Interobserver agreement in assessing plain radiographs of maxillary sinus

How Subjective is Nasal Endoscopy? A Study of Interrater Agreement using the Lund and Mackay Scoring System

Background

Scoring systems exist to standardize the recording of nasal endoscopy findings. The extent to which two surgeons will agree on the findings of nasal endoscopy in adults currently is unknown, although a study in children showed high rates of agreement between two observers. We studied the interobserver agreement of adult nasal endoscopy scored with the system proposed by Lund and Mackay. Method: A consecutive series of otolaryngology patients attending with a variety of nasal complaints were recruited. All were examined with a 0o 4-mm endoscope by two surgeons on the same clinic visit. Each independently recorded their findings using the Lund and Mackay scoring system. Neither surgeon was aware of the other's findings.

Results

Thirty patients were studied (16 women and 14 men). Because the endoscopic findings were recorded separately for each side of the nose, each surgeon made 60 observations. Interrater agreement was calculated as Cohen's κ, and suggested moderate agreement for edema (κ = 0.45), good agreement for crusting (κ = 0.62), and very good agreement for polyp (κ = 0.93) and discharge (κ = 0.84).

Conclusions

Two independent observers agree on the findings of nasal endoscopy in a high proportion of cases. Nasal endoscopy is a reliable tool, especially when a scoring system such as that of Lund and Mackay is used. (American Journal of Rhinology 18, 301–303, 2004)

The value of Waters' projection for assessing maxillary sinus inflammatory disease

OBJECTIVE

The significance of the Waters' projection for judging maxillary mucosal disease is, at the least, questionable. The aim of this study was to evaluate the diagnostic use of Waters' projection of the maxillary sinus with particular regard to sinus mucosal swelling as a consistent sign of maxillary sinusitis.

STUDY DESIGN

Forty consecutive adult patients were referred to an ear, nose, and throat surgeon for pain in the region of the paranasal sinus, recurrent mucopurulent rhinorrhea, and nasal congestion or obstruction for at least 3 months without any response to conservative treatment. Both conventional radiographs (Waters' projection) and coronal and axial computed tomography (CT) scanning of the maxillary sinuses were recorded. The conventional radiographs and CT scans, all made within an hour, were blinded and assessed in random order by 2 independent well-trained observers with standard radiodiagnostic criteria for sinus mucosal swelling. Intraobserver and interobserver agreements were quantified by calculating Cohen's kappa. The diagnostic significance of the Waters' projection was assessed with the CT scan images as criterion standard by calculating sensitivity and specificity, positive predictive value, likelihood ratio, and diagnostic odds ratio.

RESULTS

Cohen's kappa for the intraobserver agreement of Waters' examination was 0.96, and the intraobserver agreement for CT scanning was 0.92. The interobserver agreement for Waters' projection and CT scanning was 0.76 and 0.92, respectively. With CT scanning as criterion standard, the sensitivity and specificity of Waters' projections to detect maxillary sinus mucosal swelling were 83.3% and 69.2%, respectively. The positive predictive (diagnostic) value of Waters' projections was 83.3%, the positive likelihood ratio 2.7, and the diagnostic odds ratio 11.25.

CONCLUSIONS

From this study it can be concluded that Waters' projections do not necessarily rule out the presence of maxillary sinus mucosal swelling. Additional examinations may be indicated, especially in patients with compromised sinus clearance.