Objectives
The aim of this study was to determine (1) the diagnostic efficacy of orthopantomography (OPG) in the diagnosis of sinus diseases by using cone beam computed tomography (CBCT) as the imaging gold standard, (2) which diseases can be diagnosed by using panoramic radiography or CBCT, and (3) the interobserver agreement of 2 experienced dental radiologists.
Study Design
The images of 714 individuals who underwent OPG and CBCT on the same day were assessed separately by 2 dental radiologists. The results were compared by using Gwet's AC1 statistical methods.
Results
In total, 1322 maxillary sinuses were imaged. The sensitivity of OPG for the detection of any maxillary sinus pathology was poor compared with CBCT, but the specificity was high. The sensitivity of OPG for detecting mucosal thickening was 36.7%. The positive predictive value of OPG for diagnosing mucosal thickening was 79.9 %, but the negative predictive value was 51.9%. Interobserver agreement was strong (≥0.912) for all lesions except mucosal thickening.
Conclusions
Panoramic imaging has low efficacy in the diagnosis of sinus disease, even when examined by experienced dental radiologists. OPG can be useful in excluding disease, but 3-dimensional scanning is necessary for the definitive investigation of sinus lesions.
Panoramic radiography, also known as the orthopantomography (OPG), is commonly used in the investigation of facial and dental pain. This radiographic modality shows the bones of the jaws, teeth, and supporting structures including the temporomandibular joints and the maxillary sinuses, which can all be contributors to facial pain. In this tomographic technique to obtain these images, a thin “focal trough” along the dental arch is used to produce a 2-dimensional image of the jaws. A disadvantage of the technique is that structures outside the focal trough can be blurred or not visible at all. The maxillary sinuses generally extend between 28.9 mm and 47.6 mm in the anteroposterior diameter,
1
which is deeper than the focal trough that is designed to encompass the alveolar bone. The panoramic image includes only a portion of the maxillary sinuses, which might limit the value of this radiographic technique for the diagnosis of sinus lesions.Computed tomography (CT) and cone beam computed tomography (CBCT) are often considered the radiographic gold standard for imaging the sinuses.
2
, 3
Although the images are anatomically accurate, CT and CBCT have a number of disadvantages compared with OPG, including a higher radiation dose to the patient and a much higher financial cost that is not always covered by insurance. CBCT is widely available for use in dental medicine, but usually, there is limited access to CT, and this restricts referrals from general dental practitioners. For these reasons, panoramic imaging continues to be used as the first line of investigation of the maxillary sinuses in the majority of patients.There is paucity of published studies investigating the beneficial effects of OPG in the diagnosis of maxillary sinus disease. Most studies have been performed on a small number of patients, and the radiographs have not always been interpreted by a specialist in dental or oral and maxillofacial radiology. Only 4 studies have compared panoramic imaging and CT or CBCT with respect to sinus disease.
4
, 5
, 6
, 7
All these studies found CT or CBCT to be more accurate than panoramic imaging in the diagnosis of sinus disease, but there was marked variability in the sensitivity and specificity found for panoramic imaging, and none of these studies included more than 100 patients. Several articles have compared the 2 imaging modalities with specific reference to third molar root relationships and sinus septations, with similar results.7
, 8
, 9
, 10
The primary objective of this study was to determine the sensitivity, specificity, and positive predictive value (PPV) and negative predictive value (NPV) of panoramic imaging in the diagnosis of different sinus diseases with use of CBCT as the imaging gold standard. The secondary aim was to determine which sinus diseases can be detected on panoramic images and which lesions require cross-sectional imaging to make the diagnosis. The final objective was to examine the interobserver reliability between 2 experienced dental radiologists in the diagnosis of sinus pathology with the use of OPG.
Materials and Methods
The study was approved by the University of Adelaide Human Research Ethics Committee, which waived the need to obtain informed consent. This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
OPG and CBCT were performed on 714 adult patients (age 18 years or greater) for dental diagnostic purposes (Table I). Indications for imaging included orthodontic evaluation, investigation of oral pain, and implant planning. Both imaging techniques were performed on each patient on the same day, and thus both imaging techniques are representative of the same anatomy and pathology. All referrals were received from fully qualified specialist oral surgeons, orthodontists, periodontists, and endodontists. Both imaging techniques were performed at the request of the referring dentist, with the CBCT scan frequently limited to the maxilla or only a part of the maxilla.
Table IDemographic data of patients included in the study
Age | ||||
---|---|---|---|---|
Gender | N (%) | Min | Mean | Max |
Male | 312 (43.7%) | 18.8 | 51.7 | 89.7 |
Female | 402 (56.3%) | 18.9 | 54.0 | 90.3 |
All images were taken by using a Vatech Reve3-D combined panoramic/CBCT unit (Vatech, Gyeonggi, ROK) by qualified radiographers. Image parameters were selected on the basis of patient size and clinical information provided on the referral.
Data collection
A sample size calculation was undertaken on the basis of the findings of Tadinada,
4
Martinez-Gonzalez,5
and Nah,6
and the sensitivity of dental OPG for detecting sinus disease was found to be 22% to 66%. Sample size calculations were performed by assuming a prevalence of 30% (based on mucosal thickening), sensitivity of 50%, and a desired margin of error of 10% for a 95% confidence interval, which indicated that a minimum of 321 patients was required.11
Every OPG image was evaluated independently by 2 specialist dental/head and neck radiologists. The first (S.C.) had 10 years of specialist experience in dental radiology, and the other (B.C.) had 40 years of specialist experience in dental radiology. Both radiologists were experienced in the interpretation of panoramic images, as well as craniomaxillofacial CT and CBCT images. Each maxillary sinus was examined for the presence of mucosal thickening, mucosal polyps/mucous retention cysts, fluid in the sinus, odontogenic sinusitis, mucoceles, oroantral fistulas, or tumors. Each disease was recorded as “present” if the radiologist could confidently diagnose on the basis of the OPG image, or “absent” if a diagnosis could not confidently be recorded on the basis of the OPG alone. The CBCT scans were read by only one of the radiologists and served as the gold standard.
The following definitions were used:
- •Mucosal thickening: The normal maxillary sinus mucosa is between 0.2 and 0.8 mm in thickness.12Mucosa of greater than 1-mm thickness was recorded as “mucosal thickening” in the maxillary sinus. The mucosal thickening had to involve the majority of the floor of the sinus to be recorded as thickened on CBCT scans (Figure 1).Fig. 1A, Orthopantomography (OPG) image showing mucosal thickening in the left maxillary sinus (arrows). B, Sagittal cone beam computed tomography (CBCT) image in the same individual showing mucosal thickening in the left maxillary sinus (arrows).
- •Mucosal polyp/mucous retention cyst: The densities of these lesions were very similar, and it was not possible to accurately differentiate between them on imaging. A polyp/cyst was considered to be present if there was focal thickening of the sinus mucosa of greater than 5 mm (Figure 2).Fig. 2A, Orthopantomography (OPG) image showing a polyp or mucous retention cyst in the right maxillary sinus (arrows). B, Cone beam computed tomography (CBCT) image in the same individual showing a polyp or mucous retention cyst in the right maxillary sinus (arrows).
- •Sinus fluid: Fluid was deemed to be present if a horizontal shadow with a meniscal edge was present (Figure 3).Fig. 3A, Orthopantomography (OPG) image showing the meniscal edge associated with fluid in the maxillary sinus (arrows). B, Coronal cone beam computed tomography (CBCT) image in the same individual showing the meniscal edge associated with fluid in the maxillary sinus (arrows).
- •Odontogenic sinusitis: The presence of mucosal thickening or polyps/cysts that are in direct contact with periapical lesions were determinants of odontogenic sinusitis (Figure 4).Fig. 4A, Orthopantomography (OPG) image showing a periapical lesion on the left maxillary first molar with associated sinus disease representing odontogenic sinusitis (arrows). B, Coronal cone beam computed tomography (CBCT) image in the same individual showing a periapical lesion on the left maxillary first molar with associated sinus disease representing odontogenic sinusitis (arrows).
- •Mucocele: A mucocele was considered present if the sinus was opacified and expanded, with thinning of the bony walls (Figure 5).Fig. 5A, Orthopantomography (OPG) image showing the opacified and expanded left maxillary antrum caused by a mucocele (arrows). B, Axial cone beam computed tomography (CBCT) image in the same individual showing the opacified and expanded left maxillary antrum caused by a mucocele (arrows).
- •Oroantral fistula: The diagnosis of an oroantral fistula was made if a bony dehiscence was detected in the floor of the maxillary sinus and if it communicated with the oral cavity (Figure 6).Fig. 6A, Orthopantomography (OPG) image showing a small bony dehiscence in the floor of the maxillary sinus, indicating an oroantral fistula (arrows). B, Cone beam computed tomography (CBCT) image in the same individual showing a small bony dehiscence in the floor of the maxillary sinus, indicating an oroantral fistula (arrows).
- •Tumor: The presence of bone destruction is the hallmark of malignant sinus tumors. No tumors were found in this study.
The CBCT scans that were evaluated by the specialist dental radiologist (S.C.) were read by using Osirix DICOM (Digital Imaging and Communications in Medicine) viewer software enabling multiplanar reconstructions in the sagittal, coronal, axial, and oblique planes. Each sinus was evaluated by using the same criteria listed above. The OPG images were not available when the CBCT scans were evaluated, and the scans were read several weeks after the panoramic images to ensure that there was minimal crossover bias.
All data manipulations were undertaken in the open-source statistical software R v3.4.3.
13
For each sinus disease, the estimation of sensitivity and specificity of OPG (with CBCT as the gold standard) took into account the relationship between the left and right sides of each patient, using the variance inflation factor approach described by Genders et al.14
This was done because even though CBCT was considered the gold standard, it is unlikely to be 100% accurate. 15
The interobserver agreement was estimated for each sinus disease by using Gwet's AC1, which is preferable to Cohen's Kappa.16
The interpretation of AC1 is similar to Kappa—that is, AC1 varies between 0 and 1, where 0 denotes complete disagreement, and 1 denotes complete agreement, and intermediate values can be assessed by using the same scales that are used for Cohen's Kappa (≤0.2 = poor; 0.21–0.40 = fair; 0.41–0.60 = moderate; 0.61–0.80 = good; 0.81–1.0 = very good).17
This was done separately for each side because there is no readily available measure that can take clustering into account; the AC1 estimates for the 2 sides were found to be very similar, and hence their averages were reported.Results
Participant demographic characteristics
A summary of the demographic characteristics of the participants is given in Table I.
Sensitivity, specificity, and predictive values
Data were collected from 714 patients. Of these, 19 were excluded from the comparison of OPG with CBCT because neither of the maxillary sinuses had been fully imaged on CBCT. A total of 1322 maxillary sinuses were imaged (658 right sinuses, 664 left sinuses) because 68 patients had CBCT of only one side of the maxilla (37 right side only, 31 left side only).
The prevalence of maxillary sinus pathology detected on CBCT is shown in Table II.
Table IISinus pathology detected on cone beam computed tomography scanning
Number of affected sinuses | Percentage (%) | |
---|---|---|
Mucosal thickening | 744 | 56.3 |
Mucosal polyp/mucus retention cyst | 155 | 11.7 |
Sinus fluid | 15 | 1.1 |
Odontogenic sinusitis | 153 | 11.6 |
Mucocele | 2 | 0.2 |
Oroantral fistula | 4 | 0.3 |
More than half the sinuses showed mucosal disease, and disease was bilateral in 263 patients. The presence of 2 of more lesions was also common, and all involved mucosal thickening. Two diseases were observed in 208 sinuses (15.7%), and 3 lesions were observed in 22 sinuses (1.7%).
Data on the sensitivity, specificity, PPV, and NPV of panoramic imaging in the detection of sinus disease compared with CBCT are provided in Table III.
Estimates for positive and negative predictive values were calculated by using the prevalence estimates for each corresponding sinus disease (see Table II).
Table IIIEstimates of sensitivity and specificity (including 95% confidence interval [CI]), by using adjusting for clustered patient data (i.e., sides)
Sensitivity (%) | 95% CI | Specificity (%) | 95% CI | Positive predictive value (%) | Negative predictive value (%) | |
---|---|---|---|---|---|---|
Mucosal thickening | 36.7 | 32.6–40.8 | 88.1 | 85.1–91.0 | 79.9 | 51.9 |
Polyp/mucus retention cyst | 31.6 | 23.6–39.6 | 92.5 | 90.8–94.3 | 35.8 | 91.1 |
Sinus fluid | 6.7 | 0.0–19.4 | 99.9 | 99.8–100.0 | 42.7 | 99.0 |
Odontogenic sinusitis | 22.2 | 14.9–29.5 | 99.3 | 98.8–99.8 | 80.6 | 90.7 |
Mucocele | 0 | 100 | 99.8 | |||
Oro-antral fistula | 25.0 | 99.9 | 99.8 – 100.0 | 42.9 | 99.8 |
Blank cells indicate that the corresponding value(s) could not be estimated from the data.
† Could not be estimated because the intraclass correlation coefficient was zero.
‡ Could not be estimated because there was no detection using orthopantomography (OPG).
In all of our observed findings, the sensitivity of OPG for the detection of any maxillary sinus pathology was poor compared with CBCT scanning (≤36.7%), and specificity was high (≥88.1%). The PPV of OPG for correctly diagnosing mucosal thickening was 79.9%, but the NPV was only 51.9%.
Interobserver agreement
Panoramic images were reviewed for a total of 714 patients by both radiologists. The average AC1 values (based on separate calculations for left and right sides) are given in Table IV.
Table IVComparison of orthopantomography (OPG) interpretation by 2 dental radiologists
Agreement (AC1) 10 | |
---|---|
Mucosal thickening | 0.677 |
Polyp/mucus retention cyst | 0.912 |
Sinus fluid | 0.999 |
Odontogenic sinusitis | 0.971 |
Mucocele | 1.000 |
Oroantral fistula | 0.999 |
There was very high agreement (≥0.912) in all reported diseases except the presence of mucosal thickening (0.677). One radiologist (S.C.) reported mucosal thickening more often than did the other (B.C.).
Discussion
CT and CBCT have been the “gold standard” for imaging the paranasal sinuses for many years. Despite this, the higher cost and higher radiation dose, compared with plain radiography, have resulted in OPG continuing to be used for diagnosis when surgery is not immediately being considered. We found the sensitivity of diagnoses based on OPG for detecting mucosal thickening was only 36.7%, and the NPV of 51.9% was little better than flipping a coin to exclude disease. The sensitivity in detecting other sinus pathoses was even lower, although the high NPV in these conditions lends to the utility of panoramic imaging in excluding these abnormalities.
Hayfever and allergic rhinitis affect around 20% of Australians, with 8.4% reporting chronic sinusitis.
18
The incidence in South Australia is slightly higher than the national average, with 21.3% of residents reporting symptoms of hayfever/allergic rhinitis,18
a rate that is slowly increasing. Most findings of mucosal thickening in the absence of an upper respiratory tract infection can be attributed to the increasing rate of allergies in our population. The prevalence of 56.3% for mucosal thickening in our study is more than double the reported population average. This can be attributed to the percentage of individuals with odontogenic sinusitis (11.6%) who were also included in the mucosal thickening group and to the likelihood that many of the population with mild mucosal thickening or mucus retention cysts are asymptomatic and would, therefore, not report the presence of these conditions in the National Health Survey. It is well known that the correlation between sinus disease and sinus symptoms is poor.19
, 20
, 21
Tadinada et al.
4
found a high prevalence of sinus disease in their small cohort of patients, although they did not differentiate among the various diseases. When comparing OPG to CBCT, they also found an NPV that was little better than chance. Martinez-Gonzalez et al.5
reported similar results in their slightly larger cohort of patients, with OPG-detected sinus pathology in less than a third of patients with disease detected on CBCT. These findings correlate well with our results. Nah et al.6
had more success with panoramic imaging, with approximately two-thirds of OPG findings correlating with findings on CBCT, a rate that is higher than most other studies. Shahbazian et al.7
found that diagnostic results from panoramic imaging were very poor for odontogenic sinus disease compared with CBCT (7.6%). This finding is quite different from that in our study, which showed a much higher correlation. However, our patient numbers were substantially larger (714 vs 157).The present study found that OPG was more useful in the investigation of odontogenic sinusitis than that of most other sinus conditions. The PPV of 80.6% and NPV of 90.7% indicate that the dental professional can be reasonably sure of the presence (or absence) of sinus disease caused by dental pathology without proceeding to CT or CBCT scanning. The incidence of odontogenic sinusitis is reported as being between 10% and 40%,
22
, 23
although on imaging alone it can be difficult to determine the cause of sinus disease, and clinical assessment is essential. We found a prevalence of 11.6%, based on the features of periapical disease in direct contact with mucosal thickening in the maxillary sinus. This figure may change when also taking into account the clinical features. Odontogenic sinusitis is generally thought to be under-recognized,22
, 23
, 24
, 25
and the prevalence may be higher than our figure.The 2 radiologists showed good agreement in the diagnosis of most sinus conditions on the basis of panoramic imaging. There was very high agreement (≥0.912) in all reported diseases except the presence of mucosal thickening (0.677). The high agreement could be attributed, in part, to the large sample size and low prevalence of these conditions. One radiologist (S.C.) reported mucosal thickening more often than did the other (B.C.).
The recognition of mucosal thickening on OPG was the one area where there was marked disagreement. We believe this reflects the difficulty in diagnosing sinus disease by using panoramic imaging. There are multiple overlying shadows involving the maxilla, including the hard and soft palates, the tongue, and often the palatoglossal air space if the tongue is not placed on the hard palate during exposure. As previously mentioned, the sinus is longer in the anteroposterior plane than in the focal trough of the image, and this can exacerbate the confusion in diagnosis. Patient rotation can also produce asymmetry in sinus density, which, in turn, can produce artefacts over the sinuses. Finally, the decision to report disease as “mucosal thickening” or “mucus retention cyst/polyp” is not always clear.
Conclusions
This study confirmed that panoramic imaging has poor efficacy in the diagnosis of sinus disease, even in the hands of experienced dental radiologists. Our study showed that OPG can be useful in excluding sinus disease, with high specificity and NPV for sinus abnormalities other than mucosal thickening. There is high interobserver agreement between experienced dental radiologists in the interpretation of sinus disease on panoramic imaging. CT and CBCT remain the gold standard for the diagnosis of sinus disease, and despite the increased cost and radiation dose, 3-dimensional imaging is necessary for the definitive diagnosis of sinus pathology.
References
- Clinical Oral Anatomy: A Comprehensive Review for Dental Practitioners and Researchers.Springer, Basel, Switzerland2017: 170
- The accuracy of computed tomography in the diagnosis of chronic rhinosinusitis.Laryngoscope. 2003; 113: 125-129
- Technical assessment of a cone-beam CT scanner for otolaryngology imaging: image quality, dose, and technique protocols.Med Phys. 2012; 39: 4932-4942
- Radiographic evaluation of the maxillary sinus prior to dental implant therapy: a comparison between two-dimensional and three-dimensional radiographic imaging.Imaging Sci Dent. 2015; 45: 169-174
- Panoramic and tomographic implant studies: role in the diagnosis of sinus disorders.Med Oral Patol Oral Cir Bucal. 2010; 15: e611-e615
- The ability of panoramic radiography in assessing maxillary sinus inflammatory diseases.Korean J Oral Maxillofac Radiol. 2008; 38: 209-213
- Comparative assessment of panoramic radiography and CBCT imaging for radiodiagnostics in the posterior maxilla.Clin Oral Investig. 2014; 18: 293-300
- Contributions of reformatted computed tomography and panoramic radiography in the localization of third molars relative to the maxillary sinus.Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2004; 98: 342-347
- Comparison of panoramic radiography and CBCT to identify maxillary posterior roots invading the maxillary sinus.Dentomaxillofac Radiol. 2016; 4520160043
- Medication-related osteonecrosis of the jaw, osteoradionecrosis, and osteomyelitis: a comparative histopathological study.Braz Oral Res. 2018; 32: e23
- Statistical Methodology: I. Incorporating the prevalence of disease into the sample size calculation for sensitivity and specificity.Acad Emerg Med. 1996; 3: 895-900
- Diseases of the Sinuses: Diagnosis and Management. 2001;
- R: A Language and Environment for Statistical Computing.R Foundation for Statistical Computing, Vienna, Austria2017
- Methods for calculating sensitivity and specificity of clustered data: a tutorial.Radiology. 2012; 265: 910-916
- Evaluating diagnostic tests with imperfect standards.Am J Clin Pathol. 1990; 93: 252-258
- Handbook of Inter-Rater Reliability: The Definitive Guide to Measuring the Extent of Agreement Among Multiple Raters.ed4. Advanced Analytics, LLC, Chicago, IL2014
- Practical Statistics for Medical Research.Chapman and Hall, London, New York1991
- 2014–2015 National Health Survey.Commonwealth of Australia, Canberra ACT, Australia2015
- Association between computed tomography findings and clinical symptoms in chronic rhinosinusitis with and without nasal polyps.Eur Arch Otorhinolaryngol. 2017; 274: 2165-2173
- Acute radiology rarely confirms sinus disease in suspected recurrent acute rhinosinusitis.Int Forum Allergy Rhinol. 2017; 7: 726-733
- Correlation of quantitative MR imaging findings with symptoms in patients with incidentally detected inflammatory sinonasal disease.Diagn Interv Imaging. 2018; 99: 65-72
- Odontogenic sinusitis an underappreciated diagnosis: our experience.Clin Otolaryngol. 2016; 41: 284-285
- Odontogenic sinusitis: developments in diagnosis, microbiology, and treatment.Curr Opin Otolaryngol Head Neck Surg. 2018; 26: 27-33
- Odontogenic sinusitis: an underdiagnosed condition.J Mass Dent Soc. 2014; 63: 46-47
- Odontogenic sinusitis: a comprehensive review.Acta Odontol Scand. 2017; 75: 623-633
Article info
Publication history
Published online: October 18, 2018
Accepted:
October 12,
2018
Received in revised form:
October 10,
2018
Received:
March 28,
2018
Identification
Copyright
© 2018 Elsevier Inc. All rights reserved.