|Home | About | Journals | Submit | Contact Us | Français|
The anatomical features of the temporal bone can vary significantly among different individuals. These variations affect the operative view in middle cranial fossa surgery. We performed 18 middle fossa approaches in 9 cadaveric heads, with detailed morphological analysis, to identify unfavorable situations and reliable systems to avoid complications during surgery. We recorded linear, angular measurements and calculated areas. We performed a computed tomography (CT) scan with analysis of the amount of bone to remove in two temporal bones. We found that the location of the internal auditory canal (IAC) is the keystone of bone removal. We also found accuracy in the system suggested by E. and J. L. Garcia-Ibanez for its identification and that there is a smaller surgical window in female patients (statistically significant) that can be predicted on preoperative imaging studies. Our study also confirms significant individual variability in the mutual relationships of different surgical landmarks. We concluded that surgery of the middle fossa requires detailed understanding of the complex temporal bone anatomy. The surgeon has to be aware of extreme variability of the more commonly used anatomical landmarks. The method to identify the position of the IAC described by E. and J. L. Garcia-Ibanez seems to be the simplest and most reliable. When the surgical strategy includes an anterior petrosectomy, interindividual variability can critically affect the working area, particularly in females. The working area can be estimated on preoperative CT scans through the petrous bone.
The middle cranial fossa approach was originally described by Parry in 1904.1 In the 1960s, with the introduction of the operative microscope, this approach was popularized by House.2 Since then, interest in this procedure has increased rapidly because it allows effective treatment of different pathological conditions in this region while minimizing trauma to the facial nerve and hearing functions. With further advances in skull base surgery, extension of the middle fossa approach has become the key to accessing regions notoriously challenging such as the cerebellopontine angle,3,4 the petroclival area,5,6,7,8 the upper basilar artery,9,10,11 and the posterior cavernous sinus.12
Despite these advantages, the safe and effective performance of the middle fossa approach and its variations requires a detailed knowledge of the temporal bone anatomy, awareness of topographic variations, and a mandatory practice in cadaver laboratory. Therefore, this approach continues to be considered one of the most difficult surgical exercises in skull base surgery. The present study was undertaken with the goal of investigating the mutual relationships between anatomical landmarks and their variations.
The middle fossa approach was performed with the aid of the surgical microscope (Leica Microsystems GmbH, Wetzlar, Germany) in nine (3 female, 6 male) (18 sides) injected cadaveric specimens. The distance between different structures of interest exposed during this approach was measured with an electronic digital caliper under microscope view. Areas were calculated applying Carnot's theorem for triangle's resolution (a2=b2+c2−, 2bc cos α). Polygonal areas were divided into triangles and their areas were calculated with Carnot's theorem. Angle measurements were also recorded. Analysis of variance (ANOVA) was used to verify statistical significance.
High-resolution computed tomography (CT) scans (axial sections with thickness of 1 mm) of each head were obtained before dissection using a Siemens 4-slice CT scanner (Siemens, Pennsylvania). On the CT scan , the area of bone between the internal carotid artery (ICA), the third trigeminal branch (V3), the cochlea, and the internal auditory canal (IAC) (area which can be potentially removed through the middle fossa approach) was visualized. Linear measurements between these structures were taken on the CT. Using Carnot's theorem, the area of the fossa was calculated and later compared with the dissection of the same specimen to assess the predictive value of preoperative high-resolution CT scan (Fig. 1).
In each specimen the same microneurosurgical approach was performed according to the following surgical steps:
While elevating the dura, exploration of the floor of the middle fossa detects relevant anatomical variations. In 3 of the 18 dissections (17%), there was no bony covering of the geniculate ganglion; in 7 temporal bone specimens, the roof of the carotid canal was incomplete, exposing the ICA. The length of bone covering the GSPN showed great variability (mean 3.84 mm; standard deviation [SD] 2.33; range 0–7.61 mm).
The next step in the middle fossa approach is the identification of anatomical landmarks that aid the surgeon's orientation and help identify portions of bone to be removed (Fig. 2). Table Table11 summarizes the distances between different landmarks in the floor of the middle fossa measured before bone removal. For each measurement, there was a wide range with a high SD. This indicates that the anatomical conformation of the middle cranial fossa varies between subjects.
Localization of the IAC is the keystone to start bony drilling. Table Table22 summarizes angular measurements useful to clarifying the position of the IAC in relation to other middle fossa landmarks. Although the range of single measurements is wide, the ratio between the GSPN-IAC angle and the GSPN-AE angle is quite constant (mean 0.42, SD 0.08).
Table Table33 reports the distance between different structures of the middle fossa after drilling was completed. In five approaches (27.7%), there was no anatomical coincidence of the superior semicircular canal with the AE. Although the range of measurements between different structures varies by only a few millimeters, these small differences can result in significant differences in the size of removable bone in such a limited space.
Table Table44 indicates the maximal areas of bone that can be removed while preserving anatomical structures. The areas of bone analyzed were the anterior area, bounded by the gasserian ganglion, V3, ICA, cochlea, and IAC (corresponding to the area described by Kawase10); and the posterior area, bounded by the geniculate ganglion, IAC, and superior semicircular canal (corresponding to Fukushima's postmeatal triangle)12 (Fig. 3). The measurements of these two areas were then added to obtain the total bone removal area. We found a wide variability in the size of the bone windows; some were so small as to limit the efficacy of this approach.
Table Table55 compares the size of the bone windows in the middle fossa approach between males and females. We found significant differences in the anterior area (α=0.025) and the total area (α=0.01) despite the limited number of specimens (Fig. 4). A significant difference was not found for the posterior area.
In one specimen (bilateral temporal bones), the measurements of the anterior area on high-resolution CT scan were compared with the measurements taken in the same specimen through the surgical microscope. We found a very good correlation between the CT scan and the actual dissection (left side: CT=109 mm2, dissection=107.18 mm2; right side: CT=98 mm2, dissection=95.90 mm2).
Table Table66 summarizes the measurements of the petrous ICA exposed after different stages of dissection. The maximal length from the genu to the gasserian ganglion varied from 6.54 mm to 11.50 mm. After dividing connective attachments between the gasserian ganglion and the ICA, the ganglion could be mobilized to obtain further exposure of the artery (mean gain of 3.51 mm in ICA exposure). Decompressing the foramen ovale and rotundum and dividing V3 further increased the artery's exposure (mean gain of 2.55 mm and 2.35 mm, respectively).
The middle fossa approach and its variations offer a suitable surgical avenue to deal with a variety of pathological lesions located at the junction of the middle and posterior cranial fossae. A clear understanding of the anatomical landmarks and their variations is critical to effectively manage these lesions while avoiding disabling complications. Numerous authors have described frequent variations in the middle fossa anatomy. Absence of bone over the geniculate ganglion resulting from a congenital defect in the development of the roof of the temporal bone can occur in 15% of cases.13 The length of the GSPN covered by bone varies considerably,14,15 from 0 to 7.61 mm in our dissections. In ~20% of cases, the petrous ICA is exposed along the carotid canal by the absence of bone in the roof of the canal.14 Likewise, the topographic relationship between the AE and the superior semicircular canal is not consistent,16,17 and we found this to be true in 27% of our dissections. Lack of bone covering these structures may increase the risk of injury while elevating the dura off the floor of the middle fossa. Surgeons have to be familiar with these differences to avoid such complications.
In the 1960s, when the middle fossa approach was popularized by House, the main indications were small acoustic schwannomas confined to the IAC, Meniere's syndrome, and decompression of the acoustic and facial nerves.2 Thus, the initial goal of this approach was to expose the entire length of the IAC while preserving the facial nerve and the inner ear structures. In the extended middle fossa approaches, identification of the IAC still remains a crucial step in identifying portions of bone to be removed.
Different authors have described various methods to identify the position of the IAC. House suggested following the GSPN to the facial nerve, reaching it directly2; however, this may jeopardize the facial nerve by direct injury. Fisch reported that the IAC can be identified by finding the superior semicircular canal and visualizing the IAC as lying at an angle 60 degrees anteriorly.18 Unfortunately, our angular measurement showed high variability in the angle he used (mean, 77.68 degrees; SD, 13.30; range, 54.96 degrees to 102.97 degrees). The method described by E. Garcia-Ibanez and J. L. Garcia-Ibanez allows reaching the IAC without drilling over the superior semicircular canal or the geniculate ganglion, minimizing the risk of injury. They drill over the bisection of the angle formed by the GSPN and the AE (the main axis of the IAC).19 We found the ratio between the GSPN-IAC angle and the GSPN-AE angle to be quite constant (SD=0.08) and very close to 0.50 in all specimens. Thus, we found that the drilling zone indeed corresponds to the bisection of the GSPN-AE angle, making this the most reliable method in locating the IAC by anatomical landmarks. Additionally, risky maneuvers, such as drilling the superior semicircular canals and unroofing the geniculate ganglion, are avoided.
To facilitate surgical orientation, many authors have proposed subdivisions of the middle fossa in topographic areas. In 1969, Glasscock described the posterolateral triangle (bounded by the foramen spinosum, the cochlea, and the GSPN/V3 junction), allowing exposure of the horizontal petrous ICA.20 Kawase described removing a portion of petrous bone to gain access to the petroclival region and the basilar artery (bounded by the trigeminal ganglion, the IAC, the cochlea, the GSPN, and the petrous ICA).5,11 In 1994, Day et al approximated the floor of the middle fossa to a rhomboid complex of anatomical landmarks delimiting bone removal on the petrosal surface.21 These landmarks were the AE, the porus trigeminus, the GSPN/V3 junction, and the junction of imaginary lines projecting through the GSPN and the AE. They also described the postmeatal triangle (bounded by the geniculate ganglion, the IAC, and the superior semicircular canal) as an additional volume of bone that can be removed. They identified the permeatal triangle (bounded by the IAC, the carotid genu, and the geniculate ganglion) as useful for safely localizing the cochlea.21
Definition of these topographical areas relies on recognition of anatomical landmarks. Unfortunately, numerous anatomical studies have reported extreme individual variability in temporal bone anatomy.14,15,21,22,23,24,25,26,27,28 Our data confirm this variability. Even small differences in linear measurements result in a large variability of the width of the working area exposed after bone drilling. An interesting observation in our specimens was the significant differences between males and females, especially in the so-called “anterior area.” Current high-resolution neuroimaging can help in assessing the width of the working area with good correlation between the area imaged and the actual dissection, as demonstrated by one of our specimens. Calculation of bone superficies on CT is an easy and fast procedure that can help the surgeon in preoperative planning. Additionally, more sophisticated three-dimensional images29,30,31 and image-guided intraoperative navigation32,33,34 provide valuable technical adjuncts to provide more detailed information and intraoperative real-time feedback. These adjuncts might be particularly helpful in female patients because of the more limited surgical corridor through the temporal bone.
The middle fossa approach and its variations require a detailed understanding of the complex temporal bone anatomy. The surgeon has to be aware of the extreme variability of some of the more commonly used anatomical landmarks. The method described by E. and J. L. Garcia-Ibanez19 seems to be the simplest and most reliable method to identify the position of the IAC, a critical step in bone removal during these approaches. When the surgical strategy includes an anterior petrosectomy to reach deeper intracranial areas, interindividual variability can critically affect the working areas. This is particularly true in female patients because of the naturally smaller space anterior to the IAC. The working areas can be estimated on preoperative thin-cut CT scans through the petrous bone and the skull base.
We acknowledge the kind help of Ms. JoAnna Gass and Desiree Lanzino, Ph.D. in editing the manuscript. This study was supported in part by an educational grant from Midas Rex.
This article nicely confirms the usefulness of the E. and J. L. Garcia-Ibanez approach to the internal auditory canal from the middle fossa. The authors also demonstrate again that the anatomy of the middle fossa is extremely challenging and variable. However, the middle fossa remains a true “utility approach,” allowing access to several vital structures without compromising normal function, and it is also associated with minimal morbidity. Unfortunately, with this approach, there is no shortcut to the extensive temporal bone work required to develop the detailed knowledge of this complex anatomy necessary to achieve successful outcomes.