- Email: [email protected]
Multidimensional growth measurements of abdominal aortic aneurysms
BASIC RESEARCH STUDIES
Multidimensional growth measurements of abdominal aortic aneurysms Giampaolo Martufi, PhD,a Martin Auer, PhD,b Joy Roy, MD, PhD,c Jesper Swedenborg, MD, PhD,c Natzi Sakalihasan, MD, PhD,d Giuseppe Panuccio, MD,e,f and T. Christian Gasser, PhD,a Stockholm, Sweden; Vienna, Austria; Liège, Belgium; Perugia, Italy; and Münster, Germany Background: Monitoring the expansion of abdominal aortic aneurysms (AAAs) is critical to avoid aneurysm rupture in surveillance programs, for instance. However, measuring the change of the maximum diameter over time can only provide limited information about AAA expansion. Specifically, regions of fast diameter growth may be missed, axial growth cannot be quantified, and shape changes of potential interest for decisions related to endovascular aneurysm repair cannot be captured. Methods: This study used multiple centerline-based diameter measurements between the renal arteries and the aortic bifurcation to quantify AAA growth in 51 patients from computed tomography angiography (CTA) data. Criteria for inclusion were at least 1 year of patient follow-up and the availability of at least two sufficiently high-resolution CTA scans that allowed an accurate three-dimensional reconstruction. Consequently, 124 CTA scans were systematically analyzed by using A4clinics diagnostic software (VASCOPS GmbH, Graz, Austria), and aneurysm growth was monitored at 100 cross-sections perpendicular to the centerline. Results: Monitoring diameter development over the entire aneurysm revealed the sites of the fastest diameter growth, quantified the axial growth, and showed the evolution of the neck morphology over time. Monitoring the development of an aneurysm’s maximum diameter or its volume over time can assess the mean diameter growth (r [ 0.69, r [ 0.77) but not the maximum diameter growth (r [ 0.43, r [ 0.34). The diameter growth measured at the site of maximum expansion was w16%/y, almost four times larger than the mean diameter expansion of 4.4%/y. The sites at which the maximum diameter growth was recorded did not coincide with the position of the maximum baseline diameter (r [ 0 .12; P [ .31). The overall aneurysm sac length increased from 84 to 89 mm during the follow-up (P < .001), which relates to the median longitudinal growth of 3.5%/y. The neck length shortened, on average, by 6.2% per year and was accompanied by a slight increase in neck angulation. Conclusions: Neither maximum diameter nor volume measurements over time are able to measure the fastest diameter growth of the aneurysm sac. Consequently, expansion-related wall weakening might be inappropriately reflected by this type of surveillance data. In contrast, localized spots of fast diameter growth can be detected through multiple centerlinebased diameter measurements over the entire aneurysm sac. This information might further reinforce the quality of aneurysm surveillance programs. (J Vasc Surg 2013;58:748-55.) Clinical Relevance: Although the size of the aneurysm still remains the most accepted predictor of rupture risk, small abdominal aortic aneurysms (AAAs) may also rupture. Thus, the new tools presented in this study use three-dimensional aneurysm models, based on computed tomography angiography and multiple centerline-based diameter measurements over the entire aneurysm sac, could help clinicians when making decisions to treat AAAs and when choosing alternative treatments for AAAs, regardless of the diameter size. From the Department of Solid Mechanics, School of Engineering Sciences, Royal Institute of Technology (KTH), Stockholma; Vascops GmbH, Development, Viennab; the Department of Vascular Surgery, Karolinska University Hospital and Institute, Stockholmc; the Department of Vascular Surgery, University Hospital of Liege, Lièged; the Division of Vascular and Endovascular Surgery, University of Perugia, Perugiae; and the Clinic for Vascular and Endovascular Surgery, Münster University Hospital, Münster.f Author conflict of interest: none. Reprint requests: Dr Giampaolo Martufi, Royal Institute of Technology (KTH), Department of Solid Mechanics, School of Engineering Sciences, Teknikringen 8, 100 44 Stockholm. Sweden (e-mail: martufi@kth.se or http://130.237.24.173/vascumech/index.html). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214/$36.00 Copyright Ó 2013 by the Society for Vascular Surgery. http://dx.doi.org/10.1016/j.jvs.2012.11.070
748
The indication for elective abdominal aortic aneurysm (AAA) repair is strongly related to the aneurysm’s risk for rupture and is given for aneurysms that reach a maximum centerline-based diameter of 55 mm.1 However, several reports have demonstrated a rupture risk of AAAs sized <5 cm in diameter and also that the rate of expansion of the aneurysm is crucial for a certain number of AAAs, independent of their size.2-4 Therefore, a single threshold diameter is not appropriate for every patient, and the decision to perform elective AAA repair should ideally be individualized. Consequently, apart from aneurysm size, many other risk factors for aneurysm rupture have been suggested. For example, aneurysm shape, blood pressure, female gender, and other wall-weakening effects have been mentioned specifically.5
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Apart from static criteria (ie, risk factors evaluated at one time point), dynamic risk factors, such as the expansion rate, have also been suggested.2-4 Aneurysm expansion in common clinical practice is defined as the change in maximum aortic diameter over time, and an expansion rate of 10 mm/y is generally considered as an indication for elective repair.1,6 This practice is justified to some extent through an expected wall weakening effect in response to undesirable wall remodeling at fast AAA growth. Specifically, aneurysm expansion correlates with the density of the inflammatory cell,7,8 which in turn increases matrix metalloproteinase activity7,8 and leads to compromised wall strength.9,10 This biologic activity is localized, such that spots of increased expression and activation of matrix metalloproteinases might contribute to fast local aneurysm expansion.7,8 The biomechanical rupture risk assessment already integrates many static risk factors into indices, such as the peak wall stress11 and the peak wall rupture risk.12-14 Similarly, dynamic wall-weakening effect could potentially be integrated, which might further improve the biomechanical rupture risk assessment. Developing models that prescribe aneurysm wall weakening as a consequence of their expansion rate requires detailed growth information. Available data are currently based on measuring the maximum diameter at two points in time, which clearly cannot provide a comprehensive picture on how aneurysm size and shape change over time. For example, the morphology of the aneurysm neck is especially important to minimize complications related to endovascular aneurysm repair (EVAR),15,16 and its change over time can clearly not be studied from maximum diameter measurements. Measuring the expansion of AAAs also presents a number of methodologic challenges. First, the change in diameter over time is small, nonlinear, and hard to predict.17 It is characterized by periods of rapid growth, followed by quiescience.18,19 Aortic diameter measurements that use ultrasound imaging and, to a lesser extent, computed tomography angiography (CTA), have a margin of error of 2 to 3 mm,17,20,21 which is within the range of the annual expansion of many small AAAs. Finally, the development of treatment protocols based on AAA expansion is somewhat limited due to missing follow-up data for aneurysms that reach the AAA repair indication threshold. Measuring the maximum diameter at two points in time provides only limited information about aneurysm growth. Specifically, regions of fast diameter growth might be missed, axial growth cannot be quantified, and shape changes of potential interest for EVAR-related decisions cannot be captured. Motivated by these limitations, the principal aim of the present study was to explore and investigate the growth pattern of the entire AAA sac. A multidimensional analysis of AAA growth was applied to CTA follow-up data from 51 patients. The applied analysis method used multiple centerline-based diameter measurements between the renal arteries and the aortic bifurcation.
Martufi et al 749
Table I. Patient demographics and abdominal aortic aneurysm (AAA) details Variable Patients Age, years Sex Male Female Mean arterial pressure, mm Hg Surveillance time frame, months AAA diameter, mm Baseline Follow-up
No. or mean (range) 51 70 (49-85) 44 7 97 (90-146) 12 (2-24) 46.6 (34.7-56.4) 49.5 (37-63)
Sites with the highest diameter growth and axial growth, as well as the evolution of the neck morphology over time, could be extracted from the collected measurements. Finally, the derived growth information allowed a critical evaluation of conventional procedures to measure AAA growth. METHODS Cohort. Data for 51 patients with nonruptured infrarenal AAAs were retrospectively collected from the S. M. Misericordia Hospital Perugia, Italy (n ¼ 32), Karolinska University Hospital, Stockholm, Sweden (n¼13), and University Hospital Liege, Belgium (n ¼ 6). Patients were monitored between 2 and 24 months with repeated contrast-enhanced multislice CTA, where the median overall follow-up period was 12 months (interquartile range [IQR], 11 to 13 months). Some further patient data are given in Table I. Criteria for inclusion were at least 1 year of patient follow-up and the availability of at least two sufficiently high-resolution CTA scans that allowed accurate reconstruction at two different time points. CTA scans with strongly inhomogeneous lumen intensity were rejected. Other exclusion criteria were patients with chronic kidney disease, contrast agent intolerance, and symptomatic patients. Patient age, sex, blood pressure, and smoking status were taken from patient records. The collection and use of anonymized data from human subjects was approved by the local ethics committees. Aneurysm geometry representation. Aortic geometries between the renal arteries and the aortic bifurcation from 124 CTA scans were systematically analyzed by a single observer. The three-dimensional (3D) AAA geometry (Fig 1, A) was reconstructed using deformable (active) contour models22 that were implemented in the A4clinics 4.0 diagnostic software (VASCOPS GmbH, Graz, Austria). The accuracy of this type of geometric reconstruction is in the range of one in-plane pixel size.22 All reconstructions were verified by another author of the study. Reconstructed AAA geometry was used to automatically calculate the aneurysm volume (V); that is, the volume
750 Martufi et al
JOURNAL OF VASCULAR SURGERY September 2013
Fig 1. Abdominal aortic aneurysm (AAA) geometry acquisition. A, Three-dimensional anatomy is reconstructed from computed tomography angiography scans using a deformable segmentation model. B, Simplified geometry representation uses splines perpendicular to the aneurysm’s centerline.
Fig 2. Definition of the geometric properties of an abdominal aortic aneurysm (AAA). A, Centerline-based outer aortic diameters. Dmax, maximum diameter; D1, aortic diameter at the lowest renal artery. B, Centerline length segments. N, neck length; S, aneurysm sac length. C, Angulation a.
that is enclosed by the outer surface and between the transversal planes at the lowest renal artery and the aortic bifurcation, respectively. Then, the 3D model was automatically sliced orthogonally to the centerline so that the aneurysm’s luminal and outer surfaces were represented through splines (Fig 1, B). Specifically, 100 splines were used for each aneurysm, which together with the centerline, provided a simplified 3D representation. This geometric representation allowed a fully automatic analysis with a specially developed in-house algorithm using MatLab software (The MathWorks, Natick, Mass). Luminal and outer diameters were measured from each spline, which in turn defined the largest outer
diameter (Dmax) and the diameter (D1) at the lowest renal artery (Fig 2, A). The aneurysm neck length (N), the portion between the lowest renal artery and the proximal onset of the aneurysmatic enlargement (ie, where the outer diameter was 10% larger than the aortic diameter at the lowest renal artery [D1])23,24 was extracted (Fig 2, B). Similarly extracted was the aneurysm sac length (S), the length of the centerline from the proximal onset of the aneurysm down to the aortic bifurcation (Fig 2, B). Finally, the angulation (a) was computed (Fig 2, C); that is, the rotation that is required, such that the extrapolation of the straight centerline in the neck meets the aortic bifurcation.23,24
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Martufi et al 751
Aneurysm growth measurements. To quantify aneurysm growth, the changes of the geometric properties were compared between two consecutive CTA scans. Multiple comparisons were made for patients who had more than two CTA scans, and 7300 expansion observations were processed in the present study. Specifically, the nonlinear growth model g ¼
X 1-year followup 1 $100 X baseline
¼ ½Expð12 rÞ 1$100½%=y; where X baseline and X 1-year follow-up denote baseline and 1-year follow-up quantities, was used. Exp (d) is the exponential function of (d). Because the measured quantity X was not always given at the 1-year follow-up, the loga followup 1 X was used to rithmic growth factor r ¼ Ln X baseline t calculate g, where t denotes the time interval between two measurement points in months. Ln (d) is the natural logarithmic function of (d). To study the diameter expansion, g was computed at 100 cross-sections along the aneurysm. Specifically, the diameters that were at the same relative centerline position, between the two time points at which CTA scans were taken, were thought to correspond to each other. From these measurements, maximum g MAX and mean g MEAN diameter expansions were extracted. Standard maximum-diameter measures were used to define the global maximum diameter-based expansion 1 D followup which compares the rate g D , with r ¼ Ln max baseline Dmax t maximum diameters at baseline and follow-up. Note that baseline followup and Dmax are, in the maximum diameters Dmax general, not at the same relative centerline position. Finally, volume measurements were used to define global volume-based aneurysm expansion g V with followup 1 V r ¼ Ln ; and centerline length measurements V baseline t defined global longitudinal aneurysm expansion g S with followup 1 S . r ¼ Ln S baseline t Statistics. Statistical analysis was performed with SPSS 20.0 software (SPSS Inc, Chicago, Ill). The normality of the data was assessed using the Shapiro-Wilk test. The paired t-test was used for paired variables with normal distribution, and the Wilcoxon matched pairs test was used for non-normal data. For the null hypothesis test, a two-sided P < .05 was set to determine statistical significance. Correlations between the maximum diameter-based expansion rates ðg D Þ, the volume-based expansion rates ðg V Þ, and the multiple centerline-based growth measures of g MAX and g MEAN were assessed. This analysis aimed at quantifying the extent to which aneurysm growth could be extracted from measuring the maximum diameters or
Fig 3. Box and whisker plots show multiple centerline-based aneurysm expansions and global-based aneurysm expansions. g MAX , maximum diameter growth; g MEAN , mean diameter growth, which is averaged between the renal arteries and the aortic bifurcation; g D , global maximum diameter-based expansion rate; g S , global longitudinal aneurysm expansion; g V , global volume-based aneurysm expansion. The horizontal line in the middle of each box indicates the median, the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the top whiskers mark the 75th percentile þ1.5 interquartile range (IQR), while the bottom whiskers mark the 25th percentile 1.5 IQR, and the þ symbols denote outliers.
the volumes at two time points. In addition, the correlation between the site of the maximum diameter and the site of maximum growth was investigated. RESULTS AAA growth measurements. The calculated growth measurements are shown in Fig 3. The maximum diameter growth g MAX (median, 16%/y; IQR, 12.3%/y to 23%/y) was about four times larger than the mean diameter growth g MEAN (median, 4.4%/y; IQR, 2%/y to 7.4%/y). The maximum diameter growth g MAX differed significantly (P < .001) from g D (median, 6.4%/y; IQR, 3.5%/y to 8.6%/y) and g V (median, 10.7%/y; IQR, 7.7%/y to 17%/y); that is, the growth based on relating the maximum diameters or the volumes measured at the two time points, respectively. The overall aneurysm sac length (S) increased (P < .001) from 85.37 mm (IQR, 74.3 to 95.7 mm) to 89.44 mm (IQR, 81.55 to 101.75 mm) during the follow-up, which relates to a median longitudinal growth ðg S Þ of 3.5%/y (IQR, 0.84% to 11.5%). The subgroup analysis reported in Table II found that AAA expansion was not associated with age and mean arterial pressure but were increased in current smokers, women, and larger baseline diameters, similar to previous reports.17,25 However, none of the subgroup results reported in Table II reached statistical significance, probably due to the small sample size. The relative centerline position, at which the maximum diameter growth g MAX was recorded, did not coincide with
JOURNAL OF VASCULAR SURGERY September 2013
752 Martufi et al
Table II. Measured abdominal aortic aneurysm (AAA) growth by patient characteristics at baseline and risk factors AAA growth measurements, % per year (median) Factor Age at baseline, years (tertiles) 49-69 70-74 75-85 Sex Male Female baseline , mm (tertiles) Dmax 35-45 46-48 48-56 Smoking status Never Former Current Unknown MAP at baseline, mm Hg (tertiles) 90-93 94-98 99-147
Pts, No. g MAX g MEAN g D
gV
gS
20 17 14
16.7 22.4 16.8
5 5.8 4
6.1 12.9 7.8 11.7 5.2 10.7
4.3 3.24 3.79
44 7
16.4 20.9
4.6 5.2
5.6 11.8 2.8 7.9 16.9 10.2
17 17 17
15.7 18.3 18.9
3.2 4.7 5.9
4 10.3 7 10.4 7.2 16.7
2.45 3.44 5.9
20 9 11 11
17.2 17.5 22.9 16.4
3.9 4.8 5.6 4.2
4.9 6.1 6.7 7
10.6 12.7 16.7 13.8
1.33 1.97 4.7 1.53
30 4 17
16.6 18.4 18.7
4.7 5.3 4.7
6.4 11.2 4.9 14.6 7.5 14
4 3 3.45
baseline Dmax , Maximum diameter at baseline; g MAX , maximum diameter growth; g MEAN , mean diameter growth; g D , global maximum diameter-based expansion rate; g S , global longitudinal aneurysm expansion; g V , global volume-based aneurysm expansion; MAP, mean arterial pressure.
baseline the position of the maximum baseline diameter Dmax . This was confirmed by a Spearman coefficient of 0.12 (P ¼ .31) that defined the correlation between the two frequency distributions shown in Fig 4. Correlation analysis. The correlations between the maximum diameter-based expansion rates g D , the volume-based expansion rates g V , and the multiple centerline-based growth measures g MAX and g MEAN are shown in Fig 5. These variables were distributed normally, and the Pearson correlation was calculated. Specifically, poor correlation was found between g MAX and g D (r ¼ 0.43; P < .001) as well as between g MAX and g V (r ¼ 0.34; P < .001), and g MAX and g MEAN (r ¼ 0.42; P < .001), which means that measuring the maximum diameters or the volumes at two time points cannot be used to quantify the maximum diameter growth. In contrast, a correlation was seen between g MEAN and g D (r ¼ 0.69; P < .001) as well as between g MEAN and g V (r ¼ 0.77; P < .001), indicating that measuring the maximum diameters or the volumes at two time points allows an estimation of the mean diameter growth. Moreover, no correlation was found between g MAX and g S (r ¼ 0.13; P ¼ .26) or between g MEAN and g S (r ¼ 0.06; P ¼ .64). Aortic aneurysm neck evolution. There were no significance differences (P ¼ .99) between the diameters
Fig 4. Locations of maximum diameter growth ðg MAX Þ and baseline Þ as a function of the relative maximum baseline diameter ðDmax centerline position. 0 corresponds with the renal arteries location and 1 with the aortic bifurcation position.
D1 at the lowest renal artery at baseline (median, 23.93 mm; IQR, 21.74 to 27.37 mm) and at follow-up (median, 24.16 mm; IQR, 22.02 to 27.02 mm). The mean diameter growth of 2.4%/y (IQR, 3.7%/y to 6.8%/y) and the maximum diameter growth of 6.9%/y (IQR, 0%/y to 13%/y) defined the enlargement of the neck. The aortic neck length N shrank slightly over time, from 17.53 mm (IQR, 11.00 to 29.12 mm) at baseline to 15.38 mm (IQR, 9.41 to 26.73 mm) at follow-up, which statistically was not a significant difference (P ¼ .12). It corresponds to a median longitudinal growth of 6.2%/y (IQR, 33%/y to 22.3%/y; Fig 6). A slight increase in aortic neck angulations a was observed, from 28.62 (IQR, 20.10 to 39.98 ) at baseline to 29.48 (IQR, 20.78 to 43.54 ) at follow-up, which was not statistically significant (P ¼ .32). It corresponds to a median rate of 2.134 /y (IQR, 2.65 /y to 7.30 /y; Fig 6).
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Martufi et al 753
Fig 5. Association of maximum ðg MAX Þ and mean (g MEAN ) diameter growth with global maximum diameter-based expansion rate ðg D Þ and global volume-based aneurysm expansion ðg V Þ with Pearson correlation.
Fig 6. Box and whisker plots of aortic aneurysm neck morphology change during the follow-up period. The horizontal line in the middle of each box indicates the median, the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the top whiskers mark the 75th percentile þ1.5 interquartile range (IQR), while the bottom whiskers mark the 25th percentile 1.5 IQR, and the þ symbols denote outliers.
DISCUSSION Fast growth of the aneurysm wall compromises its structural integrity,7-10 which naturally increases the risk for AAA rupture. Consequently, monitoring the growth of small AAAs is critical to avoid rupture in surveillance programs, for instance. AAA growth is currently monitored by measuring the development of its maximum diameter
over time. The present study showed that aneurysms did not grow fastest at the site of the maximum diameter but randomly anywhere along the aneurysmatic sac. Thus, monitoring the development of the maximum diameter is not able to measure the maximum diameter growth of the aneurysm. Similarly, monitoring the volume of an AAA cannot predict its maximum growth. Although volume measurements account to some extent for longitudinal growth, small focal areas of fast growth cannot be distinguished from larger areas of slow growth. In contrast, monitoring the development of the maximum diameter or the volume over time provides adequate measurements for the aneurysm’s average growth rate. Here, volume measurements are slightly better than maximum diameter measurements, indicating that volume might be a more representative parameter to reflect morphologic aneurysm changes, as suggested earlier.26 However, the complex shape of aneurysms is defined by strong local changes in surface curvature,27 which naturally cannot be described by single parameters such as the maximum diameter or the volume. The maximum diameter growth of about 16%/y in our cohort was almost four times larger than the mean diameter expansion of 4.4%/y. Apart from diameter growth, the aneurysms also grew considerably in the longitudinal direction by 3.5%/y. This observation questions monitoring methods that are related to certain cross-sections. Because the aneurysm neck length shrank only minimally over time (1.09 mm/y), the measured longitudinal growth translates into (further) bending of the aneurysm. The observed absolute mean diameter growth over 1 year of 2.05 mm/y is within the margin of error for
JOURNAL OF VASCULAR SURGERY September 2013
754 Martufi et al
ultrasound measurement methods, which makes this method problematic when monitoring small AAAs. One reason for its poor accuracy is the 2D nature of standard ultrasound measurements, and analyzing a 3D geometric model naturally improves the accuracy. The present study used a general geometric analysis of 3D aneurysm models based on CTA data, which basically would be clinically applicable. However, generating CTA data is not only more expensive than ultrasound imaging but also exposes patients to ionizing radiation and intravenous contrast, which might not be suitable for all patients. If CTA scanning is contraindicated, magnetic resonance angiography or 3D ultrasound imaging could provide alternative input data for the proposed multidimensional AAA analysis. The multiple centerline-based diameter measurements applied in this study also provided a comprehensive picture of the anatomic changes of the aneurysm neck over time, which could influence EVAR treatment decisions. The aortic diameter at the lowest renal artery was 15% larger than in a healthy population of the same age and gender ratio28 and expanded at 2.4% per year, which is about 2.5 times faster than for the normal aorta.28 This result might indicate that the neck tissue too is already aneurysmatically altered. Finally, the neck length shortened by an average of 6.2% per year, which was accompanied by a slight increase in neck angulation. These results are confirmed by earlier studies reporting that despite anatomic changes in aortic aneurysm shape, suitability for EVAR and clinical management do not significantly change at midterm follow-up.29 CTA is not the standard follow-up modality for small AAAs, which explains the short follow-up time and the small number of patients in this study. This must also be seen as the main study limitation, and a prospective study with at least 3 years of follow-up and larger sample size that takes into account all the baseline characteristics of patients and AAAs would be required to get meaningful information for clinical practice, as reported in the literature.17 Regardless of this shortcoming, our results can be compared with previous results. On the basis of an average growth rate of 4.4%/y from our cohort, a 40-mm AAA would breach 55 mm at the time t ¼ Ln[55/40]/Ln [1.044] ¼ 7.4 years. Similarly, considering the quartile of fast growers in our cohort, we can estimate that 25% of 40-mm AAAs would breach 55 mm within <4.5 years. This result correlates nicely with earlier reports.6,30 On top of this “average growth analyses,” one might also consider patient individual information to assess the development of the diameter (based on average and maximum diameter growth) as well as neck length and angulation development. CONCLUSIONS Clinical trials indicate that surveillance is a safe, costeffective way of managing small AAAs,1,31-34 with annual or less frequent surveillance intervals for all AAAs <45
mm in diameter.17 However, neither maximum diameter nor volume measurements over time are able to measure the fastest diameter growth of the aneurysmatic sac. As a consequence, expansion-related wall weakening might be inappropriately reflected by this type of surveillance data. In contrast, localized spots of fast diameter growth can be detected through multiple centerline-based diameter measurements over the whole aneurysm sac. This information might further reinforce the quality of aneurysm surveillance programs. AUTHOR CONTRIBUTIONS Conception and design: GM, JR, JS, TG Analysis and interpretation: GM, TG Data collection: GM, MA, JR, NS, GP Writing the article: GM, TG Critical revision of the article: JR, JS, NS, GP, TG Final approval of the article: GM, MA, JR, JS, NS, GP, TG Statistical analysis: GM Obtained funding: JS, NS, TG Overall responsibility: TG REFERENCES 1. UK Small Aneurysm Trial Participants. Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet 1998;352: 1649-55. 2. Limet R, Sakalihasan N, Albert A. Determination of the expansion rate and the incidence of rupture of abdominal aortic aneurysms. J Vasc Surg 1991;14:540-8. 3. Lederle FA, Johnson GR, Wilson SE, Ballard DJ, Jordan WD Jr, Blebea J, et al Veterans Affairs Cooperative Study #417 Investigators. Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA 2002;287:2968-72. 4. Brown PM, Zelt DT, Sobolev B. The risk of rupture in untreated aneurysms: the impact of size, gender, and expansion rate. J Vasc Surg 2003;37:280-4. 5. Brewster DC, Cronenwett JL, Hallett JW, Johnston KW, Krupski WC, Matsumura JS. Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 2003;37:1106-17. 6. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 2002;346:1437-44. 7. Freestone T, Turner RJ, Coady A, Higman DJ, Greenhalgh RM, Powell JT. Inflammation and matrix metalloproteinases in the enlarging abdominal aortic aneurysm. Arterioscler Thromb Vasc Biol 1995;15:1145-51. 8. Anidjar S, Dobrin PB, Eichorst M, Graham GP, Chejfec G. Correlation of inflammatory infiltrate with the enlargement of experimental aortic aneurysm. J Vasc Surg 1992;16:139-47. 9. McMillan WD, Tamarina NA, Cipollone M, Johnson DA, Parker MA, Pearce WH. Size matters: the relationship between MMP-9 expression and aortic diameter. Circulation 1997;96:2228-32. 10. Vallabhaneni SR, Gilling-Smith GL, How TV, Carter SD, Brennan JA, Harris PL. Heterogeneity of tensile strength and matrix metalloproteinase activity in the wall of abdominal aortic aneurysms. J Endovasc Ther 2004;11:494-502. 11. Fillinger MF, Marra SP, Raghavan ML, Kennedy FE. Prediction of rupture risk in Abdominal Aortic Aneurysm during observation: wall stress versus diameter. J Vasc Surg 2003;37:724-32. 12. Gasser TC, Auer M, Labruto F, Swedenborg J, Roy J. Biomechanical rupture risk assessment of abdominal aortic aneurysms. Model
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
complexity versus predictability of finite element simulations. Eur J Vasc Endovasc Surg 2010;40:176-85. Maier A, Gee MW, Reeps C, Pongratz J, Eckstein HH, Wall WA. A comparison of diameter, wall stress, and rupture potential index for abdominal aortic aneurysm rupture risk prediction. Ann Biomed Eng 2010;38:3124-34. Gasser TC. Hard data about AAA. Can we predict the risk of rupture according to AAA morphology? Controversies and updates in vascular surgery (CACVS), Paris, January 19-21, 2012. van Herwaarden JA, Bartels LW, Muhs BE, Vincken KL, Lindeboom MY, Teutelink A, et al. Dynamic magnetic resonance angiography of the aneurysm neck: conformational changes during the cardiac cycle with possible consequences for endograft sizing and future design. J Vasc Surg 2006;44:22-8. van Herwaarden JA, van de Pavoordt ED, Waasdorp EJ, Albert Vos J, Overtoom TT, Kelder JC, et al. Long-term single-center results with AneuRx endografts for endovascular abdominal aortic aneurysm repair. J Endovasc Ther 2007;14:307-17. Brady AR, Thompson SG, Fowkes FG, Greenhalgh RM, Powell JT; Participants UKSAT. Abdominal aortic aneurysm expansion: risk factors and time intervals for surveillance. Circulation 2004;110: 16-21. Kurvers H, Veith FJ, Lipsitz RC, Ohki T, Gargiulo NJ, Cayne NS, et al. Discontinuous, staccato growth of abdominal aortic aneurysms. J Am Coll Surg 2004;199:709-15. Vega de Céniga M, Gomez R, Estallo L, de la Fuente N, Viviens B, Barba A. Analysis of expansion patterns in 4-4.9 cm abdominal aortic aneurysms. Ann Vasc Surg 2008;22:37-44. Norman P, Spencer CA, Lawrence-Brown MM, Jamrozik K. C-reactive protein levels and the expansion of screen-detected abdominal aortic aneurysms in men. Circulation 2004;110:862-6. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Variability in measurement of abdominal aortic aneurysms. Abdominal Aortic Aneurysm Detection and Management Veterans Administration Cooperative Study Group. J Vasc Surg 1995;21:945-52. Auer M, Gasser TC. Reconstruction and finite element mesh generation of abdominal aortic aneurysms from computerized tomography angiography data with minimal user interactions. IEEE Trans Med Imaging 2010;29:1022-8. Chaikof EL, Blankensteijn JD, Harris PL, White GH, Zarins CK, Bernhard VM, et al. Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2002;35:1048-60.
Martufi et al 755
24. Schanzer A, Greenberg RK, Hevelone N, Robinson WP, Eslami MH, Goldberg RJ, Messina L. Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair. Circulation 2011;23: 2848-55. 25. Solberg S, Singh K, Wilsgaard T, Jacobsen BK. Increased growth rate of abdominal aortic aneurysms in women. The Tromsø study. Eur J Vasc Endovasc Surg 2005;29:145-9. 26. Armon MP, Yusuf SW, Whitaker SC, Gregson RHS, Wenham PW, Hopkinson BR. Thrombus distribution and changes in aneurysm size following endovascular aortic aneurysm repair. Eur J Vasc Endovasc Surg 1998;16:472-6. 27. Martufi G, Di Martino ES, Amon CH, Muluk SC, Finol EA. Threedimensional geometrical characterization of abdominal aortic aneurysms: image-based wall thickness distribution. J Biomech Eng 2009;131:061015-1-11. 28. Sonesson B, Hansen F, Stale H, Lanne T. Compliance and diameter in the human abdominal aorta. The influence of age and sex. Eur J Vasc Endovasc Surg 1993;7:690-7. 29. Yau FS, Rosero EB, Clagett GP, Valentine RJ, Modrall GJ, Arko FR, Timaran CH. Surveillance of small aortic aneurysms does not alter anatomic suitability for endovascular repair. J Vasc Surg 2007;45: 96-100. 30. Vega de Ceniga M, Gomez R, Estallo L, Rodriguez L, Baquer M, Barba A. Growth rate and associated factors in small abdominal aortic aneurysms. Eur J Vasc Endovasc Surg 2006;31:231-6. 31. UK Small Aneurysm Trial Participants. Health service costs and quality of life for early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet 1998;352:1656-60. 32. Multicentre Aneurysm Screening Study Group. The Multicentre Aneurysm Screening Study (MASS) into the effect of abdominal aortic aneurysm screening on mortality in men: a randomised controlled trial. Lancet 2002;360:1531-9. 33. UK Small Aneurysm Trial Participants. Long-term outcomes of immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 2002;346:1445-52. 34. Powell JT, Brown LC, Forbes JF, Fowkes FG, Greenhalgh RM, Ruckley CV, et al. Final 12-year follow-up of surgery versus surveillance in the UK Small Aneurysm Trial. Br J Surg 2007;94:702-8.
Submitted Aug 21, 2012; accepted Nov 17, 2012.
Multidimensional growth measurements of abdominal aortic aneurysms Giampaolo Martufi, PhD,a Martin Auer, PhD,b Joy Roy, MD, PhD,c Jesper Swedenborg, MD, PhD,c Natzi Sakalihasan, MD, PhD,d Giuseppe Panuccio, MD,e,f and T. Christian Gasser, PhD,a Stockholm, Sweden; Vienna, Austria; Liège, Belgium; Perugia, Italy; and Münster, Germany Background: Monitoring the expansion of abdominal aortic aneurysms (AAAs) is critical to avoid aneurysm rupture in surveillance programs, for instance. However, measuring the change of the maximum diameter over time can only provide limited information about AAA expansion. Specifically, regions of fast diameter growth may be missed, axial growth cannot be quantified, and shape changes of potential interest for decisions related to endovascular aneurysm repair cannot be captured. Methods: This study used multiple centerline-based diameter measurements between the renal arteries and the aortic bifurcation to quantify AAA growth in 51 patients from computed tomography angiography (CTA) data. Criteria for inclusion were at least 1 year of patient follow-up and the availability of at least two sufficiently high-resolution CTA scans that allowed an accurate three-dimensional reconstruction. Consequently, 124 CTA scans were systematically analyzed by using A4clinics diagnostic software (VASCOPS GmbH, Graz, Austria), and aneurysm growth was monitored at 100 cross-sections perpendicular to the centerline. Results: Monitoring diameter development over the entire aneurysm revealed the sites of the fastest diameter growth, quantified the axial growth, and showed the evolution of the neck morphology over time. Monitoring the development of an aneurysm’s maximum diameter or its volume over time can assess the mean diameter growth (r [ 0.69, r [ 0.77) but not the maximum diameter growth (r [ 0.43, r [ 0.34). The diameter growth measured at the site of maximum expansion was w16%/y, almost four times larger than the mean diameter expansion of 4.4%/y. The sites at which the maximum diameter growth was recorded did not coincide with the position of the maximum baseline diameter (r [ 0 .12; P [ .31). The overall aneurysm sac length increased from 84 to 89 mm during the follow-up (P < .001), which relates to the median longitudinal growth of 3.5%/y. The neck length shortened, on average, by 6.2% per year and was accompanied by a slight increase in neck angulation. Conclusions: Neither maximum diameter nor volume measurements over time are able to measure the fastest diameter growth of the aneurysm sac. Consequently, expansion-related wall weakening might be inappropriately reflected by this type of surveillance data. In contrast, localized spots of fast diameter growth can be detected through multiple centerlinebased diameter measurements over the entire aneurysm sac. This information might further reinforce the quality of aneurysm surveillance programs. (J Vasc Surg 2013;58:748-55.) Clinical Relevance: Although the size of the aneurysm still remains the most accepted predictor of rupture risk, small abdominal aortic aneurysms (AAAs) may also rupture. Thus, the new tools presented in this study use three-dimensional aneurysm models, based on computed tomography angiography and multiple centerline-based diameter measurements over the entire aneurysm sac, could help clinicians when making decisions to treat AAAs and when choosing alternative treatments for AAAs, regardless of the diameter size. From the Department of Solid Mechanics, School of Engineering Sciences, Royal Institute of Technology (KTH), Stockholma; Vascops GmbH, Development, Viennab; the Department of Vascular Surgery, Karolinska University Hospital and Institute, Stockholmc; the Department of Vascular Surgery, University Hospital of Liege, Lièged; the Division of Vascular and Endovascular Surgery, University of Perugia, Perugiae; and the Clinic for Vascular and Endovascular Surgery, Münster University Hospital, Münster.f Author conflict of interest: none. Reprint requests: Dr Giampaolo Martufi, Royal Institute of Technology (KTH), Department of Solid Mechanics, School of Engineering Sciences, Teknikringen 8, 100 44 Stockholm. Sweden (e-mail: martufi@kth.se or http://130.237.24.173/vascumech/index.html). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214/$36.00 Copyright Ó 2013 by the Society for Vascular Surgery. http://dx.doi.org/10.1016/j.jvs.2012.11.070
748
The indication for elective abdominal aortic aneurysm (AAA) repair is strongly related to the aneurysm’s risk for rupture and is given for aneurysms that reach a maximum centerline-based diameter of 55 mm.1 However, several reports have demonstrated a rupture risk of AAAs sized <5 cm in diameter and also that the rate of expansion of the aneurysm is crucial for a certain number of AAAs, independent of their size.2-4 Therefore, a single threshold diameter is not appropriate for every patient, and the decision to perform elective AAA repair should ideally be individualized. Consequently, apart from aneurysm size, many other risk factors for aneurysm rupture have been suggested. For example, aneurysm shape, blood pressure, female gender, and other wall-weakening effects have been mentioned specifically.5
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Apart from static criteria (ie, risk factors evaluated at one time point), dynamic risk factors, such as the expansion rate, have also been suggested.2-4 Aneurysm expansion in common clinical practice is defined as the change in maximum aortic diameter over time, and an expansion rate of 10 mm/y is generally considered as an indication for elective repair.1,6 This practice is justified to some extent through an expected wall weakening effect in response to undesirable wall remodeling at fast AAA growth. Specifically, aneurysm expansion correlates with the density of the inflammatory cell,7,8 which in turn increases matrix metalloproteinase activity7,8 and leads to compromised wall strength.9,10 This biologic activity is localized, such that spots of increased expression and activation of matrix metalloproteinases might contribute to fast local aneurysm expansion.7,8 The biomechanical rupture risk assessment already integrates many static risk factors into indices, such as the peak wall stress11 and the peak wall rupture risk.12-14 Similarly, dynamic wall-weakening effect could potentially be integrated, which might further improve the biomechanical rupture risk assessment. Developing models that prescribe aneurysm wall weakening as a consequence of their expansion rate requires detailed growth information. Available data are currently based on measuring the maximum diameter at two points in time, which clearly cannot provide a comprehensive picture on how aneurysm size and shape change over time. For example, the morphology of the aneurysm neck is especially important to minimize complications related to endovascular aneurysm repair (EVAR),15,16 and its change over time can clearly not be studied from maximum diameter measurements. Measuring the expansion of AAAs also presents a number of methodologic challenges. First, the change in diameter over time is small, nonlinear, and hard to predict.17 It is characterized by periods of rapid growth, followed by quiescience.18,19 Aortic diameter measurements that use ultrasound imaging and, to a lesser extent, computed tomography angiography (CTA), have a margin of error of 2 to 3 mm,17,20,21 which is within the range of the annual expansion of many small AAAs. Finally, the development of treatment protocols based on AAA expansion is somewhat limited due to missing follow-up data for aneurysms that reach the AAA repair indication threshold. Measuring the maximum diameter at two points in time provides only limited information about aneurysm growth. Specifically, regions of fast diameter growth might be missed, axial growth cannot be quantified, and shape changes of potential interest for EVAR-related decisions cannot be captured. Motivated by these limitations, the principal aim of the present study was to explore and investigate the growth pattern of the entire AAA sac. A multidimensional analysis of AAA growth was applied to CTA follow-up data from 51 patients. The applied analysis method used multiple centerline-based diameter measurements between the renal arteries and the aortic bifurcation.
Martufi et al 749
Table I. Patient demographics and abdominal aortic aneurysm (AAA) details Variable Patients Age, years Sex Male Female Mean arterial pressure, mm Hg Surveillance time frame, months AAA diameter, mm Baseline Follow-up
No. or mean (range) 51 70 (49-85) 44 7 97 (90-146) 12 (2-24) 46.6 (34.7-56.4) 49.5 (37-63)
Sites with the highest diameter growth and axial growth, as well as the evolution of the neck morphology over time, could be extracted from the collected measurements. Finally, the derived growth information allowed a critical evaluation of conventional procedures to measure AAA growth. METHODS Cohort. Data for 51 patients with nonruptured infrarenal AAAs were retrospectively collected from the S. M. Misericordia Hospital Perugia, Italy (n ¼ 32), Karolinska University Hospital, Stockholm, Sweden (n¼13), and University Hospital Liege, Belgium (n ¼ 6). Patients were monitored between 2 and 24 months with repeated contrast-enhanced multislice CTA, where the median overall follow-up period was 12 months (interquartile range [IQR], 11 to 13 months). Some further patient data are given in Table I. Criteria for inclusion were at least 1 year of patient follow-up and the availability of at least two sufficiently high-resolution CTA scans that allowed accurate reconstruction at two different time points. CTA scans with strongly inhomogeneous lumen intensity were rejected. Other exclusion criteria were patients with chronic kidney disease, contrast agent intolerance, and symptomatic patients. Patient age, sex, blood pressure, and smoking status were taken from patient records. The collection and use of anonymized data from human subjects was approved by the local ethics committees. Aneurysm geometry representation. Aortic geometries between the renal arteries and the aortic bifurcation from 124 CTA scans were systematically analyzed by a single observer. The three-dimensional (3D) AAA geometry (Fig 1, A) was reconstructed using deformable (active) contour models22 that were implemented in the A4clinics 4.0 diagnostic software (VASCOPS GmbH, Graz, Austria). The accuracy of this type of geometric reconstruction is in the range of one in-plane pixel size.22 All reconstructions were verified by another author of the study. Reconstructed AAA geometry was used to automatically calculate the aneurysm volume (V); that is, the volume
750 Martufi et al
JOURNAL OF VASCULAR SURGERY September 2013
Fig 1. Abdominal aortic aneurysm (AAA) geometry acquisition. A, Three-dimensional anatomy is reconstructed from computed tomography angiography scans using a deformable segmentation model. B, Simplified geometry representation uses splines perpendicular to the aneurysm’s centerline.
Fig 2. Definition of the geometric properties of an abdominal aortic aneurysm (AAA). A, Centerline-based outer aortic diameters. Dmax, maximum diameter; D1, aortic diameter at the lowest renal artery. B, Centerline length segments. N, neck length; S, aneurysm sac length. C, Angulation a.
that is enclosed by the outer surface and between the transversal planes at the lowest renal artery and the aortic bifurcation, respectively. Then, the 3D model was automatically sliced orthogonally to the centerline so that the aneurysm’s luminal and outer surfaces were represented through splines (Fig 1, B). Specifically, 100 splines were used for each aneurysm, which together with the centerline, provided a simplified 3D representation. This geometric representation allowed a fully automatic analysis with a specially developed in-house algorithm using MatLab software (The MathWorks, Natick, Mass). Luminal and outer diameters were measured from each spline, which in turn defined the largest outer
diameter (Dmax) and the diameter (D1) at the lowest renal artery (Fig 2, A). The aneurysm neck length (N), the portion between the lowest renal artery and the proximal onset of the aneurysmatic enlargement (ie, where the outer diameter was 10% larger than the aortic diameter at the lowest renal artery [D1])23,24 was extracted (Fig 2, B). Similarly extracted was the aneurysm sac length (S), the length of the centerline from the proximal onset of the aneurysm down to the aortic bifurcation (Fig 2, B). Finally, the angulation (a) was computed (Fig 2, C); that is, the rotation that is required, such that the extrapolation of the straight centerline in the neck meets the aortic bifurcation.23,24
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Martufi et al 751
Aneurysm growth measurements. To quantify aneurysm growth, the changes of the geometric properties were compared between two consecutive CTA scans. Multiple comparisons were made for patients who had more than two CTA scans, and 7300 expansion observations were processed in the present study. Specifically, the nonlinear growth model g ¼
X 1-year followup 1 $100 X baseline
¼ ½Expð12 rÞ 1$100½%=y; where X baseline and X 1-year follow-up denote baseline and 1-year follow-up quantities, was used. Exp (d) is the exponential function of (d). Because the measured quantity X was not always given at the 1-year follow-up, the loga followup 1 X was used to rithmic growth factor r ¼ Ln X baseline t calculate g, where t denotes the time interval between two measurement points in months. Ln (d) is the natural logarithmic function of (d). To study the diameter expansion, g was computed at 100 cross-sections along the aneurysm. Specifically, the diameters that were at the same relative centerline position, between the two time points at which CTA scans were taken, were thought to correspond to each other. From these measurements, maximum g MAX and mean g MEAN diameter expansions were extracted. Standard maximum-diameter measures were used to define the global maximum diameter-based expansion 1 D followup which compares the rate g D , with r ¼ Ln max baseline Dmax t maximum diameters at baseline and follow-up. Note that baseline followup and Dmax are, in the maximum diameters Dmax general, not at the same relative centerline position. Finally, volume measurements were used to define global volume-based aneurysm expansion g V with followup 1 V r ¼ Ln ; and centerline length measurements V baseline t defined global longitudinal aneurysm expansion g S with followup 1 S . r ¼ Ln S baseline t Statistics. Statistical analysis was performed with SPSS 20.0 software (SPSS Inc, Chicago, Ill). The normality of the data was assessed using the Shapiro-Wilk test. The paired t-test was used for paired variables with normal distribution, and the Wilcoxon matched pairs test was used for non-normal data. For the null hypothesis test, a two-sided P < .05 was set to determine statistical significance. Correlations between the maximum diameter-based expansion rates ðg D Þ, the volume-based expansion rates ðg V Þ, and the multiple centerline-based growth measures of g MAX and g MEAN were assessed. This analysis aimed at quantifying the extent to which aneurysm growth could be extracted from measuring the maximum diameters or
Fig 3. Box and whisker plots show multiple centerline-based aneurysm expansions and global-based aneurysm expansions. g MAX , maximum diameter growth; g MEAN , mean diameter growth, which is averaged between the renal arteries and the aortic bifurcation; g D , global maximum diameter-based expansion rate; g S , global longitudinal aneurysm expansion; g V , global volume-based aneurysm expansion. The horizontal line in the middle of each box indicates the median, the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the top whiskers mark the 75th percentile þ1.5 interquartile range (IQR), while the bottom whiskers mark the 25th percentile 1.5 IQR, and the þ symbols denote outliers.
the volumes at two time points. In addition, the correlation between the site of the maximum diameter and the site of maximum growth was investigated. RESULTS AAA growth measurements. The calculated growth measurements are shown in Fig 3. The maximum diameter growth g MAX (median, 16%/y; IQR, 12.3%/y to 23%/y) was about four times larger than the mean diameter growth g MEAN (median, 4.4%/y; IQR, 2%/y to 7.4%/y). The maximum diameter growth g MAX differed significantly (P < .001) from g D (median, 6.4%/y; IQR, 3.5%/y to 8.6%/y) and g V (median, 10.7%/y; IQR, 7.7%/y to 17%/y); that is, the growth based on relating the maximum diameters or the volumes measured at the two time points, respectively. The overall aneurysm sac length (S) increased (P < .001) from 85.37 mm (IQR, 74.3 to 95.7 mm) to 89.44 mm (IQR, 81.55 to 101.75 mm) during the follow-up, which relates to a median longitudinal growth ðg S Þ of 3.5%/y (IQR, 0.84% to 11.5%). The subgroup analysis reported in Table II found that AAA expansion was not associated with age and mean arterial pressure but were increased in current smokers, women, and larger baseline diameters, similar to previous reports.17,25 However, none of the subgroup results reported in Table II reached statistical significance, probably due to the small sample size. The relative centerline position, at which the maximum diameter growth g MAX was recorded, did not coincide with
JOURNAL OF VASCULAR SURGERY September 2013
752 Martufi et al
Table II. Measured abdominal aortic aneurysm (AAA) growth by patient characteristics at baseline and risk factors AAA growth measurements, % per year (median) Factor Age at baseline, years (tertiles) 49-69 70-74 75-85 Sex Male Female baseline , mm (tertiles) Dmax 35-45 46-48 48-56 Smoking status Never Former Current Unknown MAP at baseline, mm Hg (tertiles) 90-93 94-98 99-147
Pts, No. g MAX g MEAN g D
gV
gS
20 17 14
16.7 22.4 16.8
5 5.8 4
6.1 12.9 7.8 11.7 5.2 10.7
4.3 3.24 3.79
44 7
16.4 20.9
4.6 5.2
5.6 11.8 2.8 7.9 16.9 10.2
17 17 17
15.7 18.3 18.9
3.2 4.7 5.9
4 10.3 7 10.4 7.2 16.7
2.45 3.44 5.9
20 9 11 11
17.2 17.5 22.9 16.4
3.9 4.8 5.6 4.2
4.9 6.1 6.7 7
10.6 12.7 16.7 13.8
1.33 1.97 4.7 1.53
30 4 17
16.6 18.4 18.7
4.7 5.3 4.7
6.4 11.2 4.9 14.6 7.5 14
4 3 3.45
baseline Dmax , Maximum diameter at baseline; g MAX , maximum diameter growth; g MEAN , mean diameter growth; g D , global maximum diameter-based expansion rate; g S , global longitudinal aneurysm expansion; g V , global volume-based aneurysm expansion; MAP, mean arterial pressure.
baseline the position of the maximum baseline diameter Dmax . This was confirmed by a Spearman coefficient of 0.12 (P ¼ .31) that defined the correlation between the two frequency distributions shown in Fig 4. Correlation analysis. The correlations between the maximum diameter-based expansion rates g D , the volume-based expansion rates g V , and the multiple centerline-based growth measures g MAX and g MEAN are shown in Fig 5. These variables were distributed normally, and the Pearson correlation was calculated. Specifically, poor correlation was found between g MAX and g D (r ¼ 0.43; P < .001) as well as between g MAX and g V (r ¼ 0.34; P < .001), and g MAX and g MEAN (r ¼ 0.42; P < .001), which means that measuring the maximum diameters or the volumes at two time points cannot be used to quantify the maximum diameter growth. In contrast, a correlation was seen between g MEAN and g D (r ¼ 0.69; P < .001) as well as between g MEAN and g V (r ¼ 0.77; P < .001), indicating that measuring the maximum diameters or the volumes at two time points allows an estimation of the mean diameter growth. Moreover, no correlation was found between g MAX and g S (r ¼ 0.13; P ¼ .26) or between g MEAN and g S (r ¼ 0.06; P ¼ .64). Aortic aneurysm neck evolution. There were no significance differences (P ¼ .99) between the diameters
Fig 4. Locations of maximum diameter growth ðg MAX Þ and baseline Þ as a function of the relative maximum baseline diameter ðDmax centerline position. 0 corresponds with the renal arteries location and 1 with the aortic bifurcation position.
D1 at the lowest renal artery at baseline (median, 23.93 mm; IQR, 21.74 to 27.37 mm) and at follow-up (median, 24.16 mm; IQR, 22.02 to 27.02 mm). The mean diameter growth of 2.4%/y (IQR, 3.7%/y to 6.8%/y) and the maximum diameter growth of 6.9%/y (IQR, 0%/y to 13%/y) defined the enlargement of the neck. The aortic neck length N shrank slightly over time, from 17.53 mm (IQR, 11.00 to 29.12 mm) at baseline to 15.38 mm (IQR, 9.41 to 26.73 mm) at follow-up, which statistically was not a significant difference (P ¼ .12). It corresponds to a median longitudinal growth of 6.2%/y (IQR, 33%/y to 22.3%/y; Fig 6). A slight increase in aortic neck angulations a was observed, from 28.62 (IQR, 20.10 to 39.98 ) at baseline to 29.48 (IQR, 20.78 to 43.54 ) at follow-up, which was not statistically significant (P ¼ .32). It corresponds to a median rate of 2.134 /y (IQR, 2.65 /y to 7.30 /y; Fig 6).
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
Martufi et al 753
Fig 5. Association of maximum ðg MAX Þ and mean (g MEAN ) diameter growth with global maximum diameter-based expansion rate ðg D Þ and global volume-based aneurysm expansion ðg V Þ with Pearson correlation.
Fig 6. Box and whisker plots of aortic aneurysm neck morphology change during the follow-up period. The horizontal line in the middle of each box indicates the median, the top and bottom borders of the box mark the 75th and 25th percentiles, respectively, the top whiskers mark the 75th percentile þ1.5 interquartile range (IQR), while the bottom whiskers mark the 25th percentile 1.5 IQR, and the þ symbols denote outliers.
DISCUSSION Fast growth of the aneurysm wall compromises its structural integrity,7-10 which naturally increases the risk for AAA rupture. Consequently, monitoring the growth of small AAAs is critical to avoid rupture in surveillance programs, for instance. AAA growth is currently monitored by measuring the development of its maximum diameter
over time. The present study showed that aneurysms did not grow fastest at the site of the maximum diameter but randomly anywhere along the aneurysmatic sac. Thus, monitoring the development of the maximum diameter is not able to measure the maximum diameter growth of the aneurysm. Similarly, monitoring the volume of an AAA cannot predict its maximum growth. Although volume measurements account to some extent for longitudinal growth, small focal areas of fast growth cannot be distinguished from larger areas of slow growth. In contrast, monitoring the development of the maximum diameter or the volume over time provides adequate measurements for the aneurysm’s average growth rate. Here, volume measurements are slightly better than maximum diameter measurements, indicating that volume might be a more representative parameter to reflect morphologic aneurysm changes, as suggested earlier.26 However, the complex shape of aneurysms is defined by strong local changes in surface curvature,27 which naturally cannot be described by single parameters such as the maximum diameter or the volume. The maximum diameter growth of about 16%/y in our cohort was almost four times larger than the mean diameter expansion of 4.4%/y. Apart from diameter growth, the aneurysms also grew considerably in the longitudinal direction by 3.5%/y. This observation questions monitoring methods that are related to certain cross-sections. Because the aneurysm neck length shrank only minimally over time (1.09 mm/y), the measured longitudinal growth translates into (further) bending of the aneurysm. The observed absolute mean diameter growth over 1 year of 2.05 mm/y is within the margin of error for
JOURNAL OF VASCULAR SURGERY September 2013
754 Martufi et al
ultrasound measurement methods, which makes this method problematic when monitoring small AAAs. One reason for its poor accuracy is the 2D nature of standard ultrasound measurements, and analyzing a 3D geometric model naturally improves the accuracy. The present study used a general geometric analysis of 3D aneurysm models based on CTA data, which basically would be clinically applicable. However, generating CTA data is not only more expensive than ultrasound imaging but also exposes patients to ionizing radiation and intravenous contrast, which might not be suitable for all patients. If CTA scanning is contraindicated, magnetic resonance angiography or 3D ultrasound imaging could provide alternative input data for the proposed multidimensional AAA analysis. The multiple centerline-based diameter measurements applied in this study also provided a comprehensive picture of the anatomic changes of the aneurysm neck over time, which could influence EVAR treatment decisions. The aortic diameter at the lowest renal artery was 15% larger than in a healthy population of the same age and gender ratio28 and expanded at 2.4% per year, which is about 2.5 times faster than for the normal aorta.28 This result might indicate that the neck tissue too is already aneurysmatically altered. Finally, the neck length shortened by an average of 6.2% per year, which was accompanied by a slight increase in neck angulation. These results are confirmed by earlier studies reporting that despite anatomic changes in aortic aneurysm shape, suitability for EVAR and clinical management do not significantly change at midterm follow-up.29 CTA is not the standard follow-up modality for small AAAs, which explains the short follow-up time and the small number of patients in this study. This must also be seen as the main study limitation, and a prospective study with at least 3 years of follow-up and larger sample size that takes into account all the baseline characteristics of patients and AAAs would be required to get meaningful information for clinical practice, as reported in the literature.17 Regardless of this shortcoming, our results can be compared with previous results. On the basis of an average growth rate of 4.4%/y from our cohort, a 40-mm AAA would breach 55 mm at the time t ¼ Ln[55/40]/Ln [1.044] ¼ 7.4 years. Similarly, considering the quartile of fast growers in our cohort, we can estimate that 25% of 40-mm AAAs would breach 55 mm within <4.5 years. This result correlates nicely with earlier reports.6,30 On top of this “average growth analyses,” one might also consider patient individual information to assess the development of the diameter (based on average and maximum diameter growth) as well as neck length and angulation development. CONCLUSIONS Clinical trials indicate that surveillance is a safe, costeffective way of managing small AAAs,1,31-34 with annual or less frequent surveillance intervals for all AAAs <45
mm in diameter.17 However, neither maximum diameter nor volume measurements over time are able to measure the fastest diameter growth of the aneurysmatic sac. As a consequence, expansion-related wall weakening might be inappropriately reflected by this type of surveillance data. In contrast, localized spots of fast diameter growth can be detected through multiple centerline-based diameter measurements over the whole aneurysm sac. This information might further reinforce the quality of aneurysm surveillance programs. AUTHOR CONTRIBUTIONS Conception and design: GM, JR, JS, TG Analysis and interpretation: GM, TG Data collection: GM, MA, JR, NS, GP Writing the article: GM, TG Critical revision of the article: JR, JS, NS, GP, TG Final approval of the article: GM, MA, JR, JS, NS, GP, TG Statistical analysis: GM Obtained funding: JS, NS, TG Overall responsibility: TG REFERENCES 1. UK Small Aneurysm Trial Participants. Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet 1998;352: 1649-55. 2. Limet R, Sakalihasan N, Albert A. Determination of the expansion rate and the incidence of rupture of abdominal aortic aneurysms. J Vasc Surg 1991;14:540-8. 3. Lederle FA, Johnson GR, Wilson SE, Ballard DJ, Jordan WD Jr, Blebea J, et al Veterans Affairs Cooperative Study #417 Investigators. Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA 2002;287:2968-72. 4. Brown PM, Zelt DT, Sobolev B. The risk of rupture in untreated aneurysms: the impact of size, gender, and expansion rate. J Vasc Surg 2003;37:280-4. 5. Brewster DC, Cronenwett JL, Hallett JW, Johnston KW, Krupski WC, Matsumura JS. Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 2003;37:1106-17. 6. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 2002;346:1437-44. 7. Freestone T, Turner RJ, Coady A, Higman DJ, Greenhalgh RM, Powell JT. Inflammation and matrix metalloproteinases in the enlarging abdominal aortic aneurysm. Arterioscler Thromb Vasc Biol 1995;15:1145-51. 8. Anidjar S, Dobrin PB, Eichorst M, Graham GP, Chejfec G. Correlation of inflammatory infiltrate with the enlargement of experimental aortic aneurysm. J Vasc Surg 1992;16:139-47. 9. McMillan WD, Tamarina NA, Cipollone M, Johnson DA, Parker MA, Pearce WH. Size matters: the relationship between MMP-9 expression and aortic diameter. Circulation 1997;96:2228-32. 10. Vallabhaneni SR, Gilling-Smith GL, How TV, Carter SD, Brennan JA, Harris PL. Heterogeneity of tensile strength and matrix metalloproteinase activity in the wall of abdominal aortic aneurysms. J Endovasc Ther 2004;11:494-502. 11. Fillinger MF, Marra SP, Raghavan ML, Kennedy FE. Prediction of rupture risk in Abdominal Aortic Aneurysm during observation: wall stress versus diameter. J Vasc Surg 2003;37:724-32. 12. Gasser TC, Auer M, Labruto F, Swedenborg J, Roy J. Biomechanical rupture risk assessment of abdominal aortic aneurysms. Model
JOURNAL OF VASCULAR SURGERY Volume 58, Number 3
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
complexity versus predictability of finite element simulations. Eur J Vasc Endovasc Surg 2010;40:176-85. Maier A, Gee MW, Reeps C, Pongratz J, Eckstein HH, Wall WA. A comparison of diameter, wall stress, and rupture potential index for abdominal aortic aneurysm rupture risk prediction. Ann Biomed Eng 2010;38:3124-34. Gasser TC. Hard data about AAA. Can we predict the risk of rupture according to AAA morphology? Controversies and updates in vascular surgery (CACVS), Paris, January 19-21, 2012. van Herwaarden JA, Bartels LW, Muhs BE, Vincken KL, Lindeboom MY, Teutelink A, et al. Dynamic magnetic resonance angiography of the aneurysm neck: conformational changes during the cardiac cycle with possible consequences for endograft sizing and future design. J Vasc Surg 2006;44:22-8. van Herwaarden JA, van de Pavoordt ED, Waasdorp EJ, Albert Vos J, Overtoom TT, Kelder JC, et al. Long-term single-center results with AneuRx endografts for endovascular abdominal aortic aneurysm repair. J Endovasc Ther 2007;14:307-17. Brady AR, Thompson SG, Fowkes FG, Greenhalgh RM, Powell JT; Participants UKSAT. Abdominal aortic aneurysm expansion: risk factors and time intervals for surveillance. Circulation 2004;110: 16-21. Kurvers H, Veith FJ, Lipsitz RC, Ohki T, Gargiulo NJ, Cayne NS, et al. Discontinuous, staccato growth of abdominal aortic aneurysms. J Am Coll Surg 2004;199:709-15. Vega de Céniga M, Gomez R, Estallo L, de la Fuente N, Viviens B, Barba A. Analysis of expansion patterns in 4-4.9 cm abdominal aortic aneurysms. Ann Vasc Surg 2008;22:37-44. Norman P, Spencer CA, Lawrence-Brown MM, Jamrozik K. C-reactive protein levels and the expansion of screen-detected abdominal aortic aneurysms in men. Circulation 2004;110:862-6. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Variability in measurement of abdominal aortic aneurysms. Abdominal Aortic Aneurysm Detection and Management Veterans Administration Cooperative Study Group. J Vasc Surg 1995;21:945-52. Auer M, Gasser TC. Reconstruction and finite element mesh generation of abdominal aortic aneurysms from computerized tomography angiography data with minimal user interactions. IEEE Trans Med Imaging 2010;29:1022-8. Chaikof EL, Blankensteijn JD, Harris PL, White GH, Zarins CK, Bernhard VM, et al. Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2002;35:1048-60.
Martufi et al 755
24. Schanzer A, Greenberg RK, Hevelone N, Robinson WP, Eslami MH, Goldberg RJ, Messina L. Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair. Circulation 2011;23: 2848-55. 25. Solberg S, Singh K, Wilsgaard T, Jacobsen BK. Increased growth rate of abdominal aortic aneurysms in women. The Tromsø study. Eur J Vasc Endovasc Surg 2005;29:145-9. 26. Armon MP, Yusuf SW, Whitaker SC, Gregson RHS, Wenham PW, Hopkinson BR. Thrombus distribution and changes in aneurysm size following endovascular aortic aneurysm repair. Eur J Vasc Endovasc Surg 1998;16:472-6. 27. Martufi G, Di Martino ES, Amon CH, Muluk SC, Finol EA. Threedimensional geometrical characterization of abdominal aortic aneurysms: image-based wall thickness distribution. J Biomech Eng 2009;131:061015-1-11. 28. Sonesson B, Hansen F, Stale H, Lanne T. Compliance and diameter in the human abdominal aorta. The influence of age and sex. Eur J Vasc Endovasc Surg 1993;7:690-7. 29. Yau FS, Rosero EB, Clagett GP, Valentine RJ, Modrall GJ, Arko FR, Timaran CH. Surveillance of small aortic aneurysms does not alter anatomic suitability for endovascular repair. J Vasc Surg 2007;45: 96-100. 30. Vega de Ceniga M, Gomez R, Estallo L, Rodriguez L, Baquer M, Barba A. Growth rate and associated factors in small abdominal aortic aneurysms. Eur J Vasc Endovasc Surg 2006;31:231-6. 31. UK Small Aneurysm Trial Participants. Health service costs and quality of life for early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet 1998;352:1656-60. 32. Multicentre Aneurysm Screening Study Group. The Multicentre Aneurysm Screening Study (MASS) into the effect of abdominal aortic aneurysm screening on mortality in men: a randomised controlled trial. Lancet 2002;360:1531-9. 33. UK Small Aneurysm Trial Participants. Long-term outcomes of immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 2002;346:1445-52. 34. Powell JT, Brown LC, Forbes JF, Fowkes FG, Greenhalgh RM, Ruckley CV, et al. Final 12-year follow-up of surgery versus surveillance in the UK Small Aneurysm Trial. Br J Surg 2007;94:702-8.
Submitted Aug 21, 2012; accepted Nov 17, 2012.