard deviation) of AAo was 5.2 ± 1.7 mm (range, 3.2–7.5 mm) and of VA 1.6 ± 1.0 mm (0.4–4.0 mm). While displacement of AAo was directed anteriorly, VA was displaced in cranial orientation (Figs. 2–4). The true channel DAoT was displaced by 0.9 ± 0.4 mm (0.4–1.5 mm) and the false channel DAoF by 1.1 ± 0.4 mm (0.6–1.9 mm). DAoT and DAoF did not reveal a significant direction of displacement. At all segments a phase of fast systolic displacement was followed by slow diastolic regression to the starting point (Fig. 5). Displacement was significantly greater at AAo compared to VA and both DAoT and DAoF (p < 0.005 each). VA was displaced significantly greater than DAoT (p = 0.023). No differences were found between VA and DAoF (p = 0.095) and between DAoT and DAoF (p = 0.250) (Fig. 6). Fig. 1. Representative thoracic aortic cross-sections adjusted by single-oblique MPR. To assure analysis of precise orientation of aortic displacement along (segment VA, B) or perpendicular to (segments AAo and DAo, A and C) the patient’s longitudinal axis, each target reformation was adjusted by single-oblique MPR keeping the imaging plane of the target orientation untouched.
4. Discussion This study reports on magnitude and direction of the heartbeatrelated displacement of the thoracic aorta in patients with CADB. Using a previously developed and validated ECG-gated CTA
486
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
Fig. 3. Visualization of the orientation of aortic motion at the mid-ascending aorta and the vertex of the aortic arch. While the ascending aorta underwent displacement predominantly directed anteriorly (A), the vertex of the aortic arch exhibited a slight cranial displacement (B).
protocol, a dynamic data set was acquired that provided multidimensional information on the spatial position of the thoracic aorta [13]. The cyclic displacement of the thoracic aorta is of clinical interest, because on the one hand pathogenesis of ascending aortic dissections has been previously associated with aortic motion and on the other hand CADB and other diseases of the thoracic aorta are increasingly referred to endovascular aortic repair [4,8]. TEVAR of CADB aims at sealing entry tears to initiate false lumen thrombosis and resolution of true lumen collapse [16,17]. The implanted endograft is thereafter exposed to continuous physical strains that result from pulsatile blood flow and aortic motion. For this reason, life-long surveillance is necessary after performing endovascular procedures to timely detect endograft failure, e.g., migration of the prosthesis or wire fracture. In CADB, endografts are generally deployed at a proximal landing zone within the distal aortic arch and at a distal landing zone within
Fig. 4. Representative image series displaying the displacement of the vertex of the aortic arch. Compared to the beginning of the heart cycle the aortic arch exhibited at end-systole a displacement directed cranially.
the true lumen of the descending aorta. Moreover, hybrid procedures combining endovascular and open surgical access enable endograft implantation within the proximal aortic arch or even within the ascending aorta [18,19]. For these reasons, our evaluation of aortic displacement comprised potential proximal (segments AAo and VA) as well as distal (descending aorta, true lumen, DAoT) landing zones. Image evaluation was performed by semiautomatic segmentation of the aortic lumen and calculation of the luminal centroid. Although the results of segmentation algorithms vary with the threshold selection between aortic lumen and wall, Fig. 7 demonstrates that the calculation of the geometrical centroid of the aortic cross-section remains rather unaffected from the size of the actually isolated aortic contour and, therefore, can be performed widely observer-independently. Several studies have examined the vertical motion of the aortic annulus. The contraction of the myocardium causes a shortening of the long cardiac axis that induces downward movement of the aortic root during systole and a return to the initial position during diastole [4,11]. Quantification of this motion pattern has shown a craniocaudal displacement of the aortic root of up to 14 mm [4,20]. Pathological conditions that are associated with an alteration of the cardiac stroke volume lead to an increase (aortic regurgitation) or decrease (left ventricular hypokinesis) of vertical displacement [4,20,21]. Data obtained in this investigation add to the threedimensional understanding of aortic motion associated with cardiac pulsatility. Our evaluation of ascending aortic displacement in patients with CADB generally confirms the results described by other groups in regards of the magnitude of aortic motion, but supplements the representation of its spatial orientation [12]. In addition to the vertical traction of the aortic root, a significant horizontal displacement of the ascending aorta was found in our patient group with an average of approximately 5 mm. The mid-ascending aortic displacement exhibits an ovoid motion pattern that is characterized by a fast and early systolic excursion to anterior and a slow diastolic regression to the posterior starting point (Figs. 3A and 4). In contrast to the ascending aorta, the dissected descending segments present a significantly lesser magnitude of displacement and appear rather fixed at their paravertebral position. This circumstance results in a variable distance between both locations during
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
487
Fig. 5. Time courses of segmental aortic motion. Despite of the different magnitudes of displacement a biphasic motion pattern including systolic displacement and diastolic regression was observed at the ascending aorta (a), the vertex of the aortic arch (b) and the dissected descending aorta (c). The difference between the ascending and descending aorta resulted in a distance variability between these segments (d).
the heart cycle, and, thus, concomitant cyclic strains on pivotal sites at the aortic arch may be suspected. Downward movement of the aortic root alone was shown to cause an increase in wall shear stress at the mid-ascending aorta and, therefore, was discussed as representing a potential manifestation factor for type A dissections. By adding another anterior–posterior dimension of displacement,
Fig. 6. Segmental distribution of aortic motion along thoracic vessel course. There was significant reduction of aortic motion from ascending aorta to aortic arch and dissected descending aorta.
local wall shear stress may increase even more at those critical locations and may favor – combined with, e.g., elevated blood pressure values – the formation of aortic dissection [4,10]. Results published by van Prehn et al. describe that aortic motion decreases along the course of the ascending aorta [12]. Our analysis of the aortic vertex confirms that this pattern is being carried forward distally through the aortic arch. Surprisingly, in our patient group the aortic arch seems not to be dislocated caudally like the aortic root but rather cranially by approximately 2 mm (Figs. 4 and 5B). After ruling out that this observation was an artifact from intermittent inclusion of supraaortic origins into the segmented vessel contour, we assume that it could be an effect from the stroke volume being driven through the aortic curvature resulting in considerable centrifugal forces. According to biomechanical investigations the wall stress within the aortic arch consistently reaches maximum values at its top [3,5,10]. As the distribution of the stress peaks was implicated with the onset and morphology of aortic dissection, it may be hypothesized that the craniocaudal displacement of the thoracic aortic vertex plays an additional role in the development of aortic dissection type B in loco typico. Although TEVAR is associated with favorable short- and midterm outcomes, endograft design and manufacture have not kept pace with the rapidly growing clinical ambition [6,8,22,23]. Major challenges concerning endovascular procedures are the relative rigidity of thoracic endografts and their failure to conform to the anatomy of the aortic arch. However, non-conformity of endografts
488
T.F. Weber et al. / European Journal of Radiology 72 (2009) 483–488
Fig. 7. Threshold independent identification of aortic luminal center. Differing threshold selections for segmentation of aortic cross-section do not necessarily result in deviating calculations of aortic luminal center.
may lead to device instability and endograft migration [24,25]. As this study has demonstrated that aortic displacement in patients with CADB may be distributed inhomogeneously over the thoracic aorta, magnitude and direction of aortic motion may influence design and durability requirements of dedicated thoracic endografts. The significantly lesser extent of aortic displacement at the distal landing zone (segment DAoT) compared to the proximal landing zones (VA and AAo) suggests special strains on endografts that are implanted within these segments. 5. Conclusions In addition to the downward movement of the aortic annulus during the cardiac cycle, the ascending aorta undergoes a significant displacement directed to the anterior part of the thorax. Since true and false channel at the descending aorta remain in a rather stable spatial position and the vertex of the aortic arch tends to show a cranial displacement, substantial traction forces acting on the aortic arch have to be assumed. These observations may contribute to the understanding of the onset of aortic dissections of type B at their preferred localization at the pivotal aortic arch and may be meaningful for the ongoing development and design of thoracic endografts. References [1] Larson EW, Edwards WD. Risk factors for aortic dissection: a necropsy study of 161 cases. Am J Cardiol 1984;53(6):849–55. [2] Januzzi JL, Isselbacher EM, Fattori R, et al. Characterizing the young patient with aortic dissection: results from the International Registry of Aortic Dissection (IRAD). J Am Coll Cardiol 2004;43(4):665–9. [3] Gao F, Watanabe M, Matsuzawa T. Stress analysis in a layered aortic arch model under pulsatile blood flow. Biomed Eng Online 2006:525. [4] Beller CJ, Labrosse MR, Thubrikar MJ, Robicsek F. Finite element modeling of the thoracic aorta: including aortic root motion to evaluate the risk of aortic dissection. J Med Eng Technol 2008;32(2):167–70. [5] Koullias G, Modak R, Tranquilli M, et al. Mechanical deterioration underlies malignant behavior of aneurysmal human ascending aorta. J Thorac Cardiovasc Surg 2005;130(3):677–83. [6] Lopera J, Patino JH, Urbina C, et al. Endovascular treatment of complicated type-B aortic dissection with stent-grafts: midterm results. J Vasc Interv Radiol 2003;14(2 Pt 1):195–203. [7] Böckler D, Schumacher H, Ganten M, et al. Complications after endovascular repair of acute symptomatic and chronic expanding Stanford type B aortic dissections. J Thorac Cardiovasc Surg 2006;132(2):361–8.
[8] Nienaber CA, Kische S, Ince H. Thoracic aortic stent-graft devices: problems, failure modes, and applicability. Semin Vasc Surg 2007;20(2):81–9. [9] Weber TF, Ganten M, Böckler D, et al. Assessment of thoracic aortic conformational changes by fourdimensional computed tomography angiography in patients with chronic aortic dissection type b. Eur Radiol 2008 Jul 22 [Epub ahead of print]. [10] Beller CJ, Labrosse MR, Thubrikar MJ, Robicsek F. Role of aortic root motion in the pathogenesis of aortic dissection. Circulation 2004;109(6):763–9. [11] Emilsson K, Egerlid R, Nygren BM. Comparison between aortic annulus motion and mitral annulus motion obtained using echocardiography. Clin Physiol Funct Imaging 2006;26(5):257–62. [12] van Prehn J, Vincken KL, Muhs BE, et al. Toward endografting of the ascending aorta: insight into dynamics using dynamic cine-CTA. J Endovasc Ther 2007;14(4):551–60. [13] Ganten M, Weber TF, von Tengg-Kobligk H, et al. Motion characterization of aortic wall and intimal flap by ECG-gated CT in patients with chronic B-dissection. Eur J Radiol 2008 Aug 2 [Epub ahead of print]. [14] Ganten M, Boese JM, Leitermann D, Semmler W. Quantification of aortic elasticity: development and experimental validation of a method using computed tomography. Eur Radiol 2005;15(12):2506–12. [15] Ganten M, Krautter U, Hosch W, et al. Age related changes of human aortic distensibility: evaluation with ECG-gated CT. Eur Radiol 2007;17(3):701–8. [16] Chung JW, Elkins C, Sakai T, et al. True-lumen collapse in aortic dissection: part I. Evaluation of causative factors in phantoms with pulsatile flow. Radiology 2000;214(1):87–98. [17] Chung JW, Elkins C, Sakai T, et al. True-lumen collapse in aortic dissection: part II. Evaluation of treatment methods in phantoms with pulsatile flow. Radiology 2000;214(1):99–106. [18] Schumacher H, Böckler D, Bardenheuer H, Hansmann J, Allenberg JR. Endovascular aortic arch reconstruction with supra-aortic transposition for symptomatic contained rupture and dissection: early experience in 8 high-risk patients. J Endovasc Ther 2003;10(6):1066–74. [19] Schumacher H, Von Tengg-Kobligk H, Ostovic M, et al. Hybrid aortic procedures for endoluminal arch replacement in thoracic aneurysms and type B dissections. J Cardiovasc Surg (Torino) 2006;47(5):509–17. [20] Kozerke S, Scheidegger MB, Pedersen EM, Boesiger P. Heart motion adapted cine phase-contrast flow measurements through the aortic valve. Magn Reson Med 1999;42(5):970–8. [21] Pratt RC, Parisi AF, Harrington JJ, Sasahara AA. The influence of left ventricular stroke volume on aortic root motion: an echocardiographic study. Circulation 1976;53(6):947–53. [22] Krohg-Sorensen K, Hafsahl G, Fosse E, Geiran OR. Acceptable short-term results after endovascular repair of diseases of the thoracic aorta in high risk patients. Eur J Cardiothorac Surg 2003;24(3):379–87. [23] Stone DH, Brewster DC, Kwolek CJ, et al. Stent-graft versus open-surgical repair of the thoracic aorta: mid-term results. J Vasc Surg 2006;44(6): 1188–97. [24] Böckler D, von Tengg-Kobligk H, Schumacher H, et al. Late surgical conversion after thoracic endograft failure due to fracture of the longitudinal support wire. J Endovasc Ther 2005;12(1):98–102. [25] Böckler D, Allenberg JR, Kauczor HU, von Tengg-Kobligk H. Images in vascular medicine. Postprocessing ‘fly-through’ of multislice computed tomography after thoracic endografting. Vasc Med 2005;10(1):61–2.