18F-FDG PET in squamous cell carcinoma of the oral cavity and oropharynx: a study on inter- and intraobserver agreement.

PURPOSE

Good observer agreement is mandatory for an effective imaging technique. However, little is known about the observer agreement of fluorine-18 fluorodeoxyglucose ((18)F-FDG) positron emission tomography (PET) in head and neck squamous cell carcinoma. The aim of the present study was to evaluate the inter- and intraobserver agreement of interpretations of (18)F-FDG PET in head and neck SCC and to assess the influence of observer experience, tumor localizing, and tumor size on the agreement.

PATIENTS AND METHODS

(18)F-FDG PET scans of 80 patients with oral and oropharyngeal SCC were reassessed twice by 2 experienced nuclear medicine physicians and 2 residents in nuclear medicine. The absolute agreement and Cohen's kappa were calculated by comparing the results of the 4 observers for the primary tumor, cervical metastases, and distant metastases/second primary tumor. To analyze the sensitivity and specificity, the results were compared with the findings from the histologic specimens or the follow-up data.

RESULTS

The interobserver agreement of the nuclear medicine physicians revealed an absolute agreement and kappa of 0.91 and 0.58 for detecting the primary tumor, 0.94 and 0.83 for detecting cervical metastases, and 0.85 and 0.53 for detecting distant metastases/second primary tumors, respectively. The intraobserver agreement was greater overall than the interobserver agreement. Compared with the nuclear medicine physicians, the residents scored lower in interobserver agreement. The interobserver agreement decreased when localizing the malignancy more precisely. The agreement and sensitivity increased with tumor size. However, for small metastases, a high observer agreement was found owing to the nondetection of these malignancies.

CONCLUSIONS

Good inter- and intraobserver agreement in SCC in the oral cavity or oropharynx with (18)F-FDG PET was found. Observer experience had limited influence on observer agreement. However, the agreement level decreased when a more precise anatomic tumor localization was required.