Clinical Significance of Increased Computed Tomography Attenuation of Periaortic Adipose Tissue in Patients With Abdominal Aortic Aneurysms

Abstract

Background: Recent imaging studies reported an association between vascular inflammation and progression of abdominal aortic aneurysm (AAA). This study investigated the clinical significance of periaortic adipose tissue inflammation derived from multidetector computed tomography angiography (MDCTA).

Methods and Results: Patients with asymptomatic AAA (n=77) who underwent an index and >6 months follow-up MDCTA examinations were retrospectively investigated. MDCTA analysis included AAA diameter and the periaortic adipose tissue attenuation index (PAAI). The PAAI was defined as the mean CT attenuation value within a predefined range from −190 to −30 Hounsfield units of adipose tissue surrounding the AAA. The growth rate of the AAA was calculated as the change in diameter. AAA progression (AP) was defined as an AAA growth rate ≥5 mm/year. Univariate and multivariate logistic regression analysis were performed to determine the predictors of AP. AP was observed in 19 patients (24.7%), the median baseline AAA diameter was 38.9 mm (interquartile range [IQR] 32.7–42.9 mm), and the median growth rate was 3.1 mm/year (IQR 1.5–4.9 mm/year). Baseline AAA diameter (odds ratio [OR] 1.16; 95% confidence interval [CI] 1.05–1.28; P=0.001) and PAAI (OR 1.12; 95% CI 1.05–1.20; P=0.004) were independent predictors of AP.

Conclusions: PAAI was an independent and significant predictor of AP, supporting the notion that local adipose tissue inflammation may contribute to aortic remodeling.

Abdominal aortic aneurysm (AAA) is characterized by an abnormal enlargement of the abdominal aorta and is diagnosed clinically when the diameter is >30 mm or >50% greater than the normal diameter. One of the potential dire consequences of advanced AAA is rupture of the aneurysm. Risk factors for AAA rupture have been intensively explored, with baseline AAA diameter being the best-known predictor of rupture.^1–3 An AAA diameter >50 mm is recognized as a threshold for surgical repair given the high rates of rupture with conservative observation.⁴ Nevertheless, the pathophysiological mechanisms related to the progression of AAA have not been fully elucidated. Vascular inflammation with subsequent proteolytic degradation of the aortic wall have been thought to play an important role in aortic wall weakening, leading to the progression of AAA.^5–7 In pathological examinations, the aortic wall in the aneurysm is characterized by inflammatory changes, including focal medial neovascularization, infiltration of inflammatory cells, and fragmentation of elastin and collagen fibers within the extracellular matrix.⁷ In addition, perivascular adipose tissue (PAT) interacting with vascular inflammation has been reported to potentially contribute to the progression of aneurysm.⁸

Editorial p 2181

Recent years have witnessed significant advances in imaging modalities for evaluation of vascular inflammation in AAA and its association with the progression of aneurysm. ¹⁸F-Fluorodeoxyglucose (FDG) positron emission tomography (PET) and magnetic resonance imaging (MRI) have been used to assess the susceptibility of the aortic wall to aneurysm formation.^7,9,10 Using computed tomography (CT), the standard modality for morphological assessment of AAA, a significant association between periaortic adipose tissue volume and aortic dimension has been reported.¹¹ However, CT-derived periaortic inflammation around an AAA and its clinical significance have not been investigated in depth. Recently, the assessment of perivascular inflammation using multidetector CT angiography (MDCTA) has been reported in coronary arteries.^12,13 Of note, Antonopoulos et al demonstrated that the mean attenuation of adipose tissue, defined as the fat attenuation index (FAI), was associated with adipocyte differentiation and activated macrophages on histology and ¹⁸F-FDG uptake on PET imaging.¹² In this context, periaortic adipose tissue and vascular inflammation of AAA have also been thought to play important key roles in the progression of aneurysms, which imaging modalities have targeted. However, the mean attenuation on MDCTA of adipose tissue around the aorta has not been tested as a potential predictor of AAA progression. Therefore, the aim of this study was to investigate the clinical significance of the mean attenuation of periaortic adipose tissue, as assessed by MDCTA, in patients with AAA.

Methods

Study Population

The institutional database of 320-slice MDCTAs performed in Tsuchiura Kyodo General Hospital between December 2013 and February 2019 was reviewed retrospectively to identify patients who met the following inclusion criteria, namely patients who underwent an index MDCTA for the first assessment of AAA who also had follow-up MDCTA at >6 months after the index examination. In all, 448 patients with AAA who underwent the index and follow-up MDCTAs were identified for the analysis. Thereafter, we excluded patients with a history of any surgical treatment, including endovascular and open surgical repair (n=316), a history of aortic dissection (n=32), renal insufficiency (i.e., baseline serum creatinine concentration >1.5 mg/dL; n=15), or insufficient MDCTA image quality (n=8). This left 77 patients in the final dataset for the analysis (Figure 1).

Figure 1.

Study population. In all, 448 patients with abdominal aortic aneurysm (AAA) who underwent the index and follow-up multidetector computed tomography angiography (MDCTA) examinations >6 months apart were selected from the institutional database. After applying exclusion criteria, 77 patients were left for analysis in the present study. Patients were divided into 2 groups based on AAA growth rate (≥5.0 and <5.0 mm/year) from the index to follow-up MDCTA, namely progression (n=19) and non-progression (n=58) groups.

AAA was defined as a dilation of the abdominal aorta with a diameter >30 mm. Baseline patient characteristics were collected by reviewing patients’ medical records.

Image Acquisition Using MDCTA

All patients underwent the index and follow-up MDCTAs using the same acquisition protocol and a 320-row scanner (Aquilion ONE; Toshiba Medical Systems, Otawara, Japan). The slice thickness was 0.5 mm with a reconstruction interval of 0.5 mm using a workstation (SYNAPSE VINCENT; Fujifilm, Tokyo, Japan). The scan was triggered using an automatic bolus-tracking technique with a region of interest placed in the ascending aorta. Images were acquired after bolus injection of contrast (iopamidol; 370 mg iodine/mL; Bayer Yakuhin, Osaka, Japan) at a rate of 5–6 mL/s with tube current modulation. Acquisition and reconstruction parameters were as follows: electroc120 kVp, 110–206 mAs, and 320 mm×0.5 mm slice collimation. All scans were performed during a single breath-hold.

MDCTA Image Analysis to Determine Aneurysm Size

All MDCTA analyses were performed offline using a dedicated workstation (Aquarius iNtuition Edition version 4.4.13; TeraRecon, Foster City, CA, USA) by an independent investigator who was blinded to the patients’ information and chronological sequence of the scans. Axial source images were reconstructed into curved multiplanar reconstruction images along with the center of the abdominal aorta. The major and minor diameters of the aorta were measured at 1-mm intervals through the aneurysm, with the largest value of the minor diameter defined as AAA diameter in the present study.

Assuming that the growth rate was linear and constant over the study period, the change in AAA diameter (mm) from the index to follow-up MDCTA was calculated. Annual AAA growth rate (mm/year) was defined as the change in AAA diameter (mm) divided by the duration (years) from the index to follow-up MDCTA examinations. Progression of AAA was defined as an annual AAA growth rate ≥5 mm/year.

MDCTA Analysis of Periaortic Adipose Tissue Inflammation

The periaortic adipose tissue attenuation index (PAAI) was determined around the AAA using a workstation based on the same methods used to determine the FAI around coronary arteries and validated previously.¹² However, given the significant differences in diameter and anatomy between the aorta and coronary arteries, we modified the methodology for determining the FAI in the coronary artery to determine the PAAI. Briefly, we tested 4 different regions of interest (ROI) using combinations of 2 different thicknesses (5 and 10 mm) of the cylindrical layer from the outer vessel wall and 2 different segments (renal and terminal) of the aorta to assess periaortic adipose tissue inflammation.

Although the previous study in coronary arteries examined perivascular tissue around the right coronary artery with a cylindrical layer thickness equal to the diameter of the coronary artery from the outer vessel wall,¹² in the present study we tested 2 different cylindrical layer thicknesses from the vessel wall, namely 5 and 10 mm, considering the larger lumen of the abdominal aorta compared with coronary arteries (Figure 2). Therefore, periaortic adipose tissue was defined as the adipose tissue located within a radial distance of 5 and 10 mm from the outer vessel wall.

Figure 2.

Regions of interest (ROI) for measurements of the periaortic adipose tissue attenuation index (PAAI). There were 4 different ROIs in this study based on combinations of 2 different aorta segments (renal and terminal) and 2 different thicknesses of the cylindrical layer from the outer vessel wall (5 and 10 mm) for assessment of the PAAI: PAAI-renal-5 mm, PAAI-renal-10 mm, PAAI-terminal-5 mm, and PAAI-terminal-10 mm.

Periaortic adipose tissue in the 5- and 10-mm layers was evaluated in two segments: (1) the infrarenal 100-mm segment distal to the ostium of the right renal artery (PAAI-renal-5 mm and PAAI-renal-10 mm); and (2) a 100-mm segment proximal to the terminal aorta (PAAI-terminal-5 mm and PAAI-terminal-10 mm; Figure 2). PAAI was measured semiautomatically using the dedicated workstation with additional minor manual optimization. PAAI was sampled for each ROI on 3-dimensional images, modifying the segment and thickness of the adipose tissue layer (Figure 3). For each ROI, voxel attenuation histograms were plotted and PAAI was calculated as the weighted mean attenuation of all voxels between −190 and −30 Hounsfield units (HU; Figure 3).^14,15 In the same manner, periaortic adipose tissue volume (PAAT volume) was determined using dedicated software based on the attenuation histogram of periaortic adipose tissue attenuation at the infrarenal and terminal aorta segments in the 5- and 10-mm layers (PAAT volume-renal-5 mm, PAAT volume-renal-10 mm, PAAT volume-terminal-5 mm, and PAAT volume-terminal-10 mm). In addition, to assess the heterogeneity of the value within ROIs, we determined the difference in PAAI values between the 5- and 10-mm layers in the renal and terminal segments (PAAI-renal-difference 5–10 mm and PAAI-terminal-difference 5–10 mm, respectively) by subtracting values for the 10-mm layer from those of the 5-mm layer:

Figure 3.

Representative multidetector computed tomography angiography (MDCTA) image of abdominal aortic aneurysm (AAA), (A–D) with and (E–H) without AAA progression. (A,E) Three-dimensional reconstruction of the AAA. (B,F) The periaortic adipose tissue attenuation index (PAAI)-terminal was sampled from a 100-mm segment proximal to the terminal aorta. (C,G) The PAAI-renal was sampled from a 100-mm segment distal to the ostium of the right renal artery in 3-dimensional layers of 5 and 10 mm from the outer vessel wall. (D,H) Histograms of voxel computed tomography (CT) attenuations within the volume of interest. The mean CT attenuation within the range between −190 and −30 HU, which was defined as PAAI-renal-5 mm, was −68.33 HU in the patient with AAA progression (D) and −86.10 HU in the patient without AAA progression (H).

PAAI-renal-difference 5–10 mm = PAAI-renal-5 mm − PAAI-renal-10 mm

PAAI-terminal-difference 5–10 mm = PAAI-terminal5 mm − PAAI-terminal-10 mm

Ethical Considerations

This study was approved by the Institutional Ethics Committee of Tsuchiura Kyodo General Hospital (Reference no. 886/Tsuchiura; November 11, 2019) and conformed to the Declaration of Helsinki statement on research involving human subjects. Informed consent for registration into the institutional MDCTA database and the potential use of data for future research work was provided by all participants after thorough explanation of the protocol and potential risks related to imaging before MDCTA examinations.

Statistical Analysis

Categorical data are presented as numbers and percentages, and were compared using the χ² test or Fisher’s exact test, as appropriate. Continuous variables are presented as the median with interquartile range (IQR) because none of the variables was normally distributed; continuous variables were compared using the Mann-Whitney U-test.

Binary logistic regression analyses were performed to determine the predictors of AAA progression (growth rate ≥5 mm/year). The associated variables with P<0.10 in the univariate analysis were entered in the multivariable model, and a forward stepwise regression method was used to fit the multivariable model. Odds ratios (ORs) with 95% confidence intervals (95% CIs) are presented in the text and tables.

Univariable and multivariable linear regression analyses were performed to identify the determinants of annual AAA growth rate. The associated variables with P<0.10 in the univariate analysis were entered in the multivariable model, and the stepwise regression method was used to fit the model. Receiver operating characteristic (ROC) curve analysis was used to identify the predictive values of variables and models for AAA progression. The optimal cut-off point in the ROC analysis was defined as the value with the highest sum of sensitivity and specificity. The areas under the curve (AUCs) of ≥2 ROC curves were compared by the comparison analysis method described by DeLong et al.¹⁶ Integrated discrimination improvement (IDI) and the net reclassification improvement (NRI) were determined with a category-free option among models.¹⁷

Statistical analyses were performed using JMP version 13 (SAS Institute, Cary, NC, USA) and R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Baseline Patient Characteristics and AAA Diameter

Of the 77 patients finally analyzed in this study, 19 had an annual AAA growth rate ≥5 mm and were categorized as the progression group, whereas the remaining 58 patients had an annual AAA growth rate <5 mm (non-progression group). No patient presented with AAA rupture or aortic dissection during the follow-up period.

Baseline patient characteristics at the time of the index MDCTA examination are summarized in Table 1. There were no significant differences in age, sex, comorbidity, or laboratory data between the 2 groups. Median baseline AAA diameter at the index MDCTA in the total cohort was 38.9 mm (IQR 2.7–42.9 mm), and was significantly greater in the progression than non-progression group (41.4 [IQR 38.9–48.3] vs. 37.4 [IQR 30.7–41.7] mm, respectively; P=0.005). The median annual AAA growth rate in the total cohort was 3.1 mm/year (IQR 1.5–4.9 mm/year).

Table 1. Patient Characteristics at the Index MDCTA Examination

	All patients (n=77)	AAA progression (n=19)	AAA non-progression (n=58)	P value
Age (years)	74.0 [69.5~78.0]	75.0 [67.0~79.0]	74.0 [69.8~78.0]	0.574
BMI (kg/m2)	23.2 [21.3~25.4]	21.8 [20.6~26.4]	23.3 [21.7~25.1]	0.406
Male sex	71 (92.2)	17 (89.5)	54 (93.1)	0.609
Hypertension	63 (81.8)	15 (79.0)	48 (82.8)	0.708
Diabetes	19 (24.8)	4 (21.0)	15 (25.9)	0.673
Dyslipidemia	38 (49.4)	9 (47.4)	29 (50.0)	0.842
Smoker	57 (74.0)	14 (73.7)	43 (74.1)	0.781
Laboratory data
CRP (mg/dL)	0.15 [0.06~0.36]	0.15 [0.05~0.38]	0.16 [0.06~0.39]	0.637
WBC (/μL)	6,250 [5,140~7,205]	5,450 [4,530~7,160]	6,380 [5,648~7,275]	0.453
eGFR (mL/min/1.73 m2)	68.4 [57.2~79.3]	67.4 [47.1~78.6]	70.0 [52.3~79.5]	0.551
TC (mg/dL)	178 [160~202]	172 [154~184]	182 [160~210]	0.112
LDL-C (mg/dL)	107 [91~125]	103 [89~113]	109 [91~129]	0.167
HDL-C (mg/dL)	44 [39~54]	47 [40~54]	43 [39~54]	0.561
TG (mg/dL)	122 [82~187]	120 [80~184]	125 [83~190]	0.773
HbA1c (%)	5.8 [5.6~6.1]	5.7 [5.5~5.9]	5.9 [5.6~6.2]	0.241
LVEF (%)	67 [62~71]	65 [62~69]	67 [62~72]	0.438
Medications
Statin	26 (33.7)	5 (26.3)	21 (36.2)	0.428
ACEI/ARB	40 (51.9)	9 (47.4)	31 (53.4)	0.645
β-blocker	11 (14.3)	4 (21.0)	7 (12.0)	0.331
Calcium channel blocker	40 (51.9)	12 (63.1)	28 (48.2)	0.251
MDCTA data
AAA diameter (mm)
Baseline	38.9 [32.7~42.9]	41.1 [38.9~48.3]	37.4 [30.7~41.7]	0.005
At follow-up examination	42.3 [33.9~48.7]	50.1 [45.0~55.6]	40.1 [32.0~47.1]	<0.001
Follow-up MDCTA period (years)	1.4 [1.0~2.4]	1.2 [0.8~1.9]	1.5 [1.1~2.4]	0.220
Annual AAA growth rate (mm/year)	3.1 [1.5~4.9]	7.1 [6.0~10.5]	1.9 [0.8~3.6]	<0.001
PAAI
PAAI-renal-5 mm	−74.2 [−82.6~−63.6]	−64.0 [−77.3~−60.1]	−75.1 [−78.6~−71.7]	0.015
PAAI-renal-10 mm	−77.0 [−80.4~−66.2]	−68.3 [−79.7~−62.0]	−77.7 [−80.8~−73.1]	0.034
Difference (5–10 mm)	2.4 [1.3~3.3]	2.7 [1.6~4.6]	2.2 [1.1~3.0]	0.113
PAAI-terminal-5 mm	−73.6 [−78.6~−65.8]	−65.8 [−77.4~−60.4]	−75.0 [−78.9~−80.0]	0.017
PAAI-terminal-10 mm	−77.0 [−81.4~−69.4]	−69.2 [−81.5~−62.4]	−77.7 [−81.3~−74.3]	0.053
Difference (5–10 mm)	2.6 [1.6~3.5]	2.9 [2.1~3.6]	2.5 [1.5~3.5]	0.202
PAAT volume (mL)
PAAT volume-renal-5 mm	24.2 [15.4~30.9]	18.5 [9.9~27.8]	24.2 [17.8~32.0]	0.210
PAAT volume-renal-10 mm	55.3 [34.9~66.9]	36.9 [19.2~61.0]	55.4 [41.2~67.6]	0.171
PAAT volume-terminal-5 mm	26.7 [17.1~35.0]	18.0 [12.5~34.5]	27.0 [19.3~35.6]	0.133
PAAT volume-terminal-10 mm	59.7 [38.4~76.9]	59.7 [33.2~79.0]	59.8 [43.0~76.9]	0.420

Unless indicated otherwise, data are presented as n (%), mean±SD, or as the median [interquartile range]. Periaortic adipose tissue inflammation was assessed by calculating the periaortic adipose tissue attenuation index (PAAI) in 4 different regions of interest (ROI) using combinations of 2 different thicknesses (5 and 10 mm) of the cylindrical layer from the outer vessel wall and 2 different segments (renal and terminal) of the aorta: PAAI-renal-5 mm, PAAI-renal-10 mm, PAAI-terminal-5 mm, and PAAI-terminal-10 mm. Similarly, periaortic adipose tissue (PAAT) volume was calculated at the infrarenal and terminal aorta segments in the 5- and 10-mm layers (PAAT volume-renal-5 mm, PAAT volume-renal-10 mm, PAAT volume-terminal-5 mm, and PAAT volume-terminal-10 mm). AAA, abdominal aortic aneurysm; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BMI, body mass index; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction; MDCTA, multidetector computed tomography angiography; TC, total cholesterol; TG, triglyceride; WBC, white blood cells.

Median baseline PAAI-renal-5 mm, PAAI-renal-10 mm, PAAI-terminal-5 mm, and PAAI-terminal-10 mm was −73.6 (IQR −78.6~−5.8), −77.0 (IQR −81.4~−69.4), −74.2 (IQR −82.6~−63.6) and −77.0 (IQR −80.4~−66.2), respectively. Linear regression and Bland-Altman plots indicated that these 4 measurements were comparable (Supplementary Figures 1,2), although PAAI-renal-5 mm was slightly higher than PAAI-renal-10 mm (P=0.039) and PAAI-terminal-5 mm was slightly higher than PAAI-terminal-10 mm (P=0.029). PAAI values were consistently higher in the progression than non-progression group, although the difference did not reach statistical significance for PAAI-terminal-10 mm (Table 1). There were no significant differences in PAAI-renal-difference 5–10 mm and PAAI-terminal-difference 5–10 mm between the AAA progression and non-progression groups (Table 1), and neither of these parameters was significantly correlated with annual AAA progression rate in the infrarenal region (R=0.20, P=0.08) or terminal aorta (R=0.17, P=0.15; Supplementary Figures 3,4).

Predictors of AAA Progression

Predictors of AAA progression were assessed with univariable and multivariable logistic regression analyses (Table 2). In the univariate analysis, AAA diameter at the index MDCTA and PAAI measurements were significantly associated with AAA progression ≥5 mm/year. Regarding AAA diameter at the index MDCTA, ROC analysis showed that the best cut-off threshold of AAA diameter was 37.7 mm (AUC 0.714 [95% CI 0.589–0.839], sensitivity 94.7%, specificity 43.1%, accuracy 55.8%; Figure 4A). In terms of PAAI measurements, all PAAI values were significantly predictive in univariate analysis. Thereafter, the predictive values of PAAI measurements were evaluated by comparing the AUCs of ROC analyses.

Table 2. Univariate and Multivariate Analysis of Factors Predicting Growth of AAA Diameter ≥5.0 mm/year From the Initial to Follow-up MDCTA Examination

	Univariate logistic regression			Multivariate logistic regression
	OR	95% CI	P value	OR	95% CI	P value
Baseline AAA diameter	1.00	1.01~1.26	0.002	1.16	1.05~1.28	0.001
PAAI-renal-5 mm	1.09	1.02~1.15	0.004	1.12	1.05~1.20	0.004
PAAI-renal-10 mm	1.07	1.01~1.14	0.018
PAAI-terminal-5 mm	1.08	1.02~1.15	0.006
PAAI-terminal-10 mm	1.07	1.01~1.13	0.025
TC	0.98	0.97~1.01	0.097
Calcium channel blocker	2.48	0.83~7.44	0.094

Periaortic adipose tissue inflammation was assessed by calculating the PAAI in 4 different ROI using combinations of 2 different thicknesses (5 and 10 mm) of the cylindrical layer from the outer vessel wall and 2 different segments (renal and terminal) of the aorta: PAAI-renal-5 mm, PAAI-renal-10 mm, PAAI-terminal-5 mm, and PAAI-terminal-10 mm. CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.

Figure 4.

Receiver operating characteristic (ROC) curve analyses of (A) baseline abdominal aortic aneurysm (AAA) diameter, (B) the periaortic adipose tissue attenuation index in the 5-mm layer of a 100-mm segment distal to the ostium of the right renal artery (PAAI-renal-5 mm), and (C) the univariate model (blue line, baseline AAA diameter) and multivariable model (green line; baseline AAA diameter+PAAI-renal-5 mm) to predict AAA progression. The area under the curve for predicting AAA progression was numerically greater for the multivariable than univariate model with AAA diameter. CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value.

The AUCs of PAAI-renal-5 mm, PAAI-renal-10 mm, PAAI-terminal-5 mm, and PAAI-terminal-10 mm were 0.688 (95% CI 0.541–0.834), 0.662 (95% CI 0.503–0.822), 0.683 (95% CI 0.536–0.831), and 0.648 (95% CI 0.489–0.808), respectively (Supplementary Figure 5). There were no significant differences in the AUCs of the ROC curves for the 4 different PAAI measurements for predicting AAA progression. Among the PAAI measurements, PAAI-renal-5 mm showed the numerically largest AUC (0.688 [95% CI 0.541–0.834], sensitivity 63.1%, specificity 77.6%, accuracy 74.0%) and was therefore used in further analyses. The best cut-off threshold of PAAI-renal-5 mm was −71.1 HU (Figure 4B).

In the multivariable regression analysis, AAA diameter at the index MDCTA (OR 1.16; 95% CI 1.05–1.28, P=0.001) and PAAI-renal-5 mm (OR 1.12; 95% CI 1.05–1.20, P=0.004) remained independent predictors of AAA progression. We also evaluated the incremental predictive value of PAAI-renal-5 mm on top of baseline AAA diameter. In ROC analysis, the AUC was numerically greater with the multivariable model including both baseline AAA diameter and PAAI-renal-5 mm compared with the univariable model of baseline AAA alone (0.71 vs. 0.80; P=0.110; Figure 4C), which was not statistically significant.

When patients were stratified according to the best cut-off values of baseline AAA diameter (≥37.7 mm) and PAAI-renal-5 mm (≥−71.1 HU), the incidence of AAA progression differed significantly among the 4 groups (Figure 5). Among patients with both baseline AAA diameter ≥37.7 mm and PAAI-renal-5 mm ≥−71.1 HU, AAA progression was observed in 73.3%; in comparison, no patients showed AAA progression when baseline AAA diameter was <37.7 mm and PAAI-renal-5 mm was <−71.1 HU. In the assessment of the relative IDI and NRI, adding PAAI-renal-5 mm to the univariable model of baseline AAA diameter significantly increased the accuracy to discriminate AAA progression (NRI 0.675 [95% CI 0.180–1.171; P=0.007] and IDI 0.144 [0.040–0.248; P=0.007]).

Figure 5.

Prevalence of abdominal aortic aneurysm (AAA) progression in the 4 groups stratified according to AAA diameter (≥37.7 and <37.7 mm) and the periaortic adipose tissue attenuation index in the 5-mm layer of a 100-mm segment distal to the ostium of the right renal artery (PAAI-renal-5 mm; ≥−71.1 and <−7.1 HU).

Determinants of Annual AAA Growth Rate

Univariable and multivariable linear regression analyses were used to evaluated the determinants of AAA growth rate (Table 3). In addition to baseline AAA diameter and PAAI-renal-5 mm, variables with P<0.01 (including male sex, total cholesterol level, and diabetes) were included in the multivariate model. After stepwise regression analysis, the final model included baseline AAA diameter and PAAI-renal-5 mm as independent determinants.

Table 3. Univariate and Multivariate Linear Regression Analysis of Factors Predicting the Degree of AAA Progression

	Univariate linear analysis			Multivariate linear analysis
	β	95% CI	P value	β	95% CI	P value
Male sex	0.41	−0.24~13.5	0.058
Baseline AAA diameter	0.15	0.10~0.29	<0.001	0.16	0.11~0.29	<0.001
PAAI-renal-5 mm	0.12	0.02~0.21	0.002	0.10	0.03~0.20	0.010
TC	−0.02	−3.63~0.31	0.098
Diabetes	−0.23	−0.49~0.00	0.073

PAAI-renal-5 mm, periaortic adipose tissue attenuation index in the 5-mm layer of a 100-mm segment distal to the ostium of the right renal artery. Abbreviations as in Tables 1,2.

Discussion

The progression of AAA is driven by several potential pathogenic mechanisms that are associated with inflammation and tissue degradation.^1,9 To the best of our knowledge, this study is the first to investigate the clinical significance of the mean periaortic adipose tissue attenuation value on MDCTA in patients with AAA for the prediction of AAA progression. The main findings of this study are that: (1) baseline AAA diameter at the index MDCTA and PAAI obtained by MDCTA analysis were independently associated with an increase in AAA diameter; (2) PAAI measurements were nearly identical among the different ROIs in the 100-mm segments along the aorta located between the renal artery and terminal aorta; and (3) PAAI showed an incremental predictive value for the progression of AAA in addition to baseline AAA diameter.

Aneurysm Size, Vascular Inflammation, and AAA Progression

It is well known that the baseline diameter of the AAA is one of the major predictors of AAA rupture.^1–3 In previous studies, AAA diameter was shown to correlate with AAA growth rate more than linearly, which can be explained by consideration of flow dynamics. The wall stress on an aorta without an aneurysm is relatively low and uniformly distributed. However, according to Laplace’s law, wall tension increases proportionally to the vessel radius for a given blood pressure, which promotes vessel enlargement in an aneurysmal aorta.¹⁸ Moreover, with the formation of an aortic aneurysm, different areas of high and low stress are created on the vessel wall, which may result in vessel dilatation, deformation, and weakening of the aortic wall.¹⁹ Thus, AAA diameter is the standard measure for the management of AAA in clinical practice.⁴ In line with the current consensus,^1,3,4 the present study showed that baseline AAA diameter was positively correlated with AAA growth rate (Tables 2,3).

Imaging of PAT

Vascular inflammation is considered a precursor of atherosclerosis and subsequent vascular events. There is growing research interest in the assessment of vascular inflammation using non-invasive imaging technologies. One of the promising imaging targets to evaluate the degree and extent of vascular inflammation is adipose tissue. Visceral adipose tissue, especially PAT secretes a variety of adipocytokines, which affect the biology of the adjacent vascular wall.²⁰ Moreover, PAT is thought to have a bidirectional interaction with vascular inflammation via paracrine signaling.^12,21,22 A recent study by Antonopoulos et al demonstrated a significant association between the histological characteristics of PAT surrounding the coronary artery and vascular inflammatory cytokines in the coronary arterial wall that indicated the presence of an inside-to-outside pathway around coronary arteries.¹² Thus, PAT can be considered a marker of vascular inflammation. The clinical implications of PAT imaging have been explored in recent years using PET imaging,²³ MRI,²⁴ or CT. Of note, CT attenuation of pericoronary fat has been used in clinical studies since the computational analysis methodology of CT attenuation within the adipose tissue was validated in a recent ex vivo study of coronary arteries.¹² Specifically, the pericoronary FAI has been shown to be associated with coronary flow,²⁵ myocardial ischemia,²⁶ and coronary vasospasm.²⁷

Assessment of Vascular Inflammation in the Aorta Using CT Angiography

CT is a standard modality for the morphological assessment of AAA, including the size of the aneurysm; nevertheless, qualitative assessment by CT of the pathogenesis of AAA including vascular inflammation has not been fully evaluated. Although the assessment of vascular inflammation with CT attenuation analysis was validated in coronary arteries,¹² there is no histopathological validation of using the same methodology in the aorta and periaortic adipose tissue. Unlike the coronary arteries, the abdominal aorta is surrounded by various organs, such as the intestines, kidneys, and vertebrae. In particular, tissues showing high CT attenuation, including bone tissue of the vertebrae or the contrast-enhanced renal cortex, are susceptible to artifacts caused by beam hardening effects or the partial volume effect, which increases the CT attenuation values of adjacent regions. However, in the measurement of CT attenuation of PAT proposed by Antonopoulos et al,¹² voxels with a specific range of CT attenuation values (from −30 to −190 HU) are extracted from the ROI, and the weighted mean attenuation of those voxels is calculated as the attenuation index. Considering vessel diameter and length in the abdominal aorta, in this study we arbitrary predefined 2 different aortic segments to be analyzed and 2 different thicknesses (5 and 10 mm) for periaortic adipose tissue. Of note, PAAI measured at the infrarenal and terminal aorta showed strong correlations (Supplementary Figures 1,2), and the values were almost identical (Table 1) even though the volume of interest of the PAAI-renal and PAAI-terminal segments includes a different composition of adjacent organs; this may indicate the feasibility of PAAI measurement for the abdominal aorta.

Nevertheless, PAAI values were slightly but significantly higher for the 5- than 10-mm layer measurements, which may be explained by a stronger interaction with vascular inflammation in the inner than outer layer of the PAT. The optimal thickness of a cylindrical ROI needs to be explored in further studies. Dias-Neto et al investigated the CT density of PAT on cross-sectional images of the abdominal aorta in patients with AAA, aortoiliac occlusive disease, and those without aortic disease.²⁸ These authors found that patients with AAA had a higher density of adipose tissue compared with the other patient groups.²⁸ Moreover, they demonstrated a significant correlation between the CT density of periaortic adipose tissue and aortic dimension.²⁸ In the present study, we focused on patients diagnosed with AAA and investigated the association between PAAI measured using volumetric analysis and the rate of progression of aortic expansion in serial observations. PAAIs were consistently associated with AAA progression, independent of baseline AAA diameter (Tables 2,3). Moreover, the PAAI complements the predictive value of the baseline AAA diameter, resulting in a significant net classification index. Our results suggest that the vascular inflammation status assessed by PAAI on MDCTA of aneurysmal aorta is a feasible marker for the rapid progression of AAA, and may help physicians manage patients with AAA in terms of risk stratification and determining surveillance intervals.

Study Limitations

The results of the present study should be interpreted taking into consideration several limitations. First, this study is a retrospective observational study performed in a single center and including a relatively small number of patients; therefore, a selection bias cannot be excluded. Second, although MDCTA analyses were performed by an experienced investigator completely blinded to the patients’ characteristics or the sequence of examinations, these analyses were not performed by a dedicated core laboratory. Third, in the measurement of PAAI, the specific range of CT attenuation values from −30 to −190 HU was adopted on the basis of previous studies of coronary arteries. Further studies are needed to validate these values in the assessment of aortic inflammation. The effects of beam hardening and the partial volume of adjacent organs are uncertain and remain to be determined in the abdominal aorta. Fourth, MDCTA examinations were scheduled at the discretion of the physicians, which may have influenced the results. Further large-scale, multicenter, prospective studies are needed to validate the clinical implications of the present results. Fifth, the annual growth rate of the aorta in the present study was defined as the change in AAA diameter divided by the time from the index MDCTA to the follow-up MDCTA examination, assuming that the growth rate was constant over the study period. Finally, given the retrospective nature of the analysis, connective tissue disorders or aortitis have not been routinely ruled out in the present study.

Conclusions

PAAI was an independent predictor of the progression of aortic dilation, supporting the notion that local adipose tissue inflammation contributes to aortic remodeling in the case of an AAA. Comprehensive assessment by MDCTA, including PAAI, may feasibly provide information to identify high-risk lesions that could result in future AAA progression.

Acknowledgments / Sources of Funding / Disclosures

None.

IRB Information

This study was approved by the Institutional Ethics Committee of Tsuchiura Kyodo General Hospital (Reference no. 886).

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-20-1014

References

• 1.   Kent KC. Clinical practice: Abdominal aortic aneurysms. N Engl J Med 2014; 371: 2101–2108.
• 2.   Brady AR, Thompson SG, Fowkes FG, Greenhalgh RM, Powell JT; UK Small Aneurysm Trial Participants. Abdominal aortic aneurysm expansion: Risk factors and time intervals for surveillance. Circulation 2004; 110: 16–21.
• 3.   Nordon IM, Hinchliffe RJ, Loftus IM, Thompson MM. Pathophysiology and epidemiology of abdominal aortic aneurysms. Nat Rev Cardiol 2011; 8: 92–102.
• 4.   Anderson JL, Halperin JL, Albert NM, Bozkurt B, Brindis RG, Curtis LH, et al. Management of patients with peripheral artery disease (compilation of 2005 and 2011 ACCF/AHA guideline recommendations): A report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2013; 127: 1425–1443.
• 5.   Thompson RW, Geraghty PJ, Lee JK. Abdominal aortic aneurysms: Basic mechanisms and clinical implications. Curr Probl Surg 2002; 39: 110–230.
• 6.   Britton KA, Fox CS. Perivascular adipose tissue and vascular disease. Clin Lipidol 2011; 6: 79–91.
• 7.   Courtois A, Nusgens BV, Hustinx R, Namur G, Gomez P, Somja J, et al. ¹⁸F-FDG uptake assessed by PET/CT in abdominal aortic aneurysms is associated with cellular and molecular alterations prefacing wall deterioration and rupture. J Nucl Med 2013; 54: 1740–1747.
• 8.   Sakaue T, Suzuki J, Hamaguchi M, Suehiro C, Tanino A, Nagao T, et al. Perivascular adipose tissue angiotensin II type 1 receptor promotes vascular inflammation and aneurysm formation. Hypertension 2017; 70: 780–789.
• 9.   MA3RS Study Investigators. Aortic wall inflammation predicts abdominal aortic aneurysm expansion, rupture, and need for surgical repair. Circulation 2017; 136: 787–797.
• 10.   McBride OM, Joshi NV, Robson JM, MacGillivray TJ, Gray CD, Fletcher AM, et al. Positron emission tomography and magnetic resonance imaging of cellular inflammation in patients with abdominal aortic aneurysms. Eur J Vasc Endovasc Surg 2016; 51: 518–526.
• 11.   Thanassoulis G, Massaro JM, Corsini E, Rogers I, Schlett CL, Meigs JB, et al. Periaortic adipose tissue and aortic dimensions in the framingham heart study. J Am Heart Assoc 2012; 1: e000885.
• 12.   Antonopoulos AS, Sanna F, Sabharwal N, Thomas S, Oikonomou EK, Herdman L, et al. Detecting human coronary inflammation by imaging perivascular fat. Sci Transl Med 2017; 9: eaal2658.
• 13.   Goeller M, Achenbach S, Cadet S, Kwan AC, Commandeur F, Slomka PJ, et al. Pericoronary adipose tissue computed tomography attenuation and high-risk plaque characteristics in acute coronary syndrome compared with stable coronary artery disease. JAMA Cardiol 2018; 3: 858–863.
• 14.   Tamarappoo B, Dey D, Shmilovich H, Nakazato R, Gransar H, Cheng VY, et al. Increased pericardial fat volume measured from noncontrast CT predicts myocardial ischemia by SPECT. JACC Cardiovasc Imaging 2010; 3: 1104–1112.
• 15.   Hell MM, Ding X, Rubeaux M, Slomka P, Gransar H, Terzopoulos D, et al. Epicardial adipose tissue volume but not density is an independent predictor for myocardial ischemia. J Cardiovasc Comput Tomogr 2016; 10: 141–149.
• 16.   DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988; 44: 837–845.
• 17.   Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: From area under the roc curve to reclassification and beyond. Stat Med 2008; 27: 157–172; discussion 207–212.
• 18.   Macura KJ, Corl FM, Fishman EK, Bluemke DA. Pathogenesis in acute aortic syndromes: Aortic aneurysm leak and rupture and traumatic aortic transection. AJR Am J Roentgenol 2003; 181: 303–307.
• 19.   Raghavan ML, Vorp DA, Federle MP, Makaroun MS, Webster MW. Wall stress distribution on three-dimensionally reconstructed models of human abdominal aortic aneurysm. J Vasc Surg 2000; 31: 760–769.
• 20.   Oikonomou EK, Antoniades C. Immunometabolic regulation of vascular redox state: The role of adipose tissue. Antioxid Redox Signal 2018; 29: 313–336.
• 21.   Antonopoulos AS, Antoniades C. The role of epicardial adipose tissue in cardiac biology: Classic concepts and emerging roles. J Physiol 2017; 595: 3907–3917.
• 22.   Margaritis M, Antonopoulos AS, Digby J, Lee R, Reilly S, Coutinho P, et al. Interactions between vascular wall and perivascular adipose tissue reveal novel roles for adiponectin in the regulation of endothelial nitric oxide synthase function in human vessels. Circulation 2013; 127: 2209–2221.
• 23.   Iwatsuka R, Matsue Y, Yonetsu T, O’Uchi T, Matsumura A, Hashimoto Y, et al. Arterial inflammation measured by ¹⁸F-FDG-PET-CT to predict coronary events in older subjects. Atherosclerosis 2018; 268: 49–54.
• 24.   Toya T, Corban MT, Imamura K, Bois JP, Gulati R, Oh JK, et al. Coronary perivascular epicardial adipose tissue and major adverse cardiovascular events after ST segment-elevation myocardial infarction. Atherosclerosis 2020; 302: 27–35.
• 25.   Kanaji Y, Hirano H, Sugiyama T, Hoshino M, Horie T, Misawa T, et al. Pre-percutaneous coronary intervention pericoronary adipose tissue attenuation evaluated by computed tomography predicts global coronary flow reserve after urgent revascularization in patients with non-ST-segment-elevation acute coronary syndrome. J Am Heart Assoc 2020; 9: e016504.
• 26.   Nomura CH, Assuncao-Jr AN, Guimaraes PO, Liberato G, Morais TC, Fahel MG, et al. Association between perivascular inflammation and downstream myocardial perfusion in patients with suspected coronary artery disease. Eur Heart J Cardiovasc Imaging 2020; 21: 599–605.
• 27.   Ueno H, Hoshino M, Sugiyama T, Kanaji Y, Nogami K, Horie T, et al. Pericoronary adipose tissue inflammation on coronary computed tomography in patients with vasospastic angina. JACC Cardiovasc Imaging 2020; 14: 511–512.
• 28.   Dias-Neto M, Meekel JP, van Schaik TG, Hoozemans J, Sousa-Nunes F, Henriques-Coelho T, et al. High density of periaortic adipose tissue in abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 2018; 56: 663–671.