Accuracy and reliability of oral maxillofacial radiologists when evaluating cone-beam computed tomography imaging for adenoid hypertrophy screening: A comparison with nasopharyngoscopy

Abstract

Objective To determine how accurate and reliable oral maxillofacial radiologists (OMFRs) are in screening for adenoid hypertrophy when using cone-beam computed tomography (CBCT) imaging compared with nasopharyngoscopy (NP). Study Design CBCT scans of 10 patients with distinct levels of adenoid hypertrophy were randomly selected. Fourteen board-certified OMFRs classified the levels of hypertrophy. The intraclass correlation coefficient (ICC) was used to assess accuracy by comparing their diagnosis against an NP diagnosis, which is the reference standard. OMFRs' interreliability was assessed. Kappa statistics were used to analyze dichotomous data from healthy and unhealthy patients. Results Overall, the reliability among OMFRs was good (ICC = 0.79 with confidence interval [CI] 0.63-0.93). The "statistical mode" was very good (ICC = 0.81; CI 0.43-0.94). The accuracy of OMFRs against NP was good (ICCmean = 0.69; CI 0.43-0.94). On average, the Kappa statistics (Kmean = 0.77; CI 0.62-0.92) demonstrated a good agreement between OMFRs and NP diagnoses. The individualized results from each evaluator were presented and investigated according to their performance. Conclusions Compared with the reference standard, the accuracy of OMFRs to classify adenoid hypertrophy on a four-level scale was moderate to strong and improved when adenoid hypertrophy was classified as healthy or unhealthy. The reliability of the OMFRs was greater than 80%, assuring their consistency and reliability on screening adenoids hypertrophy via CBCT.