J Neurol Surg B Skull Base 2022; 83(S 01): S1-S270
DOI: 10.1055/s-0042-1743616
Presentation Abstracts
Podium Abstracts

The Diagnostic Accuracy of Magnetic Resonance Imaging in Predicting Cavernous Sinus Invasion of Pituitary Tumors

Authors

  • Karam P. Asmaro

    1   Stanford Health Care, Stanford, United States
  • Ahmed Mohyeldin

    1   Stanford Health Care, Stanford, United States
  • Vera Vigo

    1   Stanford Health Care, Stanford, United States
  • Maximiliano A. Nunez

    1   Stanford Health Care, Stanford, United States
  • Juan Fernandez-Miranda

    1   Stanford Health Care, Stanford, United States
 

Objective: Despite major advances, pituitary tumors continue to offer unique challenges given their proximity to the cavernous sinus, whereby invasive behavior can dictate extent of resection and surgical outcome, especially in neuroendocrine-active tumors. The current classification system has been used for nearly three decades to predict cavernous sinus invasion. The aim of this study is to determine the diagnostic accuracy of the Knosp classification utilizing direct intraoperative observation and histopathology of the medial cavernous sinus wall as golden standards.

Methods: A total of 107 consecutive primary pituitary tumors were analyzed, 51% of which were neuroendocrine active. In our cohort, there were 38 corticotroph, 27 gonadotroph, 24 somatotroph, 17 lactotroph, and 11 other subtype pituitary adenomas. The golden standard was based on direct intraoperative examination of the medial wall of the cavernous sinus, and equivocal cases were followed by histopathological examination of the cavernous sinus wall.

Results: The highest degree of accuracy (73%) is found when using the Knosp score of 2 as the cut-off criteria to predict cavernous sinus invasion. The sensitivity and specificity are 36 and 94%, respectively, for all pituitary tumors combined with an area under a Receiver Operating Characteristic curve of 0.88. Diagnostic accuracy of cavernous sinus invasion varies by histopathological subtype: corticotroph (89%), gonadotroph (67%), lactotroph (76%), and somatotroph (46%). The sensitivities of detecting cavernous sinus invasion for gonadotroph, lactotroph, and somatotroph adenomas are between 20 and 25%. The negative predictive value for somatotroph adenomas is 35%.

Conclusion: Consistent with common usage, the diagnostic accuracy for the Knosp criteria is the highest at a cut-off of 2 in predicting cavernous sinus invasion (scores: 3–4). Invasion of the cavernous sinus varies by adenoma subtype which is likely related to tumor biology and invasive behavior. In patients with gonadotroph, lactotroph, and somatotroph adenomas, the criteria had a high false negative rate; failing to predict cavernous sinus disease in a large portion of patients. The probability of predicting cavernous sinus invasion was the lowest in patients with somatotroph adenomas. We suggest careful utilization of the criteria in preoperative planning, especially in cases of neuroendocrine active tumors which rely on complete excision to achieve biochemical remission.



Publication History

Article published online:
15 February 2022

© 2022. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany