High‐field MRI using standard pulse sequences has moderate to substantial interobserver agreement and good accuracy for differentiation between intracranial extra‐axial histiocytic sarcoma and meningioma in dogs

• Published in: Veterinary radiology & ultrasound Vol. 63; no. 2; pp. 176 - 184
• Main Authors: Mai, Wilfried, Burke, Emily E., Reetz, Jennifer A., Hecht, Silke, Paek, Matthew, Church, Molly E., Werre, Stephen R., Mariani, Christopher L., Griffin, John F., Glass, Eric N.
• Format: Journal Article
• Language: English
• Published: England 01-03-2022
• Subjects: Animals; brain; canine; Dog Diseases - pathology; Dogs; Histiocytic Sarcoma - diagnostic imaging; Histiocytic Sarcoma - veterinary; Magnetic Resonance Imaging - methods; Magnetic Resonance Imaging - veterinary; Meningeal Neoplasms - diagnostic imaging; Meningeal Neoplasms - veterinary; Meningioma - diagnostic imaging; Meningioma - veterinary; neoplasia; Observer Variation; reliability; Reproducibility of Results; Retrospective Studies; round cell; sensitivity; specificity; reliability; specificity; round cell; canine; sensitivity; brain; neoplasia
• ISSN: 1058-8183; 1740-8261
• DOI: 10.1111/vru.13038

Intracranial extra‐axial histiocytic sarcoma shares common MRI features with meningioma. As histiocytic sarcoma carries a generally worse prognosis than meningioma, the ability to differentiate between these two neoplasms is of clinical value. The aim of this retrospective diagnostic accuracy and observer agreement study was to evaluate the accuracy and reliability of high‐field MRI to differentiate between these two tumors, using standard pulse sequences and published MRI features. A total of 51 dogs were included (26 meningiomas and 25 histiocytic sarcomas). Magnetic resonance imaging examinations were independently assessed by three experienced board‐certified radiologists, evaluating 18 imaging features. They were asked to assign each case to one of three categories (meningioma, histiocytic sarcoma, and undetermined). Agreement for the MRI diagnosis across all three reviewers was moderate (κ 0.54) while paired interobserver agreement ranged from moderate to substantial (κ 0.58‐0.74) with percent agreement ranging between 86.1% and 87.7%. Overall, the probability of correctly diagnosing meningioma in a dog with this tumor ranged between 79.2% and 94.4%, and the probability of correctly diagnosing histiocytic sarcoma in a dog with this tumor ranged between 76.0% and 92.3%. The overall probability to diagnose the correct tumor, irrespective of type, ranged between 79.2% and 89.7%. Histiocytic sarcomas tended to have more extensive edema and more often had combined perilesional and distant meningeal enhancement affecting both pachy‐ and leptomeninges, while for meningiomas, meningeal enhancement tended to more commonly be perilesional and pachymeningeal. Imaging features that seemed more useful to make a correct diagnosis included “location/type of meningeal enhancement,” “osseous changes in the adjacent neurocranium,” “cystic changes,” and “herniation severity.”