Prediction of cavernous sinus invasion in patients with Cushing's disease by magnetic resonance imaging.

Objective: Cavernous sinus invasion (CSI) in Cushing's disease (CD) negatively affects the probability of complete resection, biochemical cure, and need for adjuvant therapy. However, the prediction of CSI based on MRI findings has been inconsistent and variable. Among macroadenomas, the Knosp classification is the most widely utilized radiographic predictor of CSI, but its accuracy in predicting CSI and the probability of gross-total resection is limited in the setting of microadenomas or Knosp grade 0-2 macroadenomas. The authors noticed that the presence of a triangular shape of adenomas adjacent to the cavernous sinus on coronal MR images is frequently associated with CSI. The authors aimed to determine the correlation of this radiographic finding ("sail sign" [SS]) with CSI.

Methods: The authors performed a retrospective review of all patients with a pituitary lesion < 20 mm and a biochemical diagnosis of CD treated with endoscopic or microscopic transsphenoidal resection from November 2007 to May 2017. Overall 185 patients with CD were identified: 27 were excluded for negative preoperative imaging, 32 for lacking tumors adjacent to the sinus, 7 for Knosp grade 3 or higher, and 4 for inadequate intraoperative assessment of the CSI. Following application of inclusion and exclusion criteria, 115 cases were available for statistical analysis. Intraoperative CSI was prospectively evaluated at the time of surgery by one of two neurosurgical attending surgeons, and MRI data were evaluated retrospectively by a neurosurgical resident and attending neuroradiologist blinded to the intraoperative results.

Results: A positive SS was identified in 23 patients (20%). Among patients with positive SS, 91% demonstrated CSI compared to 10% without an SS (p < 0.001). Using the SS as a predictor of CSI provided a sensitivity of 0.7 and a specificity of 0.98, with a positive predictive value (PPV) of 0.91 and a negative predictive value of 0.9. Among patients with positive SS, 30% did not achieve immediate postoperative remission, compared to 3.3% of patients without an SS (p < 0.001).

Conclusions: The presence of a positive SS among Cushing's adenomas adjacent to the CS provides strong PPV, specificity, and positive likelihood ratio for the prediction of CSI. This can be a useful tool for preoperative planning and for predicting the likelihood of long-term biochemical remission and the need for adjuvant radiosurgery.