Prediction of Progression or Regression of Type A Aortic Intramural Hematoma by Computed Tomography

Methods

CT Analysis

IMH was defined by CT as a localized segmental and crescent high attenuation area along the aortic wall on noncontrast CT and relatively low attenuation area without enhancement on contrast-enhanced CT.^{1 31}

The course of all 22 patients was followed up with CT examination. Twenty-one patients had the first routine follow-up CT examination within 3 weeks after the initial onset. In 3 patients, recurrence of chest pain or back pain or both resulted in repeat CT examination. Seventeen cases had another routine follow-up CT examination between 3 weeks and 3 months after the initial onset. All CT scans were obtained at 1-cm intervals from the aortic arch to the aortic bifurcation with and without the rapid intravenous bolus injection of 100 mL of contrast media. CT was performed with a model HSA-RP scanner (Kobe General Hospital) (GE Medical Systems) or TSX-0111A scanner (Saiseikai Kumamoto Hospital) (TOSHIBA).

Ascending aortic images of CT were selected at 3 levels. The aorta at the level of the left or right main pulmonary artery was chosen as the median slice. The other 2 slices, which were superiorly and inferiorly adjacent to the median image, were also selected for evaluation. With the use of computerized planimetry, measurements were taken on the basis of the accompanying calibrated scales in the contrast-enhanced CT images. Major aortic diameter and minor aortic diameter, which were the longest and the shortest transverse of best-fit ellipses, the aortic wall thickness, and lumen diameter were measured in each of the slices (Figure 1⇓). Maximum aortic diameter was defined as the largest diameter of all major and minor aortic diameters in all 3 slices. Wall thickness–to–lumen diameter ratio was calculated as the maximum wall thickness divided by the maximum aortic diameter.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Measurement methods of various aortic dimensions. Aortic diameter of the outer contour of the affected aorta is indicated in major-axis diameter (a) and minor-axis diameter (b). Aortic wall thickness from the intima to the adventitia is indicated by c; lumen diameter is indicated by d.

To determine interobserver variability for these measurements, 20 studies were randomly selected and analyzed by 2 independent observers. To determine intraobserver variability, 20 studies were also repeated by the same observer. Interobserver and intraobserver variabilities for these measurements were 1.2% and 1.0%, respectively.

On the basis of the clinical outcome, the patients were divided into 2 groups: the progression group and the regression group. We defined patients who showed increased maximum aortic wall thickness in the follow-up CT images (n=9) or died of rupture (n=1) as the progression group (n=10). The other 12 patients, who all showed decreased maximum wall thickness in the follow-up CT images, were defined as the regression group.

Results

Mean CT follow-up period was 37 days (range 7 to 113 days). There were no significant differences in CT follow-up period between the progression group and the regression group (27±19 vs 44±26 days, P=0.103).

In the progression group, the maximum aortic wall thickness significantly increased in 9 patients from 14±5 to 21±7 mm (P=0.016), as shown in Table 1⇓ and Figure 2⇓, and 1⇑ patient died of rupture of IMH. In this group, the maximum aortic diameter also significantly increased from 55±6 to 63±9 mm (P=0.002). Six of these patients had overt aortic dissection. All patients in the progression group successfully underwent surgical repair after 40±32 days (mean±SD, range 7 to 94 days) from the onset except 1 patient, who declined operation and died.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Graphs depict maximum aortic wall thickness (left) and maximum aortic diameter (right) of the progression group in CT images on admission and at follow-up. Mean values±SD on admission and follow-up are shown in the solid lines with bars.

In the regression group, the maximum aortic wall thickness significantly decreased from 9±3 to 4±4 mm (P=0.001), as shown in Table 1⇑ and Figure 3⇓. In this group, maximum aortic diameter also significantly decreased from 47±3 to 44±4 mm (P=0.007). In 6 of these patients, IMH of the ascending aorta disappeared.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Graphs depict maximum aortic wall thickness (left) and maximum aortic diameter (right) of the regression group in CT images on admission and at follow-up. Mean values±SD on admission and follow up are shown in the solid lines with bars.

Discussion

In this study we evaluated serial CT images of type A IMH and demonstrated that the maximum aortic diameter of the ascending aorta in the initial CT images was predictive of progression or regression of type A IMH.

IMH was first described in 1920 as “dissection without intimal tear” and was considered a distinct entity at necropsy. The cause of IMH was believed to be rupture of the vasa vasorum in the aorta resulting in hematoma formation.³² The presence of aortic IMH can weaken the medial layer and potentially increase the likelihood of a classic aortic dissection. It has been reported that type A IMH should require early surgery because it tends to develop classic overt dissection or rupture.^{3 6 7 8 9 11} However, the clinical data available so far are too limited to draw any conclusion.

The previous studies reported that a formal distinction between type A and type B IMH might be justified with respect to both prognosis and treatment, and the site of IMH was an important parameter for determining prognosis.⁶ In the study by Robbins et al,³ all 3 patients with type A IMH ultimately underwent surgery. In the study by Mohr-Kahaly et al,⁴ 2 of 3 patients with type A IMH developed communicating dissection or outward rupture. However, in the study by Vilacosta et al,¹⁸ 3 of 8 patients with type A IMH showed favorable response to medical treatment. Similar clinical data were demonstrated in the previous reports.^{1 11} In addition, Sueyoshi et al¹⁹ reported that 7 of 13 patients with type A IMH had been doing well without surgical intervention. Their results were similar to our results. In our study, IMH regressed or disappeared in 12 (55%) of 22 patients, whereas 10 (45%) of 22 patients with type A IMH showed progression. All patients with progression of type A IMH showed enlargement of the affected aorta, and 6 of these patients had classic aortic dissection. The differences in frequencies of progression of type A IMH between the previous early studies and our study may result from differences in medical treatment or the imaging techniques used to establish the diagnosis of IMH. The patients in our investigation had close surveillance, including initial intensive care unit monitoring, aggressive blood pressure control, and frequent serial follow-up imaging studies. Furthermore, absence of dissection was documented in our study with TEE in all patients. This was not consistently performed in the previous studies. Thus it is unclear whether those patients in the previous early studies are directly comparable to the patients that we have reported.

In the current study, a Cox regression analysis revealed that the maximum aortic diameter in the initial CT images was the only significant predictor of progression of type A IMH. Furthermore, with a cutoff value of 50 mm in the maximum aortic diameter, the positive and negative predictive values for progression of IMH were 83% and 100%, respectively. This suggests that patients with type A IMH can be treated medically without surgery when the ascending aortic diameter is <50 mm and that surgical repair may be necessary when the ascending aortic diameter is ≥50 mm. This finding is in concordance with the previous study by Ide et al,²⁴ who reported that transition to a classic dissection was found exclusively in patients with a markedly dilated ascending aorta of >50 mm in diameter. However, their finding, based on relatively small numbers of type A IMH patients, was not strong enough to offer a distinct prediction of progression or regression of type A IMH.

Previous studies suggested that hematoma formation within the media of the aortic wall caused a structural weakness of the media, and both structural weakness of the media and mechanical stress might cause fusiform aneurysm or classic aortic dissection with an intimal flap.¹⁹ With respect to aortic enlargement, there are a few studies addressing the prediction of aortic enlargement throughout the entire follow-up period in classic aortic type B dissection.^{22 23} Masuda et al²² referred to maximum diameter of the dissected aorta during the acute phase as a risk factor for survival during the chronic phase. Kato et al²³ reported that aortic enlargement in chronic type B dissection can be predicted by using 2 factors obtained at the onset of dissection: maximum diameter of the dissected aorta and location of the primary entry site. They suggested that the enlargement of type B dissection was closely correlated with wall stress on the dissected aorta. However, there are anatomic differences between type A IMH and classic type B dissection, and wall stress is related to aortic lumen diameter in IMH. Considering the results that there were no significant differences in the maximum and mean lumen diameters between the progression group and the regression group in this study, some other factors such as distensibility may play an important role in progression or regression of type A IMH.

It is difficult to clarify the mechanism of progression or regression of type A IMH because of difficulty in detecting pathological changes in hematoma with the use of CT and TEE. Other imaging modalities, such as MR imaging, may have a possibility to detect pathological changes. MR imaging not only visualizes blood sequestration but also allows assessment of the age of the hematoma based on the formation of methemoglobin.³³ Nienaber et al⁶ reported that subacute IMH revealed a high signal intensity on both T1- and T2- weighted images caused by methemoglobin formation. Murray et al⁸ reported that MR images of patients who had early subacute complications showed signal intensity changes of hematoma that were consistent with recurrent bleeding. Although MR imaging of acutely ill patients can be problematic because of life-support and monitoring equipment, close serial follow-up of IMH with MR imaging may be useful to clarify what happens to the hematoma.

Surprisingly, our results do not suggest an increased risk associated with pericardial effusion or cardiac tamponade. In the setting of classic aortic dissection, cardiac tamponade is the most common cause of death.³⁴ Isselbacher et al³⁵ reported that cardiac tamponade was associated with an early mortality rate of 60%. They also suggested that pericardiocentesis in treating cardiac tamponade might be harmful rather than beneficial. However, few data are available for the treatment of cardiac tamponade in type A IMH.¹¹ In this study, 4 patients underwent successful pericardiocentesis with neither death nor progression of type A IMH immediately after the pericardiocentesis. Although the number of patients in this study is small, our results raise the possibility that pericardiocentesis is not necessarily harmful for type A IMH patients with cardiac tamponade.

There are several potential limitations in our study. First, pathological confirmation of IMH was not available in all cases. Although diagnosis was supported with other imaging modalities, it is possible that small intimal tears could have been missed with all imaging modalities. Second, we did not take into account the variations of normal aortic dimensions. The age-related normal values for the aortic dimensions may have to be taken into account. Third, 4 patients received medical therapy after the pericardiocentesis. Although no death occurred, the numbers of patients who underwent pericardiocentesis in this study were small enough, and most physicians and surgeons would recommend surgery in such patients. Considering no operative death in the surgical group in this study, surgery in patients with cardiac tamponade may be a safer route to follow.