A New Paradigm: How to Study the Exact Location of a Paraclinoid Aneurysm and the Cavernous Sinus in the Preoperative Stage through Imaging

Abstract

Intracranial aneurysms (IAs) found in the subarachnoid compartment of the internal carotid artery are at risk of rupturing and producing hemorrhage into this space, producing all the known serious pathological manifestations of subarachnoid hemorrhage. On the other hand, those who are exclusively in the intracavernous segment have this negligible risk. It is in this context that a peculiar class of IAs fits, the paraclinoid aneurysms, which are characterized by their complex anatomical relationships with the optic apparatus, anterior clinoid process, the first supraclinoidal arterial branches of the carotid artery, the oculomotor nerve and the segment mainly said clinoid, which is delimited by the proximal and distal dural rings. It is of crucial importance, and challenging, to determine the location of paraclinoid aneurysms and their exact relationship with the cavernous sinus, given the need to establish an adequate line of treatment for each case. Through preoperative studies of 3 t Magnetic Resonance, comparison with 3D bio models and microsurgical exploration, it was possible to accurately demarcate these anatomical relationships. Therefore, today it is possible to establish the accurate location of the paraclinoid aneurysm in relation to the cavernous sinus by means of MR images and to carry out an adequate, effective, and safe operative planning.

1. Introduction

Usually, arteries in the human body have 3 layers of tissue called, from the vascular lumen to the outermost layer, the intima, media, and adventitia, respectively. The intima is composed of an endothelial tissue layer, which has direct contact with intraluminal blood, and a subendothelial layer, formed by connective tissue. The internal elastic lamina, a layer of elastic fibers, separates the tunica intima from the tunica media. The media is basically formed by concentrically organized smooth muscle fibers and collagen fibers, in which type III collagen predominates. The external elastic lamina separates the tunica media and adventitia, typical in arteries throughout the body but absent in intracranial arteries. Finally, the adventitia is mainly formed by a complex network interspersed with type I collagen, elastin, fibroblasts, nerves, and vasa vasorum. Therefore, the wall of the cerebral arteries has a different structure from other extracranial arteries, with a scarce adventitia and a low proportion of elastic fibers. Furthermore, they are immersed in the cerebrospinal fluid of the subarachnoid space instead of the surrounding connective tissue [1, 2].

In normal arteries, myointimal hyperplasia is an adaptive physiological reaction to hemodynamic stress or to a mechanism of vascular injury, resulting from a change in the phenotype of smooth muscle cells in the tunica media, which promote their migration and proliferation, outlining the lesion endothelium [3]. Once the molecular mechanisms become unable to compensate for the myointimal injury, cellular and humoral inflammatory responses are triggered (mainly responsible for aneurysm formation) [4, 5, 6], and responses mediated by inflammatory cytokines, such as tumor necrosis factor (TNF).), interleukin-β (IL-1β) and matrix metalloproteinases (MMPs), which promote macrophage influx and continuous degradation of collagen and elastin fibers [6, 7, 8, 9]. IA is, therefore, an encephalic vascular lesion characterized by an abnormal dilation of the blood vessels of the brain, affecting, in general, about 5% of the population [7, 10, 11, 12, 13], resulting from a molecular and hemodynamic imbalance, the which would explain its formation in arterial junctions and bifurcations, where an excess of hemodynamic stress is exerted on the vessel wall, followed, then, by a local inflammatory process that leads to disruption of the internal elastic lamina [14, 15, 16, 17].

Once the IA ruptures, blood leaks into the subarachnoid space, the natural space between the arachnoid mater and the pia mater, producing spontaneous subarachnoid hemorrhage (SAH). It is an entity with high mortality, reaching 50% of affected individuals, and considerable morbidity among survivors. It is noteworthy the fact that approximately 30% of subarachnoid hemorrhages resulting from intracranial aneurysm rupture occur during sleep. This denotes a multifactorial pathophysiological nature fundamentally related to inflammation and not exclusively hemodynamic, since in this period hemodynamic stress tends to be less intense than during the day. In the natural history of IA there are known modifiable and non-modifiable risk factors for rupture. As an example of the former, we can mention smoking and high blood pressure. Non-modifiable factors include advanced age, genetic profile, and family history of SAH, growth of the aneurysmal sac, among others [7, 13].

Bouthillier and van Loveren’s classification (Figure 1), separates the internal carotid artery into segments according to its branches and anatomical relationships with adjacent structures. This makes it possible to classify carotid saccular ais according to these segments and their respective vessels. Thus, IA’s can be in the cavernous, clinoid, ophthalmic and/or posterior communicating segments of the carotid artery and are primarily related to an arterial branch of the carotid artery or to an anatomical structure of interest.

Figure 1.

Anatomical structure of the internal carotid artery (ICA). Segments: petrous (C2), lacerus (C3), cavernous (C4), clinoid (C5), ophthalmic (C6), proximal and distal dural rings, III cranial nerve and cavernous sinus. By the classification of Bouthhillier and van Loveren. On the left, schematic drawing and on the right, a piece obtained through anatomical dissection (adapted from Boutillier and Van Loveren left and right anatomical dissection by Hugo Doria, MD PhD).

As for the paraclinoid IAs, specifically, since 1968 with the work of Drake et al. [18], anatomical patterns are studied to try to classify them. However, until today we find confounding factors that add difficulty in understanding these classifications. Much of this stems from the fact that the existing nomenclature of aneurysms arising from the clinoid and ophthalmic segments of the ICA is contradictory, mainly for anatomical reasons [19, 20]. First: the ophthalmic artery can arise both from the clinoid segment (C5) and from the ophthalmic segment (C6) [21] of the ICA as previously exposed [12, 19]. Second: aneurysms in this region do not necessarily arise in relation to a named arterial branch [19, 20]. Third: aneurysms in this area can be intradural, extradural or transitional and sometimes it is impossible to make this determination using radiographic investigations currently available [20]. Fourth: the recognition of the carotid cavum as an entity further complicated the issue, as cavum aneurysms are located below the plane of the distal dural ring (DDR), but are intradural [20, 22, 23]. In a summarized and practical way, the proposed classifications try to establish some standard for the surgical technique, always in search of the basic principles of the management of cerebral aneurysms, regardless of their location (establishing proximal and distal vascular control; adequate exposure of the neck and complete obliteration of the aneurysm with maintenance of cerebral blood flow distal to the aneurysm) [23].

Paraclinoid aneurysms are lesions that originate in the cavernous, clinoid, or ophthalmic segments of the ICA, defined by Bouthillier as segments C4, C5 and C6, respectively (Figure 2).

Figure 2.

A - Angiography with digital subtraction in the AP showing a large multi-lobulated paraclinoid aneurysm with a medial conformation. B – Microsurgical image of the same aneurysm after anterior intradural clinoidectomy. 1 – Optic Nerve, 2 – Aneurysm, 3 – Left Carotid Artery, black arrow – emergence of the ophthalmic artery (angiography and microsurgical photography kindly provided by Sergio Tadeu Fernandes, MD PhD).

These are aneurysms that may arise proximally to the proximal dural ring (PDR), between the dural rings or distally to the DDR, between the distal dural ring and the posterior communicating artery and may be intra or extracavernous. Extracavernous paraclinoid aneurysms present a risk of SAH and usually require treatment, while unequivocally intracavernous aneurysms are located completely below the proximal dural ring, rarely coursing with SAH and present lower morbidity than aneurysms originating from the intradural space [24].

Approximately 33–59% of paraclinoid IAs are associated with the ophthalmic artery; 27–47% are associated with the superior hypophyseal artery and between 14% and 20% are not associated with any arterial branch. Paraclinoid IAs comprise between 1.4% to 9.1% of all ruptured aneurysms [25].

They are considered uncommon, accounting for approximately 5% of IAs, reaching up to 14% of IAs, in some studies, with an increased prevalence in women [26, 27, 28]. These aneurysms affect the ICA between the cavernous segment and the origin of the posterior communicating artery.

As for the diagnosis, the available methods are classified as invasive and non-invasive. The invasive method is digital subtraction angiography (DSA) performed in a hemodynamic suite through selective arterial catheterization of the intracranial vascular tree, which allows, in addition to the structural study, the analysis of hemodynamic behavior in real time. Despite being invasive and with intrinsic risks, this method is still consolidated as the gold standard for this purpose [29, 30]. On the other hand, non-invasive methods include the use of images processed by computer graphics and three-dimensional reconstructions obtained by Computed Tomography (CT-Angio) or Magnetic Resonance (MR-Angio) devices that are increasingly sensitive and specific, often used not only as screening, but replacing DSA in selected cases with the additional benefit of contributing, not only with the visualization of the target pathology, but its relationships with adjacent structures, such as the anterior clinoid process [31, 32].

However, despite all the technological advances and modern imaging techniques validated so far, both by invasive and non-invasive methods, paraclinoid aneurysms still represent a separate challenge. In view of the difficulty in determining whether the aneurysm studied is located exclusively in the cavernous compartment of the carotid artery, or whether it has a relationship, even if partial, with the subarachnoid space. The practical significance of such information is that each type of aneurysm requires a different surgical strategy. Aneurysms identified as being completely intradural may not require anterior clinoidectomy. On the other hand, transitional aneurysms may require a wide opening of the dural rings and adequate management of the roof of the cavernous sinus. Still, those located completely in the intracavernous space rarely require any kind of approach. Until then, it was only possible to determine this exact relationship through microsurgical exploration.

Many strategies have emerged with the purpose of resolving this dilemma. The proposal to use the origin of the ophthalmic artery as a marker for the intradural ICA, making a distinction between the intra and extradural segments, had some relevance, but it was soon found that this anatomical marker had low accuracy, since the origin of the artery Ophthalmic is extradural, that is, proximal to the distal dural ring, in 2–16% of cases [11, 12, 19, 23, 33, 34, 35]. There was a proposal to use the base of the ACP in lateral radiographs, serving as a more reliable marker than the origin of the OphA in angiograms, which also proved to be of low accuracy because, for example, carotid cavum aneurysms can be observed below the level of the ACP and, even so, they are inside the intradural space. Subsequently, Oikawa et al. [36] proposed that the use of the anterior clinoid process (ACP) in lateral projection radiographs should be replaced by the sellar tubercle in the same projection, when evaluating aneurysms on the medial side of the dural ring, as this is more proximal than the lateral one [36]. Kim et al. stated that they were not aware of any combination of radiographic exams that allowed the reliable identification of the distal dural ring, reaffirming that “surgical exploration is the only solution in these cases” [19]. In 2001, Murayama et al. proposed the use of 3D CTA as an indirect method for identifying the distal dural ring, noting that in 84.8% of the evaluated images it was possible to identify a concave impression on the anterior curve of the ICA and suggested that this concavity coincides with the location of the distal dural ring, because of ring fixation to the ICA [37]. Gonzalez et al. postulated that, if the optic pillar could be reliably identified with high-resolution CTA, it could represent an anatomical landmark for evaluating aneurysms in this critical region [24]. Hashimoto et al. applied the same methodology for analysis of the optical pillar, through CTA images, comparing images and intraoperative findings and stated that the optical pillar is the most useful landmark for operative planning of aneurysms in this region [38].

With the purpose of changing the indirect reference of the dural rings previously studied by previous methods, magnetic resonance imaging (MRI) begins to be used for direct visualization of the distal dural ring and the limits of the cavernous sinus in relation to paraclinoid aneurysms. However, the method initially ran into difficulties such as low spatial resolution and the need to improve signal acquisition powers. Thines et al. proposed the improvement of the resolution in 3 Tesla weighted in T2, demonstrating in thin and contiguous sections, the dural folds of the roof of the cavernous sinus and the distal dural ring, however, they were not compared with surgical findings or post-mortem dissections [39, 40]. In 2019, Obusez et al. evaluated the use of MR imaging of the vessel wall (“VW-MR”) to determine the exact location of unruptured paraclinoid aneurysms in relation to DDR but, once again, the study ran into low statistical power and there was no comparison with surgery or necropsy study to verify the findings [41]. Therefore, until now, there is no knowledge of accuracy studies for the diagnostic tests of paraclinoid aneurysms and their relationship with the cavernous sinus. The studies found are series of cases that did not determine sensitivity, specificity, predictive values, and likelihood ratios [24, 37, 38, 39, 40, 42, 43, 44, 45, 46].

Therefore, the preoperative identification of the distal dural ring and the actual definition of the limits of the roof of the cavernous sinus in relation to paraclinoid aneurysms remains an unresolved problem. This dilemma stimulated the studies of the authors of this chapter to develop more adequate preoperative evaluation protocols and proposed the adequate formatting of 3-tesla MRI studies with sensitivity and specificity sufficiently capable of determining the exact location of these AI in relation to the cavernous sinus, the which allows an effective diagnosis and enables a more adequate, safe, and efficient surgical planning. Next, we will describe the formulation of these protocols of great interest for neuroradiological and neurosurgical practice.

2. Original studies that define the precise anatomical relationship between the intradural and intracavernous compartments applied to the context of paraclinoid aneurysms

2.1 Method based on 3t MR images compared with 3D models, anatomy in cadaveric and microsurgical specimen

This method documented the correlation between the oculomotor nerve and the internal carotid artery, assuming their intersection, visualized in a 3 T magnetic resonance study and confirmed from printed three-dimensional biomodels and microsurgery, as a new anatomical-radiological paradigm that marks the upper limit of the cavernous segment of the internal carotid artery, distinguishing paraclinoid intracranial aneurysms. This is a retrospective study carried out in four stages: anatomical, radiological, 3D printing stage and surgical stage. The internal carotid arteries were dissected in their clinoid topography of 10 cadaveric specimens, totaling 20 cerebral hemispheres. Magnetic resonance images and 3D biomodels of 42 aneurysms from 34 patients were analyzed [47].

Magnetic resonance imaging and MRI angiography were performed in a 3 Tesla (3 T) (Siemens-Skyra Evolve, Erlangen, Germany) and GE (GE Health care, HDXT, Milwaukee, USA) machine using a 32- and 8-channel dedicated skull coil, with the objective of identifying the course of the III nerve when crossing laterally with the ICA, inferring at this point the presence of the carotid-oculomotor membrane and correlating this point to the paraclinoid aneurysm under study (Figure 3).

Figure 3.

(A) Bilateral identification of NO on 3 T MRI in T2, coronal sections. In this section, both ONs are contained in the interpeduncular cistern between the posterior cerebral artery, superiorly, and the superior cerebellar artery, inferiorly. We used this cut to begin following ON on its path from the midbrain to the cavernous sinus, in view of the ease with which ON could be identified without a doubt. (B) Location of ON (yellow arrows) and ICA (red arrows) left and right, on 3 T MRI T2 sequence coronal thin slices. Emphasis on the cerebrospinal fluid cistern adjacent to the ON, halo of hypersignal provided by the cerebrospinal fluid, circumcising the ON in the imminence of its entry into the posterior roof of the cavernous sinus, in the topography of the ON triangle. (C) Location of ON (yellow arrow), ICA (red arrow) and left AI (orange arrow) on 3 T MRI in T2 sequence, axial view. Again, emphasis on the halo provided by the cerebrospinal fluid circumcising the ON, a cistern adjacent to the ON, exactly on the verge of its entry into the posterior roof of the cavernous sinus, within the ON triangle. We used this cistern as a marker to identify cranial nerve III in its path from posterior to anterior, in its transition from the cisternal portion (posterior) to the intracavernous portion (anterior). (D) Identification of ON (yellow arrow) at its intersection with the paraclinoid ICA (red arrow) in the cavernous sinus. The ACP (blue arrow) can be seen just above the NO. The dashed line connecting the upper border of the ON, laterally, to the ICA, medially, is used for evaluation and classification of the AI under study, classified as superior or extracavernous in this image – the entire neck of the aneurysm is found above the upper border of the ON. Caption: ICA – internal carotid artery, NO – oculomotor nerve, ACP – anterior clinoid process, OphA – ophthalmic artery, ON – optic nerve, OS – optic strut, DS – diaphragm sellae, IA – intracranial aneurysm, CSF – cerebrospinal fluid (images kindly provided by Dr. Hugo Doria, MD PhD [47]).

High-resolution images were acquired, following the established protocol, in 2D Coronal T2 Fast Spin Echo (FSE) sequences with thin sections, intracranial arterial magnetic AngioMRI in the 3D TOF technique with and without gadolinium, for vascular analysis. The detailed protocol of imaging studies can be found in [47].

The following structures were analyzed: Identification of the internal carotid artery ICA, identification of the oculomotor nerve in all its extension. Identification of anatomical repairs - Anterior Clinoidal Process (ACP), Ophthalmic Artery (OphA), Optic Nerve (ON), Optical Strut (OS), Diaphragm Sellae (DS); Identification of the paraclinoid aneurysm in the patient under study and the relationships of its neck and dome with the III cranial nerve. The intersection between the internal carotid artery (ICA) and the III nerve was identified on CORONAL T2 / CORONAL FIESTA / volumetric T1 with intravenous contrast and post-gadolinium AXIAL 3D TOF sequences.

According to their relationship with the III nerve, paraclinoid aneurysms were classified as follows: Superior to the superior border of the III nerve, whose neck and dome are located distal to the intersection ICA X ON, in an extracavernous location; at the level of the superior border of the III nerve or in a transitional location (with a part of the aneurysmal neck or dome located superiorly and another part inferiorly to the superior border of the ON in its intersection with the ICA); and inferior to the upper border of the oculomotor nerve, in an intracavernous location, when the aneurysm neck and the dome are located below the ICA x III nerve intersection.

To obtain the 3D model, it was necessary to compile Computed Tomography (CT) and 3-tesla Magnetic Resonance images and computational processing for conversion into STL format (STereoLithography – Stereolithography or triangular pattern language - 3D file of the region of interest). These data were downloaded to the 3D printer, which deposited the chosen material layer by layer, thus forming the desired object on a scale of 140% of the original size for a better visualization of the structures (Figure 4).

Figure 4.

3D biomodel made from radiological images of patients, showing the structures: ICA – internal carotid artery, printed in red; IA – intracranial aneurysm, printed in black; NO – oculomotor nerve, printed in yellow; bone at the base of the skull, printed in white. In A: evidence of IA classified as extracavernous or greater than the upper limit of the ON at its intersection with the ICA. In B: evidence of IA classified as intracavernous or inferior to the upper limit of the ON at its intersection with the ICA. In transparent acrylic, the supports for supporting the anatomical structures in their exact anatomical positions after three-dimensional printing were printed (images kindly provided by Dr. Hugo Doria, MD PhD [47]).

Of the 34 patients participating in the study, with a total of 42 intracranial aneurysms, 20 patients, totaling 23 aneurysms, underwent intracranial vascular microsurgery for clipping the paraclinoid aneurysm. In the comparative analysis between the radiological study and the 3D model of these 42 cases, 40 (95.23%) were considered compatible and of these, 36 (90%) obtained total compatibility in the 3 (neuroradiologist 1 / neuroradiologist 2 / 3D biomodel) evaluations and classifications and 4 (10%) obtained compatibility between 2 of the 3 evaluators (neuroradiologist 1 or neuroradiologist 2 compatible with the 3D biomodel), and the three-dimensional biomodel is the parameter of success in view of its total accuracy and anatomical-radiological reliability (Figure 5).

Figure 5.

Comparison between the classification of paraclinoidal aneurysms by the two neuroradiologists, in a blind and independent manner, and the classification by the printed three-dimensional biomodel. The radiologists, blindly and independently, classified the aneurysms according to their relationship between the aneurysmal neck and the intersection between the ON and the ICA, as superior, transitional, or inferior to the superior limit of the III cranial nerve, while the author of the work classified the aneurysms based on the processing of CT, MR and Angio RM 3 Tesla images and analysis of the three-dimensional biomodel from the 3D printing (Modification and Publication authorized by Dr. Hugo Doria, MD PhD [47]).

The results together indicate that, in the impossibility of directly identifying the ADP, the identification of the upper limit of the III cranial nerve immediately lateral to the ICA, in all its diameter, and the distance between the III cranial nerve and the ICA can be considered a landmark for delimitation of the roof of the cavernous sinus distinguishing intracavernous ICA and extracavernous ICA since the measurements were close both in the 20 brain hemispheres dissected in 10 anatomical specimens (average distance of 1.19 mm - ranging from 0.6 mm to 1.7 mm) and in the 34 patients studied radiologically (average distance of 1.09 mm - ranging from 0.4 mm to 2.6 mm).

Of the 42 aneurysms studied, twenty-three (54.76%) underwent intracranial vascular microsurgical treatment by clipping, which confirmed the classification of the aneurysm as extracavernous, corroborating the findings in the anatomical specimens and in the radiological analyzes and printed 3D biomodel and indicated that the distance between the III cranial nerve and the ICA can be a landmark for delimiting the ceiling of the cavernous sinus, distinguishing intracavernous ICA and extracavernous ICA.

In summary, in cadaveric specimens, totaling 20 cavernous sinuses studied, we identified that the upper limit of the cavernous sinus is determined by the carotid-oculomotor membrane (COMM), which closely correlates to the intersection between the internal carotid artery and the oculomotor nerve, crossing it, transversely across its entire diameter.

Corroborating the anatomical step, we identified the intersection between the oculomotor nerve and the internal carotid artery in 3 Tesla brain magnetic resonance images of 42 aneurysms. The intersection between the oculomotor nerve and the internal carotid artery was established as a new anatomical-radiological landmark for paraclinoid aneurysms in terms of the carotid segment in which they are contained, intra or extracavernous; The 3D biomodel confirmed the radiological precision for the exact location of the paraclinoid aneurysms, showing high compatibility for the location of the analyzed paraclinoid aneurysms. The surgical procedure performed in 23 aneurysms confirmed this legend and allows formulating a new paradigm for classifying paraclinoid aneurysms, between: extracavernous or superior, intracavernous, or inferior, or transitional in the preoperative stage, thus avoiding surgical exploration and its associated risks.

2.2 Method based on 3t MR images compared with microsurgical anatomy

This is a cross-sectional clinical study of diagnostic accuracy that analyzed a prospective cohort of 20 patients totaling 25 paraclinoid aneurysms in a single hospital center in São Paulo in the period between 2014 and 2018 [48].

The patients underwent Cerebral Angiography with Digital Subtraction, which characterized the sample with 10 cavernous and 15 non-cavernous aneurysms, as shown in the table below (Table 1).

		Localization
Variable	General (25)	Cavernous (10)	Not Cavernous (15)	P Value
Age (average ± DP)	51,4 ± 11,5	48,5 ± 13,8	53,3 ± 9,7	0,313
Feminine Gender	24 (96)	10 (100)	14 (93,3)	1000
Size (median e quartiles)	5,0 (4,0 – 5,5)	5,0 (3,0 – 5,3)	5,0 (4,0 – 6,0)	0,511
Guidance				0,322
Superior	7 (28,0)	2 (20,0)	5 (33,3)
Medial	15 (60,0)	7 (70,0)	8 (53,3)
Lateral	2 (8,0)	0 (0,0)	2 (13,3)
Inferior	1 (4,0)	1 (10,0)	0 (0,0)

Table 1.

Characterization of the sample according to age and gender of patients and size and location of aneurysms, according to Kristh et al [20].

Data presented as n (%) unless specified. DP (standard deviation)

The same sample of patients underwent a 3-tesla MRI study with the specific protocol detailed in [48]. The following structures were analyzed: distal dural ring (DDA); proximal dural ring (PDR); Anterior Clinoid Process (ACP), ICA, Ophthalmic Artery (OphA), Optic Nerve (ON), Optic Strut (OS), Diaphragm Sellae (DS); identification of the paraclinoid aneurysm and the relationships of its neck and dome with adjacent structures. DDR has been identified as the reflection of the dura mater surrounding the ICA as it leaves the roof of the cavernous sinus. It is contained in a curved dural plane that projects inferomedially between the median crest of the superior surface of the PDA and the diaphragm sellae [39, 40, 42, 49, 50, 51]. The DDR also extends infero-posteriorly between the floor of the optic canal and the posterior part of the roof of the cavernous sinus as illustrated in Figures 6 and 7.

Figure 6.

3-T MRI in T2-weighted turbo-spin sequence in the coronal plane (anterior to posterior, from “A” to “I”, respectively) according to the protocol of the present study, demonstrating the anatomo-radiological markers of the paraclinoid region. Note that the optic strut (OS) is identified on MRI as the shade in green and a paraclinoid aneurysm on the right is considered transitional – note in “I”, blue arrow, how the most posterior portion of the aneurysm projects into the subarachnoid space. ACP and OS, green; ON, golden; ICA, anterior loop of the internal carotid artery – “C” to “H”, red; ICA, cavernous internal carotid artery, horizontal segment – “I”; Aneurysm – “E” to “I”, red arrow (adapted and published with permission of Sergio Tadeu Fernandes, MD PhD).

Figure 7.

Removed markers from the previous figure to exercise the identification of structures of interest. Paraclinoid aneurysm on the right is considered transitional – note on “I”, blue arrow, how the most posterior portion of the aneurysm projects into the subarachnoid space – aneurysm – “E” to “I”, red arrow (adapted and published with permission of Sergio Tadeu Fernandes, MD PhD).

After analysis by the neuroradiologist, the following results were obtained: 11 (44.4%) were classified as intracavernous, 1 (4%) as transitional and 13 (52%) as intradural. Finally, the patients underwent microsurgical treatment of AI clipping. Of the 25 aneurysms analyzed during the microsurgical procedure and exploration of the paraclinoid region, 10 (40%) were classified as intracavernous, 2 (8%) as transitional and 13 (52%) as intradural. The comparative analysis of these data can be seen in Table 2.

Classification		Surgery		Total
		Cavernous	Not Cavernous
MRI	Cavernous	9 (36,0)	2 (8,0)	11 (44,4)
	Not Cavernous	1 (4,0)	13 (52,0)	14 (56,0)
	Total	10 (40)	15 (60)	25 (100)

Table 2.

Agreement analysis between MRI and surgery.

Data presented as n (% of total). Kappa: 0.754, p < 0.001. MRI: magnetic resonance imaging

Data processing showed that the accuracy of magnetic resonance imaging in terms of the intracavernous or intradural location of the aneurysm, with the intraoperative finding as the gold standard and the characteristic “presence of disease” or “positive test” for non-cavernous aneurysm, found sensitivity of 86.7% (95% CI, 59.5–98.3), specificity of 90.0 (95% CI, 55.5–99.8), positive and negative likelihood ratios of 8.7 (CI 95%, 1.3–56.2) and 0.15 (95% CI, 0.04–0.6), respectively, and positive and negative predictive values of 92.9 (95% CI, 66.1–99.8) and 81.8 (95% CI, 48.2–97.7), respectively. The inter-observer agreement by Cohen’s Kappa method was almost perfect (κ = 0.901; p < 0.001; 95% CI, 0.71–1.00) between MRI and surgical procedure findings. The diagnostic test in individuals with no history of SAH had a sensitivity of 92.3% and specificity of 100%. In this circumstance, the 100% specificity demonstrates the superiority of MRI when the aneurysm is intracavernous, that is, it is a method free of false negatives and can be considered the gold standard in ruling out the presence of disease (transitional or intradural aneurysm).

Transferring the issue to daily practice, it can be stated that, when considering the preoperative MRI result in the decision-making process for conservative treatment of paraclinoid aneurysms with no history of SAH, it is likely that all aneurysms considered cavernous, fact they are. The practical significance of these findings is that absolutely all patients eligible for preventive treatment of SAH (transitional or intradural aneurysms) will be so diagnosed and, eventually, only 1 in 10 of treated cases would not need treatment (cavernous aneurysms – false negatives).

3. Conclusion

MRI acquisition protocols demonstrated here are able to accurately define aneurysms in the paraclinoid region in terms of their location in relation to the cavernous sinus and, therefore, provides more appropriate and reliable recommendations for the management of paraclinoid aneurysms smaller than 10 mm regarding conservative or interventional (coiling or clipping) treatment options.

Acknowledgments

The authors would like to thank all collaborators of HTEJZ - SPDM, on behalf of its CEO, Otávio Monteiro Becker Jr., MD, PhD.

Conflict of interest

The authors declare no conflict of interest.