Question

• In children who have symptoms and history consistent with chronic sinusitis, what is the usefulness (sensitivity and specificity) of obtaining a CT scan compared to sinus radiography in making a diagnosis?

Clinical Bottom Lines

1. If the goal is to confirm the diagnosis of sinusitis, for instance, in preparation for endoscopic surgery, a sinus CT is superior to radiography in diagnostic accuracy and anatomic definition.
2. If the goal is to decide between allergic or non-allergic rhinitis versus sinusitis, a positive radiograph (or empiric therapy) may be sufficient to discriminate between a therapeutic choice of antibiotics or agents to manage rhinitis.
3. Comparing sinus x-rays to a gold standard of a limited sinus CT scan, the positive likelihood ratio is 4.05 and the negative likelihood ratio is 0.28.

Summary of Key Evidence

1. Seventy children were evaluated for sinus infection by sinus radiographs as well as CT scans.¹  Note:  Although 91 children entered the study, only the 70 children who had both radiographs and CT exam were included in the analysis. 
2. Inclusion criteria - Children with greater than three months of upper respiratory tract symptoms (rhinorrhea/sniffling, nasal congestion, sneezing, postnasal drip, headache, coughing and wheezing). 
3. Outcomes - Both radiographs and CT exams were interpreted according to pre-set criteria and inter-rater reliability was good.  Thirty-nine children were found to have sinus disease by CT exam while only 30 had evidence for disease by radiographs.  In addition, 6 children had sinus disease on radiographs that was not confirmed on CT (false positive).  Nine children with no disease on radiographs actually had disease by CT (false negative).  In this study, the pre-test probability of sinus disease was 56% (39/70).  The sensitivity of radiograph compared to CT was 77% (30/39) and the specificity of the radiograph to CT was 81% (25/31).  The positive likelihood ratio is 4.05 and the negative likelihood ratio is 0.28.
4. Conclusions - The difference between radiographs and CT for diagnosing sinus disease in this population is relatively small but favors CT exam.  With a positive likelihood ratio of  4.05, examination by radiographs in similar population of children will alter the probability from 59% pre-test to 84% post-test.  With a negative likelihood ratio of 0.28, a negative radiograph will mean 26% of children may still have significant sinus disease.

Additional Comments

• Another article was found that compares sinus radiographs to sinus CT exams,² but data were not presented in such a way that a 2x2 table could be constructed in order to calculate sensitivity, specificity and likelihood ratios. The article was written with an ENT evaluation and possible surgery in mind; thus, anatomic details were emphasized.  The conclusions still favored CT exam over radiographs for diagnostic accuracy and anatomic detail and the article contains important contributions regarding why radiographs may be more difficult than CT scans to interpret accurately.

Citation

1. Garcia, DP, Corbett, ML, Eberly, SM, et al. Radiographic imaging studies in pediatric chronic sinusitis.  J Allergy Clin Immunol, 94:523-530, 1994.
2. McAlister, WM, Lusk, R, and Muntz, HR. Comparison of plain radiographs and coronal CT scans in infants and children with recurrent sinusitis.  American J Radiol, 153:1259-1264, 1989.

CAT Author: Thomas Shope, MD

CAT Appraisers: General Pediatrics Faculty Development Team

Date appraised: March 20, 1999