Abstract
PURPOSE We aimed to evaluate the frequency and neuroimaging features of Rathke's cleft cysts (RCCs) in children examined for endocrine-related diseases and to determine changes in the neuroimaging features of RCCs during the follow-up of children. We hypothesize that RCCs are being more commonly diagnosed in children with endocrine-related diseases and most of the RCCs show neither fluid intensity nor intensity due to high protein content on magnetic resonance imaging (MRI). METHODS After approval by the local ethics committee, the medical records and contrast-enhanced pituitary MRI of 833 children (boys/girls, 338/495; mean age±SD, 9.4±3.7 years) were retrospectively reviewed between January 2016 and January 2019. The size, location, signal intensities, and postcontrast enhancement pattern of RCCs were assessed by a pediatric radiologist. Same imaging features were also independently reviewed by another radiologist to determine the interobserver agreement by using the kappa statistics (κ) and intraclass correlation coefficient (ICC). RESULTS RCC was evident on MRI in 13.5% of the patients (boys/girls, 39/74; mean age±SD, 9.8±3.9 years). The mean size of RCCs was 5.5 mm (range, 3.1-8.5 mm). An RCC frequency higher than expected was found in patients with central precocious puberty, diabetes insipidus, and hypersecretion of prolactin (P = 0.007). The mean size of RCCs did not show significant differences among the clinical indications for MRI (P ≥ 0.461). All RCCs showed abnormal signal on T2-weighted image and most (89%) showed neither fluid intensity nor intensity due to high protein content (i.e., isointense on T1-weighted imaging and hypointense on T2-weighted imaging compared with the normal anterior pituitary gland). Eighty-four patients with RCCs (74%) had follow-up MRI and the mean follow-up was 1.5 years. In follow-ups, five RCCs disappeared; the mean size of 10 RCCs increased and that of 6 RCCs decreased. These size changes were not statistically significant (P = 0.376). No signal intensity changes of RCCs were seen during the follow-up, except for 4 RCCs, whose protein content increased over time and T1 signals increased on imaging. Interobserver agreements were almost perfect for the MRI findings of RCCs (κ and ICC range, 0.81-1, P < 0.001). CONCLUSION RCCs were not uncommon in patients examined for endocrine-related diseases, and nearly 1 in 10 patients had an RCC. The size and signal intensities of RCCs may change over time and the evolution of RCCs is unpredictable. Most RCCs showed neither fluid intensity nor intensity due to high protein content on MRI, and all RCCs had an abnormal signal on T2-weighted imaging, thus eliminating the need to administer a contrast agent at follow-up imaging of the patients.