Unruptured intracranial aneurysms occur in 7% of adults aged from 35–75 years old in China [1]. Rupture of these aneurysms may lead to subarachnoid hemorrhage that may be life-threatening to patients. The therapeutic approaches (such as surgical and endovascular interventions) are also associated with specific risks [2]. Therefore, it is important to evaluate the rupture risk of intracranial aneurysms. The rupture risk of aneurysms in symptomatic patients is higher than those in asymptomatic ones. Sentinel headaches and third nerve palsy are well-known symptoms that can precede aneurysmal rupture [3, 4]. However, both the specificity of sentinel headaches and sensitivity of third nerve palsy are low. Some clinical risk factors (such as gender, age, hypertension, dyslipidemia, diabetes, cigarette smoking and alcohol consumption) predispose to aneurysmal rupture [2].

Structural changes and the inflammatory response within the wall occur before aneurysm rupture [5]. Therefore, a clinical decision-making process depending solely on luminal characteristics (such as sac size, location, and morphology among other characteristics) observed by computed tomography angiography (CTA), magnetic resonance angiography (MRA) and digital subtraction angiography (DSA) may be misleading. High-resolution magnetic resonance imaging (HRMRI) has been used to investigate intracranial vascular lesions such as aneurysms, dissection and atherosclerosis. Circumferential uptake of superparamagnetic particles of iron oxide (ferumoxytol) in unruptured intracranial aneurysm walls indicating in vivo inflammation has been demonstrated, which may predict the rupture [6]. It has been reported that circumferential uptake of gadolinium (Gd) in aneurysm walls might be an indirect marker of inflammation, which could be related to the rupture [7]. The clinical significance of circumferential uptake of gadolinium in unruptured aneurysms has been mentioned but not reported separately [7, 8]. In the present study, the authors used a 3T gadolinium-enhanced HRMRI to investigate circumferential aneurysmal wall enhancement (CAWE) of the intracranial unruptured aneurysms of patients with sentinel headaches and third nerve palsy and compared it with clinical risk factors and aneurysm-related characteristics to demonstrate that CAWE may be associated with these symptoms.

Materials and Methods

Study Object

A retrospective review of the data of patients with unruptured aneurysms who underwent gadolinium-enhanced HRMRI examination from October 2014 to November 2016 in the Department of Interventional Neuroradiology of a tertiary institute was performed. The inclusion criteria were as follows:

  • in patients without any symptoms, the unruptured intracranial aneurysms were discovered as an incidental finding at the time of physical examination, and

  • symptomatic patients (ischemic cerebrovascular events were not included because of their various pathogenesis and symptoms), sentinel headaches (no history of headaches, a sudden headache on the ipsilateral sides of the aneurysms within 2 weeks prior to admission) [3, 9] or third nerve palsy (a sudden headache with one or several symptoms of pupillary light reflex disappearing, ptosis or extraocular myoparalysis on the ipsilateral sides of the aneurysms within 1 month prior to admission) [3, 10, 11] caused by the unruptured intracranial aneurysms.

In both case presentations, the possibilities of cerebral ischemia and subarachnoid hemorrhage (by CT and/or lumbar puncture) were excluded. Exclusion criteria were as follows:

  • patients who experienced a dissecting aneurysm,

  • patients who did not undergo digital angiography imaging,

  • patients who underwent MR imaging that was unqualified in evaluating CAWE,

  • patients who suffered from an unruptured aneurysm as well as a ruptured aneurysm, and

  • patients who had incomplete clinical records.

The study was approved by an institutional review board. Aneurysm status was also categorized as symptomatic or asymptomatic.

Data Acquisition

According to the findings of digital subtraction angiography (FD20, Philips, The Netherlands), the following aneurysm-related characteristics were recorded:

  • sac size

  • location (anterior circulation or posterior circulation) and

  • the number of aneurysms (single or multiple).

Based on medical records and laboratory findings, the following clinical risk factors for intracranial aneurysm (growth/rupture) were recorded:

  • age

  • sex (male or female)

  • history of subarachnoid hemorrhage (yes or no)

  • hypertension (yes or no)

  • dyslipidemia (yes or no)

  • diabetes (yes or no)

  • cigarette smoking (yes or no) and

  • alcohol consumption (yes or no).

Imaging Protocol and Analysis

HRMRI was performed using a 3T MR imaging scanner (Verio, Siemens Healthcare, Erlangen, Germany) with a 16-channel head coil. The protocol included three-dimensional (3D) time of flight (TOF) MR angiography, a T1-weighted imaging sequence and a T2-weighted imaging sequence. The scanning parameters of the HRMRI sequences (T1 WI, and T2 WI) were the same, with 130 × 130 mm field of view, an acquired matrix of 512 × 512, a slice thickness of 2.0 mm, a slab thickness of 0.2 mm, a flip angle of 180°; a scan comprising 5 slices, a T1 WI repetition time/echo time of 861/18 ms, a T2 WI repetition time/echo time of 903/83 ms. Gadopentetate dimeglumine (Magnevist; Bayer HealthCare Pharmaceuticals) was administered intravenously (0.1 mmol/kg), and the T1 WI was repeated 5 min after the contrast agent was infused. A fat saturation technique was used in T1 WI and T2 WI.

In this study, two experienced neurovascular radiologists (10 and 15 years of experience in neurovascular imaging) analyzed pre- and postcontrast T1-weighted images, retrospectively, and determined whether there was a CAWE. Before analysis, the readers were blinded to the clinical data of the patients and other sequences except for 3D-TOF imaging. Discordances between two readers were resolved by consensus. The CAWE on postcontrast T1-wighted images would have to fulfill both criteria, namely, circumferential and clear enhancement.

Statistical Analysis

Categorical and continuous variables were expressed as frequencies with percentages and means±SD, respectively. The agreement between two observers for the presence of CAWE was evaluated by a kappa value. The abovementioned aneurysm-related characteristics and clinical risk factors were compared between two groups using χ2 or Fisher’s exact test for categorical variables and Student’s t‑test for continuous variables where appropriate. Multivariable logistic regression was performed to determine factors independently associated with the symptoms of patients, applying variables that achieved a significance of P < 0.2 in the univariable analysis. Significance was defined as a 2-sided P value < 0.05. Data were analyzed with the JMP statistical package (Version 9.2; SAS Institute, Cary, NC, USA).

Results

During this study, among the 39 included patients (47 aneurysms), 2 patients were excluded (dissecting aneurysm). Subarachnoid hemorrhage was excluded by both computed tomography (CT) and lumbar puncture in 4/39 (10.3%) patients and by either CT or lumbar puncture in 35/39 (89.7%) patients. The final study sample included 37 patients (59 ± 10 years; 28 females and 9 males) with 45 aneurysms (22 asymptomatic and 23 symptomatic). Among the 37 patients, 20 patients did not have symptoms and 17 patients were symptomatic (10 sentinel headaches and 7 third nerve palsies). Interobserver agreement for CAWE presence was good (kappa = 0.8225; 95% CI 0.657–0.988). Discordances between two readers were resolved by a consensus in 4 cases. CAWE was observed in 22/45 (48.9%) unruptured aneurysms (Table 1). Patient demographics including aneurysm rupture risk factors and aneurysm-related characteristics are summarized in Table 2. CAWE was more frequently observed in symptomatic aneurysms than in asymptomatic ones (16/23, 75% versus 6/22, 23.8%, respectively; P < 0.05). CAWE was observed in 9/15 aneurysms in patients with sentinel headaches and 7/8 aneurysms in patients with third nerve palsy (Fig. 1). Multivariable regression analysis (Table 3) demonstrated that CAWE was the only independent factor associated with symptoms (odds ratio = 5.17, 95% CI 1.30–20.52; P = 0.02).

Table 1 The aneurysm count and circumferential aneurysmal wall enhancement (CAWE) results in both symptomatic and asymptomatic patients
Full size table
Table 2 Patients characteristics and aneurysms of symptomatic and asymptomatic groups
Full size table
Table 3 Multivariable logistic regression analysis for factors associated with symptomatic aneurysms
Full size table
Fig. 1
figure 1

DSA (first column), precontrast T1-weighted HRMRI (second column), and postcontrast T1-weighted HRMRI (third column) of 3 patients with or without symptoms (a no symptoms, b sentinel headaches and c third nerve palsy). a Female, 60 years old, left internal carotid artery terminus unruptured aneurysm of a patient with no symptoms. DSA exhibits the aneurysm (a i, arrow). Precontrast T1-weighted HRMRI shows the aneurysm wall (a ii, arrow). Postcontrast T1-weighted HRMRI shows aneurysm without circumferential wall enhancement (a iii, arrow). b Female, 52 years old, left internal carotid artery terminus unruptured aneurysm in a patient with sentinel headaches. DSA shows the location of an aneurysm (b i, arrow). Precontrast T1-weighted HRMRI shows the aneurysm wall (b ii, arrow). Postcontrast T1-weighted HRMRI shows an aneurysm with circumferential wall enhancement (b iii, arrow). c Male, 53 years old, right internal carotid artery-posterior communicating artery unruptured aneurysm from a patient with ipsilateral third nerve palsy. DSA exhibits the aneurysm (c i, arrow). Precontrast T1-weighted HRMRI shows the aneurysm wall (c ii, arrow) and partial thrombus in the sac. Postcontrast T1-weighted HRMRI shows an aneurysm with circumferential wall enhancement (c iii, arrow) DSA digital subtraction angiography, HRMRI high resolution magnetic resonance imaging

Full size image

Discussion

Our study used 3T gadolinium-enhanced HRMRI to demonstrate that circumferential aneurysm wall enhancement is more frequently observed in symptomatic unruptured aneurysms than in asymptomatic ones.

The risk of rupture of symptomatic aneurysms is likely higher than that of asymptomatic ones. It has been demonstrated that sentinel headaches or third nerve palsy caused by unruptured aneurysms can precede rupture. Aneurysm growth was found to be associated with rupture [12]. The increased size of aneurysms may lead to sentinel headaches and third nerve palsy [3]. A sentinel headache occurs in 10–43% of subarachnoid hemorrhage (SAH) patients [13]. It seems that the warning leak or stretching of the aneurysmal wall are the cause of these headaches. These symptoms usually occur within 2 weeks preceding SAH. The direct compression or the pulsatility of the aneurysms near the nerve have been considered as the causes of third nerve palsy, which can lead to diplopia, eye pain, and ptosis, among other symptoms [4]. However, diabetic microvasculopathy could also lead to third nerve palsy [11, 14]. There was one patient with third nerve palsy in our study who also had diabetes. Her symptoms disappeared after interventional therapy, which were then favored not to be related to diabetes. Therefore, it is meaningful to evaluate the risk of rupture of aneurysms by imaging instead of clinical assessment alone. The CAWE may provide additional clinical information. It has been demonstrated as being a sign of aneurysmal instability [7, 15, 16]. Also, it is related to inflammation, as previously demonstrated by histopathological–radiological correlation studies [8], one of these studying the extracranial carotid artery [17]. The inflammation could be the cause of the symptoms. It may then be appropriate to use contrast-enhanced images in clinical practice.

Recently, wall enhancement of a cerebral aneurysm has been revealed as a marker of rupture using vessel wall MR imaging. Matouk et al. [18] investigated 5 patients (2 patients with single aneurysms and 3 patients with multiple aneurysms) with aneurysmal subarachnoid hemorrhage and observed that wall enhancement was associated with ruptured aneurysms. They inferred that aneurysm wall inflammation may lead to the enhancement. Nagahata et al. [15] investigated 144 aneurysms including 61 ruptured and observed wall enhancement in 60 ruptured aneurysms on postcontrast MRI. They thought inflammation within the aneurysm wall may be one of the possible mechanisms underlying the enhancement. Edjlali et al. [7] investigated 108 aneurysms including 17 ruptured aneurysms and observed wall enhancement in 16 ruptured aneurysms. These reports also considered aneurysmal wall inflammation as the cause of wall enhancement. Bhogal et al. [19] investigated 1 patient (multiple bilateral aneurysms) and confirmed vessel wall enhancement was related to ruptured aneurysm during the surgery. Omodaka et al. [16] investigated 104 aneurysms including 28 ruptured ones and found that the wall enhancement was greater in ruptured aneurysms. Hu et al. [8] investigated 30 aneurysms including 6 ruptured ones, and observed wall enhancements in all of these. Histological analysis demonstrated inflammatory cells in the aneurysms with wall enhancement. All the studies described above suggest that wall enhancement is a feature of aneurysm rupture.

The clinical significance of wall enhancement on MR imaging of unruptured aneurysms remains ambiguous. Hasan et al. [6] investigated 30 unruptured aneurysms using superparamagnetic particles of iron oxide (ferumoxytol) as a contrast agent and observed that the aneurysm might rupture within 6 months if the wall showed enhancement within the first 24 h. Their histological findings demonstrated that inflammation of the aneurysmal wall was the reason for the enhancement. The studies of Edjlali [7], Omodaka [16] and Nagahata [15] included unruptured aneurysms, but they did not analyze the relationship between CAWE and symptoms (sentinel headaches and third nerve palsy). Hu et al. [8] suggested that the wall enhancement of unruptured aneurysms may be related to patients’ symptoms. Vakil et al. [20] investigated 27 unruptured aneurysms and demonstrated that the contrast permeability rate of the aneurysm wall had a positive correlation with the risk of rupture. Liu et al. [21] investigated 61 unruptured aneurysms and observed wall enhancement in 33 aneurysms. They found that wall enhancement was related to sac size and suggested that inflammation caused the enhancement. To take this a step further, CAWE helps in distinguishing which aneurysm(s) is (are) symptomatic in patients with multiple aneurysms. In our study, CAWE was not observed in 7/12 aneurysms in 6 patients with symptoms (6/7 aneurysms in the patients with sentinel headaches and 1/7 aneurysms in a patient with third nerve palsy). Among the 6 aneurysms without wall enhancement in 5 patients with sentinel headaches; 4 aneurysms were in 3 patients harboring multiple aneurysms (1/3 aneurysms, 2/3 aneurysms and 1/2 aneurysms) and 2 aneurysms were in 2 patients with a single aneurysm. These 4 aneurysms may not be the culprits because they were contralateral to the headaches and did not have CAWE, and the other coexisting multiple aneurysms in the corresponding patients were ipsilateral to the headaches and had CAWE (making them responsible for the symptoms). The aneurysms in the 2 patients with a single aneurysm were on the ipsilateral sides of the headaches and did not have CAWE, so the warning leak may not be the cause of the headaches. Also, the only aneurysm without wall enhancement in a patient with third nerve palsy was in a case of mirror aneurysms. This aneurysm may not be the culprit because it was contralateral to the third nerve palsy and headaches and did not have CAWE, and the other existing aneurysm in this patient was ipsilateral to the third nerve palsy and headaches and had CAWE (indicating it is the symptomatic one). However, 6 of 22 asymptomatic aneurysms also demonstrated wall enhancement. The wall enhancement index may have a positive relationship with the risk of rupture of aneurysms [16]. Maybe the inflammation caused by the histopathological change of the aneurysm wall was not enough to lead to the patient symptoms.

This study has several limitations. First, there was a selection bias because the samples were obtained in a tertiary center, where there was a high proportion of patients with symptomatic aneurysms. This selection bias could be why nearly half of the aneurysms were symptomatic in our study. Second, there was an information bias because aneurysm-related characteristics could be obtained by MR imaging when they were analyzed. Third, we did not conduct a histopathological–radiological correlation study because all patients underwent endovascular treatment. Fourth, our study did not conduct adequate follow-ups of the aneurysm wall enhancement because of the necessary interventions and/or lost to follow-up. Finally, we did not analyze the enhancement index of the aneurysm wall due to lack of postprocessing software. To confirm the clinical significance of wall enhancement of the unruptured aneurysms, further study is needed to investigate the relationship between the enhancement index of the aneurysm wall and the symptoms associated with unruptured aneurysms.

Conclusion

Our study demonstrates that CAWE of aneurysms on 3T Gd-enhanced HRMRI correlated with sentinel headaches and third nerve palsy caused by unruptured aneurysms, which may be helpful to distinguish the cause of these symptoms. Further studies are needed to confirm whether CAWE can provide useful information to assess the rupture risk of unruptured aneurysms through supportive evidence (lumbar puncture, histopathology and/or follow-up).