Accuracy of Tumor Surveillance of Pediatric Adamantinomatous Craniopharyngioma Using Noncontrast Brain MRI

BACKGROUND AND PURPOSE: Adamantinomatous craniopharyngioma (ACP) is a benign but histologically complex brain tumor that arises in the sellar or suprasellar region. Contrast-enhanced MRI of the brain is the standard of care for radiologic follow-up. Numerous studies have demonstrated gadolinium retention in the brain and body as a result of gadolinium-based contrast agent (GBCA) usage in MRI. While the clinical consequences of GBCA retention remain unknown, the concern for potential injury combined with advances in noncontrast imaging has promoted efforts to determine the necessity for contrast enhancement in the care of pediatric patients, particularly those with brain tumors. This study aims to evaluate the agreement and interrater reliability between noncontrast and contrast-enhanced MRI in the routine surveillance of children with ACP.

MATERIALS AND METHODS: We conducted a retrospective review of MR imaging for 25 patients with ACP undergoing routine imaging follow-up assessment. Pre- and postcontrast-enhanced MRI sequences were evaluated by 2 pediatric neuroradiologists. Three consecutive 1-year follow-up scans were graded based on the characteristics of the solid and cystic components of the residual lesion. Percent agreement and interrater reliability were determined for grading between contrasted and unenhanced MRI scans.

RESULTS: The mean age of patients was 8.72 ± 4.59 years, and the mean time between scans was 12.60 ± 4.08 months and 14.48 ± 7.77 months from the first to second and second to third scans, respectively. Interrater reliability between grading by using contrast-enhanced and unenhanced MRIs was high. We observed a Cohen κ of 0.961 (95% CI, 0.92–1.0) when evaluating the cystic tumor and a Cohen κ of 0.80 (95% CI, 0.68–0.92) when assessing the solid tumor. The percent agreement between grading by using contrast-enhanced and unenhanced MRI for the cystic tumor component was 98.7% (74/75) for radiologist 1, and 97.3% (73/75) for radiologist 2. The percent agreement between grading by using contrast-enhanced and unenhanced MRI for the solid tumor component was 93.3% (70/75) for both radiologists.

CONCLUSIONS: For children with ACP and known residual disease, noncontrast MRI may be sufficient for the assessment of solid tumor or cyst growth.