OBSERVER VARIATION IN THE SCINTIGRAPHIC DIAGNOSIS OF SOLITARY COLD THYROID LESIONS

Abstract
In order to evaluate the reproducibility of the diagnosis of solitary cold thyroid lesions, two specialists in nuclear medicine and two specialists in endocrinology independently twice read 240 thyroid 99mTc pertechnetate scintigrams. No significant difference or interaction between the results obtained from the right and the left lobe was found. A solitary cold lesion was diagnosed in 100 of the 480 lobes; however, in only 30 did all four observers agree upon the diagnosis. Interobserver variation was determined by pairwise comparison of observers. The observed agreement was between 0.91 and 0.94. After adjusting for expected chance agreement, kappa values between 0.57 and 0.70 were found. Intraobserver variation was smaller than interobserver variation, revealing agreement rates of 0.93–0.96 and kappa values between 0.71 and 0.77. Agreement was related to large lesions, lesions located centrally in the lobe, and ovoid-shaped lesions. Clinicians should be aware to what extent they can rely on the information they use in their decisions. The considerable inconsistency in the evaluation of cold lesions on thyroid scintigrams should be taken into consideration, and calls in question the value of using thyroid scintigrams for deciding whether a patient should be referred for operation or biopsy.