Introduction

Atherosclerosis is a systemic, progressive, and chronic condition that can affect the entire vascular tree [1]. Calcium in the artery wall is considered a direct marker of atherosclerotic disease [1] and can be easily evaluated through computed tomography (CT) [2]. There are remarkable mass of robust data supporting the prime role of coronary artery calcium (CAC) in cardiovascular risk assessment of the intermediate-risk population, as well as specific subgroups, as patients with diabetes and family history of premature coronary heart disease (CHD) [3]. Several studies have shown that thoracic aortic calcium (TAC) is also a marker of subclinical atherosclerosis [4]. Distinct associations of TAC arouse interest in its particularities compared with CAC analysis. TAC also impacts the CV system, as aortic wall calcium worsen arterial stiffening [5], which is associated with several implications for end-organ damage [6]. CAC and TAC prevalence also seem to differ between men and women and race/skin color [7,8,9,10,11], though results are inconsistent. Moreover, unlike CAC, TAC has not been evaluated through standard CT protocol, mainly with regard to TAC anatomical extension [12,13,14] and the use of ECG synchrony during exam [15].

Because differences in TAC definition and acquisition might impair the evaluation of study’s results on the predictive value of TAC both at individual and population levels, our aim was to review recent studies about TAC, discussing the particularities of the aortic wall calcium formation and the differences between the aortic segments. And, finally, emphasize the anatomical references and the extension of the aorta included in the TAC studies.

Mechanisms Related to the Calcium Formation in the Thoracic Aorta Wall

The distribution of calcium along the aorta is usually very heterogeneous. It is possible to identify coarse calcium in one segment, while there is no calcium in another segment from the same individual, as shown in Fig. 1. In the first case (images 1a and 1b), there was calcium in large amount in the arch and descending thoracic segments, while there was no calcium in the ascending aorta. In the second case (images 1c and 1d), the calcium concentration was much higher in the aortic arch compared with ascending and descending thoracic portions.

Fig. 1
figure 1

Heterogeneous distribution of calcium along the aorta. a, b CT reconstructions in the parasagittal plane. In this case, ascending aorta had no calcium (arrow in a), whereas in the arch and descending portions (arrow in b) there were circumferential plaques covering almost all aortic wall. c, d CT reconstructions in the axial plane. The most calcium concentration was in the aortic arch (arrow in c), while in ascending (superior arrow in d) and descending (inferior arrow in d) segments calcium were coarse, but sparse

Full size image

The variation in the distribution of calcium across aorta segments may be in part associated with different embryonic origin of the vascular smooth muscle cells colonizing the aorta, which in the aortic arch derives from cardiac neural crest cells, whereas the calcium found in the descending aorta derives from the mesoderm [16]. The Leroux-Berger et al. study found correlation between the embryonic origin of vascular smooth muscle cells and the timing of the appearance of calcium [16]. Thus each aortic segment differs in their embryonic origin and is subject to different hemodynamic stress, which also appears to affect susceptibility to calcium [16], as the rate of calcium seems to differ among individuals [17]. Therefore, the calcium found in each aortic segment may be associated differently to cardiovascular risk factors [18] and probably has distinct predictive value for cardiovascular (CV) and non-CV morbidity and mortality, as suggested in some studies [12, 13, 19, 20•, 21•].

Another important particularity refers to the molecular mechanism of plaque calcium in the aortic wall, which is mainly composed by two mechanisms:

  1. 1)

    Intimal calcium: atherosclerosis, inflammatory response of tunica intima;

  2. 2)

    Medial calcium: occurs independently of intimal calcium in the tunica media.

The intimal layer consists of endothelial cells that eventually form atheromatous plaques which can rupture and cause thromboembolic events, whereas the medial layer consists of smooth muscle cells and elastic fibers that are associated with blood flow and arterial pressure regulation [2]. Medial calcium is thought to cause arterial stiffening, reduce compliance, and limit distensibility [2]. Actually the way to distinguish intimal and medial calcium is through ex vivo histological analysis [22]. Then CT scans cannot define if the calcium is in the intimal or medial layer of aortic wall [2].

However, the patterns of calcium distribution observed in CT scans may suggest the predominance of intimal or medial calcium. Intimal calcium usually has a patchy distribution within atherosclerotic lesions and is most commonly amorphous without distinct architecture [23•]. On the other hand, vascular medial calcium is generally concentric, appears more circumferential, and has a diffuse distribution [23•]. Figure 2 shows schematically the patterns of calcium distribution in the tunica intima and media.

Fig. 2
figure 2

Patterns of calcium distribution in the aortic wall. Intimal calcium has a patchy distribution in the atherosclerotic lesion and medial calcium is generally diffuse and circumferential

Full size image

Frequently, medial calcium is associated with uremia, radiotherapy, or vascular inflammation which induces a phenotypic change of vascular smooth muscle cells into osteoblasts, a process of metabolite-induced (toxic) vascular changes in the absence of lipid deposits [24•]. However aging, chronic kidney disease, diabetes mellitus, and mediastinal radiation are also associated with accelerated intimal atherosclerosis [24•]. Therefore, the overlap of these two processes in the aortic wall might explain some differences in findings on cardiovascular risk factors associated with calcium in different vascular beds, and mainly between distinct aorta territories.

Differences in the Anatomical References Used at TAC Evaluation

The differences in TAC evaluation can impact both the identification of calcium as well as its quantification either in volume or using the Agatston score. So, caution is advisable when interpreting and comparing studies that used different TAC extensions. Table 1 shows selected studies published in the last 5 years evaluating calcium in the thoracic aorta. They were grouped based on the aortic segments included in the analysis. The first group evaluated calcium in three segments: ascending thoracic aorta (ATA), aortic arch, and descending thoracic aorta (DTA). The aortic root calcium was not included in this group. The second group represents the largest one, and evaluated TAC in the ATA, DTA, and in the aortic root, but not in the aortic arch. The third, fourth, and fifth groups included each a single study that used distinct anatomical references, respectively: the extended versions of ATA plus DTA, aortic arch, and aortic root. These studies shown in Table 1 also differ with respect on how they measured or analyzed the presence of calcium: yes/no [11], and/or Agatston [4, 12, 13, 25•,26,27,28,29,30,31,32] and/or volume [5, 33,34,35, 36••], and/or density [19, 20•, 21•], and/or semi-qualitative evaluation [14].

Table 1 Recent thoracic aorta calcium studies organized according to each group of thoracic aorta anatomical references used in the computed tomography evaluation
Full size table

Does the Aortic Arch Add Relevant Information to TAC?

As shown by Craiem et al., the inclusion of aortic arch in combination with ATA and DTA doubled the TAC prevalence, mainly in middle-aged women [25•]. Besides the impact on the overall and sex-specific prevalence of TAC, the inclusion of the aortic arch in TAC evaluation might also relate to TAC predictive value on morbidity and mortality. Bos et al., for instance, analyzed only the aortic arch and found that the volume of calcium in this segment was related to increased CV and non-CV mortality, after adjustment for many CV risk factors including CAC, intracranial, and extracranial internal carotids calcium [33••]. A recent study of Cho et al. used the same TAC extension described by Craiem et al. and followed 702 patients without obstructive coronary artery disease (CAD) during 64 months and also found TAC as an independent predictor for outcomes, especially stroke [34]. Thus, taking account these latter results, it appears that calcium in the aortic arch might contribute to TAC prediction, but we cannot rule out that it may be also a marker of calcium in other thoracic aorta segments.

What Do We Know About the Presence of Calcium in ATA and DTA?

The largest study group presented in Table 1 is the ones that assessed TAC using the same scan performed for CAC assessment. Thus, only the presence of calcium in the ascending thoracic aorta (ATAC) and descending thoracic aorta (DTAC) were evaluated. Eight studies are from Multi-ethnic Study of Atherosclerosis (MESA), and they measured calcium using Agatston, volume, and/or density [19, 20•, 21•, 27,28,29,30, 35]. The others are from Heinz Nixdorf Recall Study [12, 31], Framingham Heart Study [37], and EISNER [4], and all of them used Agatston to measure TAC. All these studies included the aortic root in the TAC and excluded the aortic arch. As demonstrated by Tesche et al., calcium in the aortic root is a stronger and independent predictor of CAC and of obstructive CAD [36••], suggesting that a similar process lead to calcium in these vascular beds. Thus, like the CAC [38], it is possible that calcium in aortic root reflects more localized than generalized atherosclerosis, differently from other thoracic aorta segments.

When taken together, results on ATAC plus DTAC associations and predictive value are controversial [27,28,29,30,31,32, 35]. However, when Thomas et al. and Kälsch et al. studied the ATAC and DTAC separately, they found that while greater ATAC volume predicted the incidence and progression of CHD and CVD [12, 21•], DTAC was associated with the occurrence of non-CV morbidity and mortality [20•]. These authors also showed that greater ATAC density, contrary to greater volume, was associated with lower risk of CAD [19, 21•], and explained such differences between aorta segments in terms of embryology, wall constitution and pathophysiologic mechanisms of calcium formation. It is thus, possible, that such differences in DTAC and ATAC also account for the controversial results reported by the other studies included in this group, as they are based on ATAC plus DTAC [27, 28, 30, 32]. In light of these recent findings, further research using the same anatomical references for each thoracic aorta segment should be stimulated.

Anatomical References for TAC Segmentation

Based on the current anatomical references used in some TAC studies, and understanding the possible value of studying each aortic segment separately, including the aortic arch, we created the video 1 to show each portion of the aorta slice-by-slice in axial CT images. Since aorta has an oblique path, some details are of importance. The following anatomical references were used for TAC segmentation:

  1. 1)

    ATAC: from the sinutubular junction to the lower edge of pulmonary artery bifurcation (Some caution with the first slices, because of the initial curvature of ascending aorta above aortic root, where there are some slices that both appear in the same axial slice).

  2. 2)

    Aortic arch calcium: from ascending to descending thoracic aorta at the same anatomical reference, which is the level of the lower edge of pulmonary artery bifurcation.

  3. 3)

    DTAC: from the lower edge of pulmonary artery bifurcation to the apex of the heart.

What Is the Best Way to Measure TAC?

In addition to anatomical definitions, other methodological TAC parameters deserve to be considered. Agatston method has been widely used; however, the quantification of TAC can vary considerably between different CT systems once the acquisition of CAC scans, usually used to measure TAC, was not created for this application [39]. Mori et al. in 2015 described and validated a new volume-rendering approach to quantify TAC that demonstrated an excellent agreement of the pixel-based TAC score with volumetric TAC score and observed that volume-based score was less influenced by slice thickness as compared with pixel-based score [40••]. Agatston score depends nonlinearly on the measured Hounsfield Unit density of each pixel in the calcium, which changes with different x-ray energies, while the calcium volume is only slightly affected by scanning at different energies [41]. Since TAC is in the early development phase, perhaps now is the time to think about more accurate measures of quantifying the TAC [39].

Is TAC Radiation Exposure Justified?

The last, but a very important consideration to be made, refers to the radiation dose involved in TAC extended exams (all segments). Although the increase in the radiation dose of extended CAC, necessary to include the aortic arch, is lower than that delivered for a bilateral mammogram [25•], its value remains uncertain. So far, there appear to be no doubt regarding the value of evaluating DTAC and ATAC on CAC scans, as CAC clinical indication is already established. Lung cancer screening trials [42] seem to offer some opportunity to evaluate the predictive value of all segments, especially the aortic arch.

Limitations

The current review of the literature is limited mostly due to the high variability across the studies included in the analysis. As previously detailed, there is no current standards to define which aortic segments to include or the most appropriate tool to quantify the presence and extent of TAC. Moreover, the outcomes included in each analysis are not similar. Collectively, those issues limit the comparison between studies and the potential to fully interpret those results in other populations or scenarios.

Future Directions

Future studies should focus on the standardization of image acquisition, areas of the thoracic aorta to be included and most appropriate tools to quantify TAC. Moreover, detailed investigation on the different role of each thoracic aorta segment for the prediction of different outcomes, including separate analysis for coronary artery disease events, cerebrovascular events, and incidence of acute aortic syndromes.

Additionally, more studies on the implications of such findings for clinical management are needed. Currently, TAC is understood to be atherosclerosis. However, the clinical management of asymptomatic individuals with atherosclerosis is currently based on the individual’s clinical risk profile with the potential use of other diagnostic tools, such as CAC scores in selected individuals. Yet, not clear role for TAC in selecting the most appropriate management strategy for those individuals exist.

Conclusions

TAC has been considered as subclinical marker of atherosclerosis; however, the lack of standard protocol regarding the anatomical segments included and measurement analytical unit have contributed to controversial results and studies comparability. The accumulated evidences indicate that each aorta segment should be evaluated separately, as they differ in terms of structural characteristics, embryologic origin, and pathophysiologic mechanisms of calcium formation along the aorta and predictive value.