Inter-rater Reliability and Standardization on Neuroimaging for Idiopathic Intracranial Hypertension
Nivethitha Arunkumar1, Areeba Nisar2, Isha Sharan1, Jenny Chan1, Scott Faro3, Prabath Mondel3, Neelu Jain3, Kiran Talekar3, Joga Chaganti3, Nicole Spare2, Reid Gooch4, Hsiangkuo Yuan2
1Thomas Jefferson University, Sidney Kimmel Medical College, 2Department of Neurology, 3Department of Radiology, 4Department of Neurosurgery, Thomas Jefferson University Hospital
Objective:
To evaluate the inter-rater reliability of idiopathic intracranial hypertension (IIH) neuroimaging criteria in patients with refractory (IIH).
Background:
The lack of operational cut-off values for the current IIH imaging diagnostic criteria causes potential variation in radiological interpretations. A standardized neuroimaging protocol was developed and investigated for its utility.
Design/Methods:
Brain MRI/MRV from 15 refractory IIH patients were reviewed independently by three experienced neuroradiologists at Thomas Jefferson University Hospital. Standardized neuroimaging interpretation criteria were utilized: sella herniation grade, posterior globe flattening (PGF), protrusion at the optic nerve head (ONP), optic nerve horizontal/vertical tortuosity (ONHT/ONVT), distention of the optic nerve subarachnoid space (ONSAS)≥2mm, distention of the optic nerve sheath diameter (ONSD) ≥5.5mm, index of transverse sinus stenosis (ITSS), and transverse sinus of either unilateral ≥50% stenosis with contralateral hypoplasia (≥40% narrower) or bilateral stenosis. Sequence utilization and inter-rater reliability (intraclass correlation coefficient) were assessed.
Results:
Several IIH imaging features showed high agreement: ONSAS≥2mm (1.00), ONSD≥5.5mm (0.72, 95%CI 0.29-0.91), ONVT (0.88, 95%CI 0.68-0.92), sella herniation grade (0.75, 95%CI 0.38-0.92), ITSS (0.85, 95%CI 0.63-0.95), 50% venous sinus stenosis (0.87, 95%CI 0.69-0.95) and pituitary height<4.8mm (0.75, 95%CI 0.38-0.92). Other imaging features showed low agreement: tonsil descent≥5mm, ONHT, PGF, and ONP. Quality assessment of imaging sequences showed only 5 (35.7%) axial orbit T2 thin-cut fat-saturated, and 6 (42.8%) coronal orbit T2 thin-cut fat-saturated were available.
Conclusions:
IIH imaging features showed high agreement justifying the use of standardized criteria in IIH assessments. Inter-rater discrepancy exists in particularly assessing orbits. Standardizing imaging protocol, especially including axial high-resolution T2w sequence, in our experience, will not only improve inter-rater reliability for certain low agreement criteria, including ONHT and PGF, but also will further improve objective measurements of ON sheaths reported above.