2026 Poster Presentations
P191: A NOVEL NOMENCLATURE SYSTEM FOR RADIOGRAPHIC ASSESSMENT OF PITUITARY MICROADENOMAS IN CUSHING'S DISEASE
Silky Chotai, MD; Alan Gordillo; Chris Karakassis; Matt Kiczek; Jonathan Lee; Divya Yogi-Morren; Pablo F Recinos, MD; Varun R Kshettry, MD; Cleveland Clinic Foundation
Introduction: Interpretation of pituitary microadenomas (<6 mm), specifically in Cushing’s disease, remains inconsistent, leading to uncertainty in clinical decision-making. An effective nomenclature must be simple, reproducible, and transparent about radiographic limitations. We introduce a novel system designed to standardize reporting and improve communication in the care of patients with pituitary microadenomas.
Methods: Patients with microadenomas and biochemically confirmed Cushing’s disease undergoing surgical resection were reviewed. Additional cases of normal MRIs in patients without pituitary pathology were included to avoid bias in rater scoring but not included in the analysis. Three blinded neuroradiologists scored MRIs using a proposed nomenclature: Definite (tumor visible in >1 orthogonal plane and >1 sequence), Probable (meets 1 criterion), Possible (tumor visible, but meets neither criterion), and Negative (no tumor visible). Cases were scored first on pituitary protocol (including fat saturation 3mm slices with 0.3mm spacing on T1 pre- and post-contrast, T2 coronal) then after a 4-week washout on skull base protocol (0.8 mm slice with isotropic voxels on T1 pre/post and CISS). Reliability was assessed using intraclass correlation coefficients with a two-way random-effects model. Sample size calculation indicated 34 patients for 80% power. Univariate analysis was performed.
Results: A total of 37 cases met the inclusion criteria. Figure 1 demonstrates individual radiologist scoring per nomenclature. On the MRI pituitary protocol, across the three raters, the mean scores were 25% (n=9) definitive, 33.3% (n=12) probable. 25% (n=9), and 16.7% (n=6). On MRI skull base protocol, the mean scores were 46.8% (n=17) definitive, 27.9% (n=10) probable. 10.8% (n=4), and 12.6% (n=5). Operative reports documented a tumor in 84.2% (n=32) of patients, pathology reports confirmed a tumor in 68.4% (n=26), and 73.7% (n=28) achieved biochemical remission postoperatively. Inter-rater reliability for the nomenclature was 0.76 using the MRI pituitary protocol and 0.79 using the MRI skull base protocol (p<0.001). Scoring a definitive tumor showed average sensitivities of 52.3%, and average specificities of 66.6% among the raters.
Conclusion: The proposed nomenclature for microadenomas demonstrates good inter-rater reliability and provides a standardized approach to MRI interpretation. In patients with Cushing’s disease, use of the skull base protocol increased the detection of definitive tumors compared with a pituitary protocol. Scoring a definitive tumor showed high specificity but moderate sensitivity across radiologists, highlighting that while definitive findings reliably indicate tumor presence, some microadenomas may be missed. This system may improve preoperative assessment and aid in consistent communication and reporting in patients with ACTH-secreting microadenomas.