Maturation of the internal auditory canal and posterior petrous bone with relevance to lateral and posterolateral skull base approaches

Anatomic knowledge of the internal auditory canal (IAC) and surrounding structures is a prerequisite for performing skull base approaches to the IAC. We herein perform a morphometric analysis of the IAC and surgically relevant aspects of the posterior petrous bone during pediatric maturation, a region well-studied in adults but not children. Measurements of IAC length (IAC-L), porus (IAC-D) and midpoint (IAC-DM) diameter, and distance from the porus to the common crus (CC; P-CC) and posterior petrosal surface (PPS) to the posterior semicircular canal (PSC; PPS-PSC) were made on thin-cut axial CT scans from 60 patients (grouped by ages 0–3, 4–7, 8–11 12–15, 16–18, and > 18 years). IAC-L increased 27.5% from 8.7 ± 1.1 at age 0–3 to 11.1 ± 1.1 mm at adulthood (p = 0.001), with the majority of growth occurring by ages 8–11. IAC-D (p = 0.52) and IAC-DM (p = 0.167) did not significantly change from ages 0–3 to adult. P-CC increased 31.1% from 7.7 ± 1.5 at age 0–3 to 10.1 ± 1.5 mm at adulthood (p = 0.019). PPS-PSC increased 160% from 1.5 ± 0.7 at age 0–3 to 3.9 ± 1.2 mm at adulthood (p < 0.001). The majority of growth in P-CC and PPS-PSC occurred by ages 12–15. Knowledge of these patterns may facilitate safe exposure of the IAC in children.

The dimensions and anatomic relationships of the internal auditory canal (IAC) to the components of the inner ear within the petrous temporal bone have important implications for the performance of lateral and posterolateral skull base approaches, in particular for the treatment of pathologies involving the IAC such as acoustic neuromas [1][2][3][4][5] . Familiarity with this anatomy is particularly important when hearing preservation is a surgical goal, such as with the retrosigmoid transmeatal and middle fossa approaches 2,3,5-8 .
In adults, numerous cadaveric and radiographic studies have defined the anatomic dimensions and variability of this region, including the size and shape of the IAC 2,[9][10][11][12][13][14][15] , as well as the distance of the IAC to surgically relevant inner ear structures such as the posterior semicircular canal (PSC), vestibule, and common crus (CC) 2,3,13 . Descriptions of these parameters in children are nonetheless limited 16,17 , despite other known differences in skull base anatomy between children and adults [18][19][20][21][22][23] , and an increasing application of skull base approaches in pediatric patients 18,19,[24][25][26][27][28][29][30][31] . We herein perform a granular analysis of the IAC and surgically relevant parameters of the posterior petrous bone during normal pediatric maturation. www.nature.com/scientificreports/ Methods Study population. Retrospective review identified neurosurgical patients at Rady Children's Hospital of San Diego and the University of California-San Diego during a 4-year period who underwent high resolution head computed-tomography (CT) imaging for focal cranial pathologies or trauma. Patients with traumatic fractures or other pathologies affecting the skull base, a history of hydrocephalus and/or a shunting procedure/ endoscopic third ventriculostomy, or those with a CT scan not aligned to the skull base (as defined below) were excluded. Data was collected on a total of 50 randomly selected pediatric patients and 10 adult patients, out of a potential patient pool of more than 500 meeting the inclusion/exclusion criteria. Patients were evenly distributed by sex and the following ages groups: 0-3, 4-7, 8-11, 12-15, 16-18 years, and > 18 years, with an n of 10 patients per age category. These groupings were chosen to approximate developmental milestones including infancy/ toddler (ages 0 to 3), early childhood (ages 4 to 7), middle childhood (ages 8 to 11), and adolescence (ages 12 to 18; split into pre-and postpuberty age groups).
The following variables were collected: age, sex, distance from the IAC porus acusticus to the fundus (IAC-L), IAC diameter at the porus (IAC-D), IAC diameter at the midpoint between the porus and fundus (IAC-DM), distance from the posterior aspect of the porus to the common crus (CC; P-CC), and the distance from the posterior petrosal surface (PPS) to the posterior semicircular canal (PSC; PPS-PSC).
Internal auditory canal parameters. IAC-L, IAC-D, and IAC-DM were measured on axial CT slices, using the slice with the largest visible cross-section of the IAC (Fig. 1). IAC-L was defined as the distance from the IAC porus (as marked by IAC-D) to the most lateral aspect of the fundus, along the center of the long axis of the IAC. IAC-D was defined as the distance from the most anterior to the most posterior aspect of the IAC porus. IAC-DM was defined as the anterior-posterior IAC diameter at the midpoint between the porus and fundus (defined as the midpoint of IAC-L).
IAC volume (IAC-V) was estimated based on the formula for a cone (volume = πr 2 × height/3), with height defined by IAC-L and base radius by IAC-D/2. This volumetric approximation was based on our observations of IAC shape during data collection and prior studies suggesting the most common IAC morphology in children is cone-(or funnel) shaped 32 .
Posterior petrous bone parameters. P-CC and PPS-PSC were measured on axial CT slices (Fig. 1).
P-CC was defined as the distance from the posterior aspect of the IAC porus to the most mesial aspect of the CC. PPS-PSC was defined as the distance from the PPS to the mid-PSC. [21][22][23] , all measurements for analysis were made by a single author (R.C.R) after technique optimization on 10 CT scans from 10 excluded patients (data not included). Measurements were repeated on 10 random patients across a variety of ages by a second author (M.G.B.) to assess data reproducibility. Although gantry angles varied by scan based on patient head positioning, included head CTs were based on institutional conventions wherein axial slices were oriented parallel to the base of the skull and the supraorbital meatal line.

Statistical analysis.
Data on continuous variables is reported as mean ± standard deviation (SD) or standard error of the mean (SEM). IAC and posterior petrosal bone parameters were analyzed by laterality, sex, and age using paired t-tests, unpaired t-tests, and linear regression. Age was considered a categorical variable, broken into six groups (0-3, 4-7, 8-11, 12-15, 16-18 years, and > 18 years). After testing for significant differences by laterality, left and right data for individual patients were averaged to reduce statistical comparisons. Corrections for multiple comparisons were made to reduce the likelihood of Type 1 error using the Student-Newman-Keuls (SNK) method. Scatterplots of standardized fit parameters by age group (obtained by subtracting the mean and dividing by the SD) were compared via regression line coefficient analysis by using Stata's "suest" postestimation command. To assess for differences in measurements by author (R.C.R. versus M.G.B.), paired two sample t-tests and Pearson correlation coefficients were calculated between datasets (after left and right data were averaged). All statistical analyses were performed using Stata MP Version 14.1 (StataCorp LP). Statistical significance was defined as a p-value < 0.05.

Consent. All methods were performed in accordance with relevant guidelines and regulations of the Rady
Children's Hospital-San Diego/University of California-San Diego Institutional Review Boards (IRBs) and Health Insurance Portability and Accountability Act. The Rady Children's Hospital-San Diego/University of California-San Diego Institutional Review Board (IRBs) approved this study, including waiver of individual consent for collection and analysis of the retrospective de-identified data in this work. Individual patient consent for procedures was obtained for all patients.
Supporting the fidelity of this data, no significant differences in any parameter were seen between the primary and test datasets as measured by different authors, with Pearson correlations for all parameters > 0.8.

Discussion
The known variability of the IAC and the petrous temporal bone has important surgical implications. For example, the orientation and position of the IAC within the petrous bone can significantly affect visualization of the facial nerve from a translabrynthine approach 1 . The length of the IAC and distance from the IAC porus to the common crus, as well as the depth of the posterior semicircular canal from the surface of the posterior petrous bone can similarly affect the maximum safe exposure of the IAC via a retrosigmoid transmeatal approach (Fig. 1A) 2 . The length and diameter of the IAC is also relevant to more recently described techniques, such as the minimally invasive expanded transcanal transpromontorial approach to the IAC through the external auditory canal 33,34 . Conversely, knowledge of structures that are relatively constant, such as the relationships between the cochlea, superior semicircular canal, and the facial nerve at the fundus of the IAC is critical to optimize exposure of the IAC while preserving hearing with a middle fossa approach 35 . An awareness of how the anatomy of this region changes with maturation is a pre-requisite for the performance of these skull base approaches in children.
This work builds on our prior studies detailing the surgical implications of maturation of the occipital condyle, jugular tubercle (JT), anterior clinoid process (ACP), and anterior petrous apex (APA) [21][22][23] , In line with our previous findings, maturation of the IAC and posterior petrous bone is nonlinear. Unlike the biphasic growth pattern of the JT 21 , or the early growth then plateau of the APA 23 , the development of the IAC most closely resembles the dynamic elongation and remodeling seen with the ACP 22 . The increase in IAC-L up until ages 8-11, combined with the stability of IAC-D and decrease in IAC-DM from ages 12 to 15 to adult, effectively remodels the IAC from a short, wide cone to a longer, narrower canal over time, while keeping its net volume relatively stable.  www.nature.com/scientificreports/ The maturation of the posterior petrous bone is more straightforward, with P-CC and PPS-PSC demonstrating concentrated growth from birth until early adolescence, reaching near-adult dimensions by ages 12-15, with little growth thereafter. Highlighting the variations in skull base development by region, the growth of the measured posterior petrosal dimensions is slightly prolonged as compared the APA region, which aside from the depth of the IAC from the middle fossa floor is largely completed by ages 8-11 23 . It is nonetheless completed earlier than the final maturation of the JT, which occurs by ages 16-18 21 , or the ACP which matures up until adulthood 22 . Supporting our methodology, the parameters measured in the adult control group in this study are in line with previous adult cadaveric and imaging-based works. Specifically, our adult IAC-L of 11.1 ± 1.1 mm fits within the 9.8 to 11.6 mm reported means (range 6-16.5 mm) of prior studies 2,[9][10][11][12][13][14][15]32 , and our adult IAC-D (7.7 ± 0.7 mm) and IAC-DM (4.0 ± 0.5 mm) measurements closely match prior reports (mean 6.2 to 7.5, range 2.6-12.4 mm for IAC-D and mean 4.2 to 6.5, range 2.3-8.1 for IAC-DM) 2,10,13,14,32,36 . Our adult posterior petrous measurements P-CC (10.1 ± 1.5 mm) and PPS-PSC (3.9 ± 1.2 mm) are also supported by prior dissection and imaging-based temporal bone analyses (mean 7.5 to 10.4 mm, range 5.3-13.5 mm for P-CC and mean 3.3 to 3.6 mm, range 0.6-6.5 mm for PPS-PSC, despite variations in measurement technique for the latter parameter) 2,3,13 . The wide ranges reported within these prior works nonetheless underscores the variability possible with this anatomy.
Although studies of the IAC and posterior petrous bone in pediatric populations are more limited, our data largely fit with prior observations and increase the granularity of our understanding of the development of these structures. In a study by Bonaldi et al., IAC molds from 98 dried temporal bones on subjects ranging from 7-month-old fetuses to adults were grouped (I: 7-9-month fetuses, II: 2-11 month old infants, III: 1-6 year old children, and IV: adults) and measured 16 . Despite a focus on the prenatal, infancy and early childhood periods (≤ 6 years of age) in this study, a similar elongation of IAC-L (measured along the anterior, posterior, superior, www.nature.com/scientificreports/  www.nature.com/scientificreports/ and inferior walls) and overall stability of IAC-DM was observed (with the narrowing of IAC-DM from ages 12 to 15 to adulthood in our study outside of the analyzed populations). The reported increase in IAC-D (notably in the axial but not coronal plane) in this work (from 4.02 in group I to 7.03 mm in group IV) was nonetheless concentrated within groups I-III, indicative of an early widening of the porous in the late prenatal and early postnatal period. This growth was already largely present in the 0-3 age group in our study (mean 7.4 ± 1.5 mm versus a mean of 6.09 ± 0.92 mm in group III in Bonaldi et al.), before leveling off throughout the rest of development. IAC-D growth that stopped by approximately 1 year of age and an overall stability of IAC-DM throughout development was also reported by Sakashita and Sando in a study of 20 temporal bones from subjects aged 1 month to 72 years 17 , with the narrowing of the IAC-DM during later development in our data potentially not captured by the small n of this study. Although it is unclear why the observed changes in IAC-DM lag behind IAC-L maturation, it is conceptually consistent with the conclusion reached by Sakashita and Sando that the IAC develops mainly along its long axis away from the early matured otic capsule (elongating and potentially narrowing the medial canal over time) 17 . Further supporting our findings, these authors report a similar, significant increase in IAC-L from birth to age 10, after which no further growth was noted through adulthood. Sakashita and Sando also report IAC-V using computer aided reconstructions, with their volumetric growth curve and maturation by age 10 closely matching our data. Although IAC-V growth during development reached significance in the Sakashita and Sando study (compared to the near-significant trends observed in our data), this difference is likely driven by concentrated volumetric growth within the first year of life that was not granularly captured herein. Our results differ somewhat from the findings of Marques et al., who studied IAC morphology using CT scans of 110 subjects across an age range of 1 to 92 years 32 . When comparing the IACs of children (up to 14 years) and adults (19 to 92 years) in this work, IAC-D was largely stable and IAC-DM decreased with maturation, mirroring our data. However, rather than elongation Marques et al. report a decrease in IAC-L during development, with a mean of 11.17 mm in children to 9.84 mm in adults. The mean and distribution of ages within these cohorts is nonetheless not reported in this work, and given the maturation to near adult parameters of IAC-L in our data by ages 8-11, if a relatively older cohort of children was analyzed the IAC-L would be expected to be closer to adult size (supported by the 11.17 mm mean IAC-L in children in Marques et al. resembling the 11.1 mm mean IAC-L in adults in our study, and the 9.84 mm mean IAC-L in adults in Marques et al. representing the low end of previous analyses) 2,[9][10][11][12][13][14][15]32 . To the best of our knowledge, prior granular data on the effects of maturation on posterior petrous bone parameters were not available for comparison.
These data, combined with a close pre-operative imaging review, may aid the selection and performance of surgical approaches to lesions near to or involving the IAC in pediatric patients. For example, in the older pediatric population where the anatomic dimensions of the IAC and surrounding structures are largely adult size www.nature.com/scientificreports/ by ages 12-15 23 , few modifications may be needed in the performance of skull base approaches to this region as compared to adults. Conversely, in younger patients, awareness of a shorter IAC-L, P-CC, and PPS-PSC is important to avoid an inner ear breach during exposure of the IAC via a middle fossa or retrosigmoid transmeatal approach. The relative proportion of the IAC available for safe exposure is also smaller in this population, as IAC growth occurs preferentially medially to the foramen singulare, a surgical landmark of the labyrinth during a retrosigmoid transmeatal exposure of the IAC 11,17 . The predominately open conical shape and largely preserved volume of the IAC in younger patients may nonetheless aid visualization of the proximal IAC despite these constraints. Accordingly, in our experience the safe performance of these approaches is possible, even in patients less than 2 years old 27,31 . A limitation of this work is the use of imaging-based rather than cadaveric measurements. Prior works have nonetheless demonstrated the comparability of these approaches for studying this region 2 . Obtaining cadaveric specimens of adequate numbers within each of the selected age groups was also impractical. As described in previous studies using a similar methodology by our group [21][22][23] , multiple steps including the use of strict patient inclusion/exclusion criteria and statistical analyses to minimize selection biases were taken to optimize the fidelity of this data. To limit interrater variability, all measurements used for analysis were performed by a single author after a training period for familiarization with variations in bony anatomy across patients. Supporting the accuracy of our data, repeat measurements on randomly selected patients across age groups by a second author demonstrated no significant differences and a high level of correlation between authors.

Conclusion
Maturation of the IAC and surrounding region is largely complete by ages 12-15. Awareness of the growth patterns of this area may facilitate the safe performance of skull base approaches to the IAC in pediatric patients, including the hearing-preserving middle fossa and retrosigmoid transmeatal approaches.