Introduction

Nasal septal deviation is usually attributed to trauma but can be present in patients without a documented history of injury. The nasal septum divides the internal nose in the sagittal plane and extends from the maxillary crest (MC) at the floor to the cribriform plate (CP) and skull base at the roof. It comprises of a cartilaginous section caudally (quadrangular cartilage) and a bony section made from the MC, perpendicular plate of the ethmoid (PPE) and vomer [20].

Nasal septal deviation could be secondary to trauma or septal growth and development. The septum begins its development from the neural crest and starts during the third month of gestation [21]. Endochondral ossification forms the first bony segment (PPE) at six months gestation. Prior to this, the septum is completely cartilaginous. Studies have shown that the cartilaginous septum develops to roughly its maximum size in the first two years after birth with only limited cell proliferation up to the age of 35 in the anterior central septum [25,26,27]. This corresponds to rapid growth of the septum in the first two years of life and subsequent minimal increases in size by early adulthood [3]. The exact point and regulatory mechanisms that determine stabilisation in septal growth rate is not well understood [3]. The endochondral ossification continues into the 4th decade of life (average 36 years of age). Therefore, while septal growth shows no significant change in total area between patients in their twenties compared to those over 70, there is a decrease in overall septal cartilage area as this ossification process continues [9, 12, 17, 25].

There have been multiple studies examining the influence of midface growth on the septum and vice-versa [2, 6, 8, 15, 22]. Other studies have assessed various internal nasal dimensions including nasal septal deviation with CT scans [4, 5, 7, 10, 17]. However, there are no known studies examining the deviation of the septum at its insertion to the roof and floor and its potential relationship with overall septal deviation. Given the fact that the septal development has such a close relationship with mid-face structures, it is likely to have a similar relationship with its insertion points into the facial skeleton.

This study therefore aimed to assess septal deviation at the CP and MC and to see if there was any correlation with overall septal deviation.

Materials and methods

Study design

All adult (> 18 years) patients who had a sinus CT scan, for any indication, between 1st January 2020 and 31st December 2021 at an inner-city hospital affiliated with a tertiary university hospital were identified by the coding department. Of the identified patients, scans and case notes were analysed retrospectively. Data collected included age, gender, ethnicity, indication for CT scan (categorised to sinus or non-sinus disease), any documented relevant injury (head, nasal or facial trauma) and any previous ENT surgery. Patients were excluded if their scans had landmarks that were not clearly identifiable to undertake the required measurements, or had a radiological or documented history of head, nasal or facial trauma, or any previous nasal surgical procedure. If patients had duplicate imaging over the study period, then only the earliest scan was included for analysis.

Computed tomography (CT) protocol

All CT scans were performed using either the Aquilion Prime 160 slice or Aquilion Prime SP 160 slice scanners (Canon Medical Systems, Otawara, Japan). A non-contrast navigation protocol was utilised with coverage from the vertex to the lower lip with the patient in a supine position. The scan parameters were 120kVP, 80 mA with a gantry rotation time of 0.5 s. Slice thickness was 0.5 mm with axial bone and soft tissue kernels sent to the picture archiving communications system (PACS). The 0.5 mm bone reconstructions were reviewed in the coronal plane to obtain the measurements. All included patients had their CT scans retrieved by the radiology department and reviewed on the hospital imaging platform (Agfa Impax, Agfa-Healthcare, Belgium).

Radiological measurements

In order to evaluate septal insertions to the floor and roof of the nose, as well as overall septal asymmetry, the following measurements were conducted:

  1. 1.

    Angle between septum and maxillary crest (SMC) (floor of nose) – left (LSMC) and right (RSMC).

  2. 2.

    Angle between septum and cribriform plate (SCP) (roof) – left (LSCP) and right (RSCP).

  3. 3.

    Maximal angle of septal deviation (MSD).

All measurements were undertaken independently by a head and neck radiology specialty trainee (D.D) and an ENT specialty trainee (K.K.G) using digital callipers and rulers available on the hospital radiology imaging viewer software. The measurements were taken in coronal CT planes with 0.5 mm slices and viewed in the bony tissue window. Figure 1 summarises how these measurements were taken. SMC and SCP angles were measured at the level of the first slice after the crista galli disappears and the cribriform plate forms, by drawing a vertical line down the septum and horizontal line across the maxillary crest and cribriform plate and then measuring the angles at which these lines intersect (Fig. 1a). MSD was measured by drawing a vertical line from the roof midline (at the first slice where the crista galli forms to allow for a more reliable and consistent midline measurement) to the floor midline at the maxillary crest, and another line from the roof midline point at the crista galli to the maximal point of septal deviation and measuring the angle created between these two lines (Fig. 1b). An angle of 90° was used to define no deviation at the cribriform plate (CP) or maxillary crest (MC).

Fig. 1
figure 1

Computed tomography (CT) scan examples to demonstrate how internal nasal measurements were taken using digital callipers and rulers. (A) Coronal CT scan showing measurement of the septum-cribriform plate angles and septum-maxillary crest angles on the left and right at the level when the crista galli disappears and the cribriform plate forms. Measurements are taken by drawing a vertical line down the septum and horizontal line across the maxillary crest and cribriform plate and then measuring the angles at which these lines intersect. (B) Coronal CT scan showing measurement of the maximal septal deviation (MSD) by drawing a vertical line from crista galli to the maxillary crest and another line from crista galli to the maximal point of septal deviation and measuring the angle between these lines

Full size image

Deviations at the CP and MC were assessed by calculating the difference between the measured angle on the scan and no deviation (i.e. 90°). This was then compared to the overall septal asymmetry (using the MSD) to assess for correlations between the MSD and deviations at CP and MC.

Statistical analysis

Descriptive statistical analysis was performed with mean and standard deviations calculated to summarise continuous variables after confirming normal distribution with Shapiro-Wilk test. Categorical variables were summarised using frequencies and percentages. Inter-rater reliability for measurements was tested using interclass correlation coefficient (ICC). Mean values of both raters were used in the final analysis if the ICC was greater than 0.75 (‘excellent’ inter-rater reliability as defined by Cichetti) [19]. Correlation coefficients were calculated to assess for any statistical correlation between measurements. All statistical analyses were performed using SPSS (version 27.0, Armonk, NY:IBM Corp). No patient identifiable information was retained or stored, and all scans were analysed anonymously. Adhering to the guiding principles of the Declaration of Helsinki this was an anonymous, retrospective analysis of case notes and investigations and as such no ethical approval was sought or required.

Results and analysis

A total of 100 scans were conducted during the study period (this low number compared to other years is likely a reflection of reduced outpatient work during the coronavirus pandemic). Seventy scans were included in the final analysis after exclusion of 13 cases with previous nasal surgery (all endoscopic sinus surgery), 4 cases with previous documented trauma and the 13 cases under 18 years of age.

The study population (n = 70) had the age range of 18–93 years and the majority were female (58.6%). Forty patients were European (57.1%) and 30 were South Asian (42.9%). The majority of scans were undertaken for sinus or polyp disease (n = 65, 92.9%). These findings are summarised in Table 1.

Table 1 Baseline demographics for included patients and scan indications (n = 70)
Full size table

Inter-rater reliability was deemed as ‘excellent’ with an ICC of 0.85. Therefore, the mean values for both raters were used in the final analysis. The mean results for MC and CP deviation angles to the right or left as well as the MSD measurements are presented in Table 2. The overall direction of septal asymmetry was toward the right in 36 patients (51.4%) and to the left in 34 patients (48.6%). The results are summarised in Table 2.

Table 2 Overall results from nasal measurements (n = 70). LSMC = left septum maxillary crest, RSMC = right septum maxillary crest, LSCP = left septum cribriform plate, RSCP = right septum cribriform plate, MSD = maximal septal deviation, SD = standard deviation
Full size table

When calculating the septal deviation at the roof (CP) and floor (MC), the mean deviation was 0.89° at the CP and 2.02° at the MC. When comparing this to the MSD, the correlation coefficient between the deviation at the CP and MSD was 0.025 and the correlation coefficient between the deviation at the MC and MSD was 0.321. These results are summarised in Figs. 2 and 3.

Fig. 2
figure 2

A scatter plot with trend line demonstrating the relationship between septal deviation at the cribriform plate (CP) and the maximal septal deviation (MSD) (correlation coefficient 0.025)

Full size image
Fig. 3
figure 3

A scatter plot with trend line demonstrating the relationship between septal deviation at the maxillary crest (MC) and the maximal septal deviation (MSD) (correlation coefficient 0.321)

Full size image

Discussion

Our results demonstrate that in a cohort of patients with no documented history of trauma or nasal surgery, there was some degree of septal deviation with an average of 8.14° when measured at its maximum. There was also an element of deviation at the septal insertion to the CP (0.89°), but the septal deviation was more prominent at its insertion to the MC (2.02°). The deviation at the CP had no correlation with the MSD. However, there was a correlation observed between MSD and the septal deviation at the MC. Our results suggest that greater deviations at the septal insertion to the floor of the nose are related to an increased overall septal deviation.

An explanation for the lack of correlation observed between the angle of deviation at the CP and MSD could be explained by tilted cribriform plates. This could also explain why there was an overall much smaller mean deviation at the CP (0.29°) compared to the MSD (8.14°). If the cribriform plate is tilted, even if there was a near perpendicular (90°) septal insertion into the roof, there would still be an overall septal asymmetry and apparent deviation due to the tilt at the point of septal insertion to the roof. This may have relevance in patients undergoing submucosal resection of the septum while preserving the dorsal bony and cartilaginous strut. While the septum may inherently be straight, it might still be leaning to one side (due to the angulated cribriform plate) resulting in residual septal deviation observed clinically. The patient may never truly have a straight septum in the midline due to variance in their cribriform plate and therefore still experience symptoms. Our findings therefore can have a clinical and practical relevance as if the insertion and anchoring points of the septum are asymmetrical, then no amount of surgery can result in a perfectly straight septum. It is important to be aware of this in terms of pre-operative assessment and patient counselling with respect to their expectation regarding the outcome of surgery.

Another explanation might be related to the septal development. Lindahl described developmental deviations as being usually smooth (‘C’ or ‘S’ shaped) with traumatic deviations being more irregular [14]. As our cohort had no documented history of trauma or surgery, the septal deviations observed are likely to be related to developmental issues. The development of the septum in isolation is not well researched, with most studies in the literature evaluating its development in relation to midfacial growth [3].

The nasal septum mostly grows in the first two years of life with endochondral ossification occurring until around 36 years of age [3]. It therefore follows that developmental nasal septal deviation would therefore have to occur in utero or during the early stages of life as the cartilaginous septum is developing. Excluding genetic causes and other craniofacial abnormalities, various explanations for this have been put forward in the literature [24]. These include intra-uterine microtrauma due to prolonged contact of the uterine wall and the fetal head during key stages of development, or various positions of the extremities of the fetus with the uterine wall from month four onward [23]. There have also been studies demonstrating correlation between nasal pyramid deformity, intrauterine stress and intramembranous ossification [11]. Furthermore, mammalian studies have that the nasofacial skeleton may act to constrain growth of the nasal septum and therefore induce deviation due to restriction; an effect that may be true in humans [6].

This does not however explain why the septal deviation observed in this study was more related to MC deviation and not CP deviation. A recent scoping review evaluated the literature pertaining to the nasal cartilage and its development, but it did not comment on such a relationship or find any studies examining it [3]. A potential theory for this relationship between overall septal deviation at the MC but not the CP is earlier ossification of the septum close to the CP. This would allow further growth of the lower part of the septum at its insertion at the floor of the nose which leaves it at risk to deviation for a longer time than the upper part of the septum where its insertion at the roof is ossified and fixed earlier.

A recent study also assessed high anterior septal deviation using CT scans [13]. This study found a similar angle of septal deviation (8.9°) to our study (8.14°) however their measurements were conducted in a more anterior plane. While CT scans have been used as a measurement tool in other studies, these have focussed on other aspects such as the nasal aperture [16, 19], septal surface area [4], relationship with the anterior nasal spine [7], measurements post septoplasty [18], and anatomical relations of the ethmoid roof [1], but have not examined the septum in terms of its deviation and insertion points.

Our study is not without limitations. The use of a retrospective cohort brings with it recall bias especially with respect to trauma and surgical histories. It is known that childhood trauma can have a high influence on overall septal growth. While we tried to minimise this by excluding patients with a documented trauma history, some patients may have not disclosed this and therefore have been included. In addition, while our sample represented all patients undergoing a CT sinus at our centre over a two-year period, the overall number included for analysis was fairly small. Despite this, our n value was comparable to similar studies conducted in this field thus far. In addition, the authors note that that there were some CT scans where the anatomical landmarks were not wholly consistent.

Conclusions

Overall, we provide a reliable method using CT scans for measuring the overall septal deviation as well as its deviation at the point of septal insertion at the roof and at the floor of the nose. Using this method, we have demonstrated an overall septal deviation of 8.14° in patients with no documented history of trauma or nasal surgery. Our data reveal that there is a positive correlation between septal deviation at the floor of the nose and overall septal deviation; this correlation was not observed between the overall septal deviation and deviation of the septum at its insertion to the roof. This could be explained due to the inherent tilt in the cribriform plate or by the developmental aspect of the septum such as earlier ossification and fixation of the septum at its insertion to the roof, thereby allowing further growth and potential for deviation of the lower part of the septum and its insertion to the floor of the nasal cavity.