Otolaryngologists’ radiological assessment of nasal septum deviation symptomatology

Abstract

Objectives

Nasal Septal Deviation (NSD) is one of the most common causes of nasal obstruction. This study aims to further examine the clinical utility of imaging assessment in the workup and management of symptomatic nasal septal deviation, across all levels of medical training.

Study design

Cross-sectional survey.

Methods

CT scans of 10 confirmed NSD patients and 36 healthy controls (HC) were mixed and emailed through anonymous REDCap surveys to otolaryngologists in the US. The HC had no reported sinonasal obstruction symptoms—NOSE (NSD: 62.2 ± 12.5; HC: 5.69 ± 5.99, p < 0.05); SNOT-22 (NSD: 31.4 ± 14.5; HC: 9.72 ± 10.76, p < 0.05). The images consisted of a coronal slice at each subject’s most deviated location. Participants were instructed to choose the patients suspected to present with symptoms of sinonasal obstruction.

Results

88 otolaryngologists responded to the survey. 18 were excluded due to incomplete responses. On average, they identified 64.2 ± 29.8% of symptomatic NSD subjects correctly, but misidentified 54.6 ± 34.6% of HC as symptomatic. Their decisions were strongly correlated to degree of NSD (r = 0.69, p < 0.05). There exists a significant degree of NSD among HC (38.7 ± 17.2%), which does not significantly differ from symptomatic subjects (51.0 + 18.7%, p = 0.09). Residents and fellows performed similarly, with responses correlated between levels of training (r = 0.84–0.96, p < 0.05).

Conclusions

The incorporation of a substantial number of otolaryngologists, large patient sample, and blind mixing with HC gives us greater insight to the relative contribution of the extent of septal deviation to symptoms of nasal obstruction. Although NSD is a common factor contributing to nasal obstruction, the results of this study suggest that it is difficult to reliably infer obstructive symptoms based on degree of NSD on CT.

Level of evidence

Three.

Introduction

In the US, a septoplasty is one of the most common surgical procedures, with roughly 260,000 being performed annually [1]. The primary objective for most patients undergoing a septoplasty is to reduce symptoms of nasal obstruction. In contrast to a rhinoplasty, which alters the external and/or internal structures of the nose for cosmetic purposes or symptomatic relief, a septoplasty focuses solely on the internal structures to alleviate symptoms.

When the internal nasal structures present with a bony pyramid abnormality and/or septal deformity, they are globally defined as nasal septum deviation (NSD) and are often associated with symptoms of nasal obstruction [2]. These deviations can be classified into three groups: convex/concave deformities, concave dorsal deformities, and caudal septal deviations [3]. Patients with NSD commonly present with a history of nasal obstruction, although concomitant rhinitis or rhinosinusitis may contribute to additional symptoms, such as post-nasal drip, rhinorrhea, hyposmia, and facial pressure. In some cases, NSD may also lead to rhinogenic headache or dryness of the septal mucosa. NSD is quite common, with estimates in literature suggesting that some degree of septal deviation is present in up to 80% of the general population [1]. Despite the prevalence, not all NSDs contribute to nasal obstruction. In addition, turbinate hypertrophy, nasal valve collapse and inflammation are all common contributors to nasal obstruction that may be independent from the contribution of NSD to nasal obstruction.

Clinically, the ability to understand the driving factors that result in NSD symptomatology are essential in both diagnosing and planning for corrective surgery. However, there exists a void in identifying the cause behind symptomatic NSD in patients, as well as a gap in identifying specific anatomic anomalies that are more likely to present with symptoms.

Understanding the etiology of nasal obstruction is essential in developing both clinical and surgical plans. The inherent problem is that no single measure allows clinicians to definitively conclude that a patient will benefit from a septoplasty. Objective measures available for clinician use include acoustic rhinometry and rhinomanometry. Acoustic rhinometry provides a geometric measurement of the nasal cavity which serves to investigate the physiological and pathological changes of the nose [4]. Rhinomanometry measures the pressure and flow during inspiration and expiration to diagnose and objectively evaluate the respiratory resistance of the nose [5]. Patient reported scores of symptoms such as Sino-Nasal Outcome Test (SNOT-22) [6] and Nasal Obstruction Symptom Evaluation (NOSE) scores can serve as a simple screening instruments [7], with the NOSE score in particular being a strong indicator of the extent of nasal obstructive symptoms [8]. Another measure that can be used by the clinicians is the degree of NSD deviation present on a patient’s CT [9]. These scans are analyzed subjectively by physicians, but there is no indication as to whether or not the degree of NSD correlates with the symptoms that the patient experiences.

The goal of this paper is to gain a better insight into the extent to which degree of septal deviation contributes to nasal obstructive symptoms. Furthermore, we hope to understand the sort of qualitative characteristics which trained otolaryngologists look for when determining whether or not septoplasty would be beneficial.

Materials and methods

Research design

The study was based on an anonymous REDCap survey sent via email to the American Rhinologic Society membership list. The email recipients were able to access the survey via link or QR code. The study and survey were explained in detail to all prospective participants through a disclosure page on the link. Any additional questions were answered by members listed on the IRB. Enrollment criteria was based on status as an otolaryngologist. There was no identifiable information collected in the REDCap survey; therefore, it was completely anonymous. As such, the study protocol is approved by the Ohio State University Institutional Review Board as an exempt study.

The survey design is displayed in Fig. 1. The participants were asked to review CT scans of 10 patients with symptomatic septal deviation, randomly interspersed with 36 healthy control (HC) subjects (Fig. 2). The symptomatic individuals presented with complaints of nasal obstruction, and had no other definitive causes of obstruction other than a deviated septum and varying degrees of compensatory turbinate hypertrophy. The images consisted of one coronal slice at each subject’s most deviated location. In patients with septal spurs, the spur was included as the patient’s most deviated location. An example of a healthy control and a symptomatic patient are presented in Fig. 3, and all CT images used in this study are included in the supplementary files. The participants were instructed to choose whether they thought each subject would present with symptoms of nasal obstruction, or not.

Fig. 1.

Survey design

Fig. 2.

Measurement of nasal septal deviation. Percentage deviation was calculated by taking the distance along the x axis between the central plane (dotted line) and point of greatest deviation (p2–p1) and dividing it by the total width of that side of the nasal cavity (p3–p1), resulting in a percentage. This approach reflects the percentage of airway affected by deviation out of the potential available airway width, which should effectively control for variations in nasal cavity sizes and location of septal deviation

Fig. 3.

CT scan of healthy control (A) vs. symptomatic nasal septal deviation (B)

The healthy control subjects included in the survey were recruited from a local, healthy population in Columbus, Ohio. Symptomatic NSD patients were recruited from the patient population visiting the rhinology clinics at the Ohio State University Wexner Medical Center. All subjects underwent screening for confounding pathologies, including a CT scan, at the time of assessment. All subjects completed a NOSE and an SNOT-22. Healthy control subjects who scored above 25 (out of 100) on the NOSE scale were excluded from this study [10]. NSD patients with other confounding factors at the time of assessment (collapsible nasal valve, prior nasal surgery, etc.) were also excluded. Each patient also underwent rhinometry measurements, where the minimum cross section and nasal resistance were obtained. The unilaterally minimum (narrowest) cross-sectional area (MCA) in the anterior 5 cm of nasal airway was collected for each subject using acoustic rhinometry (SRE21000, RhinoMetrics A/S, Denmark). Nasal resistance during normal breathing was measured unilaterally by anterior rhinomanometry (SRE21000, RhinoMetrics A/S, Denmark) at reference pressure drop of 75 Pascals.

The results of the questionnaires were recorded, transferred into excel, and analyzed for differences in assessment among otolaryngologists. The primary confounding variables are as follows: (1) Level of Education, (2) Years of Practice, (3) Region of the Country, (4) Type of Practice, and (5) Specialty.

Results

88 otolaryngologists of various levels of training from across the country responded to the survey. 18 were excluded due to incomplete responses, leaving 13 residents, 6 fellows, and 51 attending physicians in total. Overall, the 70 otolaryngologists accurately identified 64.2% of symptomatic NSD subjects but misidentified 54.5% of the HC as symptomatic. Their decisions were strongly correlated to the degree of NSD measured on their CT (r = 0.69, p < 0.05). The degree of NSD was measured based on the coronal CT image slice at which the deviation was greatest (Fig. 4). Although there were moderate differences in degree of NSD among HC (38.7 ± 17.2%) vs. symptomatic NSD (51.0 ± 18.7%, p = 0.08), the differences were not statistically significant.

Fig. 4.

A Percentage of participants who determined each subject (dot) to be symptomatic when graded by attending physicians vs. resident physicians. B Percentage of participants who determined each subject (dot) as symptomatic, grouped by residents (triangle), fellows (square), and attending physicians (diamond), plotted against degree of measured Nasal Septal Deviation (NSD) of each subject. C Average percentage of participants who determined each subject (dot) to by symptomatic, plotted against degree of measured NSD of each subject. D Percentage of participants who determined each subject (dot) to be symptomatic based on years of practice experience (only attending physicians)

There were no significant differences in performance based on level of training. Residents, fellows, and attending physicians performed similarly, with responses correlated between levels of training (r = 0.85–0.94, p < 0.05) (Fig. 4a). In addition, while all levels of training appeared to take the percent NSD into account when predicting which subjects were symptomatic, there were likely other factors in the decision-making process since even the most deviated septa were not unanimously considered to be “symptomatic”. All training levels showed significant correlation between percent NSD and percent marked symptomatic, but at different levels (r = 0.55–0.69, p < 0.05) (Fig. 4b).

Among attending physicians, there were no significant differences between years of practice and accuracy in identifying symptomatic NSD via CT scan (Fig. 4d). While the accuracy of determining symptom burden from CT was relatively steady over time, there was a slight upward trend suggesting that physicians who have been in practice longer may be more likely to consider a patient with NSD to be symptomatic, but further studies are needed to confirm this. Other demographic variables among participating otolaryngologists, including region, type of practice, and subspecialty, did not yield significant findings.

Furthermore, HC reported fewer sinonasal obstruction symptoms—NOSE (NSD: 62.2 ± 12.5; HC: 5.69 ± 5.99, p < 0.05); SNOT-22 (NSD: 31.4 ± 14.5; HC: 9.72 ± 10.76, p < 0.05). Neither NOSE score, nor SNOT-22 score correlated strongly with percent NSD (Fig. 5a, b). Symptomatic patients were similar to HC when looking at percent NSD, despite having slightly greater deviation on average than those who were asymptomatic (Fig. 4c). Minimum nasal cross-sectional area and nasal resistance, as determined by acoustic rhinometry and rhinomanometry, were not significantly different between symptomatic NSD and HC, looking at both the deviated and non-deviated sides of the nose (Fig. 6b, c). These observations fall in line with the previous findings in the literature suggesting that objective measurements alone are insufficient in estimating the symptom burden of NSD.

Fig. 5.

A NOSE score vs. percent Nasal Septal Deviation (NSD) among healthy controls and symptomatic patients. B SNOT-22 score vs. percent NSD among healthy controls and symptomatic patients

Fig. 6.

A Average NOSE and SNOT-22 scores among patients with symptomatic NSD vs. healthy controls. Both scales were significantly greater in symptomatic individuals than in asymptomatic individuals. B Average nasal cross-sectional area among patients with symptomatic NSD vs. healthy controls, divided into deviated vs. non-deviated nostril. C Average nasal resistance among patients with symptomatic NSD vs. healthy controls, divided into deviated vs. non-deviated nostril

Discussion

Nasal airway obstruction is one of the most common complaints in primary care offices, causing significant impact on the quality of life in patients regardless of age [11]. Patients with symptoms of nasal obstruction will often seek out medical evaluation and be diagnosed with NSD. Despite a multitude of papers characterizing septal deviaton [12–14], there is still a lack of correlation that a higher degree of NSD will lead to a higher degree of nasal obstruction.

The current clinical workflow involves, first, consulting with patients to identify their complaints and other symptoms, followed by the clinical examination, and potentially, radiologic examination to rule out other causes of obstruction. It is well known that without clinical examination, radiologic evaluation alone cannot yield a diagnosis. Our study implicates that even within this conservative workup, CT findings can still be misleading in directing the management of septal deviation. We showed that while symptomatic patients as a whole had slightly more deviated septa than HC, the difference was not significant. In fact, several HC with no complaints of nasal obstruction had greater degrees of septal deviation than some individuals with symptoms.

The current logic of starting with symptom assessment and then progressing to clinical assessment is that; if a patient is truly symptomatic, has a deviated septum, and no other obvious confounding factors, then they should be treated for septal deviation. The problem with this logic is: how do we know that treating the septal deviation will lead to symptom improvement when so frequently, deviated septa are completely asymptomatic in the general population? Perhaps there are additional factors beyond the physical deviation that may contribute to the patients’ symptoms, e.g., the sensing of airflow [15], the effect of nasal cooling [16, 17] or certain aspects of NSD resulting in symptoms, rather than the overall degree of NSD. NOSE and SNOT-22 scores were effective at assessing symptomatic septal deviation, which was expected [18]. However, several patients with low degrees of septal deviation scored in mild to moderate obstruction ranges on the scales, which is counterintuitive. Future studies will be necessary to determine which other characteristics of septal deviation make a patient more likely to suffer from symptoms, since the degree of deviation is clearly not the only factor.

Several otolaryngologists who responded to the survey commented that, during their medical training they did not have the opportunity to view diverse CT images from general, healthy populations. This may represent a bias during medical training: that a healthy patient always presents with “perfect anatomy”, e.g., with a completely straight septum, even though significant anatomical variations exist in the normal asymptomatic population. Without adequate exposure to such normal variations, when an otolaryngologist views an image with anatomical deviations, such as NSD, there is an unconscious bias to assume it is causing symptoms, which is further reinforced by lack of negative controls, since a healthy person without symptoms wouldn’t have gotten imaging in the first place. This limitation could potentially apply to endoscopic examination as well if physicians were blinded of the subjects’ symptom status. This use of heuristics is reasonable and wouldn’t lead to problems in most contexts. However, in the setting of a condition that has a significant number of asymptomatic variations, it complicates decision-making on treatment options.

In addition, the population selected for this study differs from what the typical rhinologist might see on a normal day, since many patients visiting clinics have complaints of nasal obstruction, often with contribution from NSD. Yet, only 10 of the 46 subjects in this study were actually symptomatic. This may help to explain why the study participants predicted so many of the HC as being symptomatic.

The management of patients with nasal obstruction focuses on relieving the symptoms, with surgical options including septoplasty being the last resort. Considering the results of our study, it is important to acknowledge that there may be additional yet-to-be discovered factors that may determine if a patient’s nasal obstruction will be alleviated by septoplasty. Further studies examining certain patterns of NSD (i.e., caudal deviation, septal spurs, C-shaped vs. S-shaped deviation [19]) and symptom status after surgery, could help physicians identify which patients are more likely to benefit from a septoplasty, and which would not.

Study Limitations

It is important to note that none of the surgeons had access to the full CT stack of each subject when they made their decision, due to our survey format.

Conclusions

The novel incorporation of a substantial number of otolaryngologists, large patient sample, and blind mixing with HC gives us greater insight to the relative value of degree of septal deviation in clinical decision making for the treatment of NSD. Our results suggest that the degree of NSD cannot be used to reliably infer the symptoms and plan of action in patients with nasal obstruction. Therefore, we cannot definitively determine the extent to which septoplasty will benefit a given patient’s nasal obstructive symptoms. Further studies examining patterns of septal deviations and resultant nasal obstruction would build upon these findings in identifying which factors of NSD are most likely to cause obstructive symptoms.

Abbreviations

NSD

Nasal septal deviation

CT

Computed tomography

NOSE

Nasal obstruction symptom evaluation

SNOT-22

Sino-nasal outcome test