Predictors of Recurrence after Gamma Knife Radiosurgery for Cavernous Sinus Meningiomas

Background The factors associated with tumor progression of cavernous sinus meningiomas after stereotactic radiosurgery have been poorly studied.

Methods We retrospectively reviewed our 25-year radiosurgery database identifying 1,982 radiosurgery procedures for skull base meningiomas. Of those, 292 were performed on radiosurgery-naive patients with cavernous sinus meningiomas. To identify factors associated with tumor progression, we then performed a subgroup analysis by matching the 33 tumors which progressed despite radiosurgery with 33 tumors of similar volume which did not progress. Cox proportional hazard ratios (HR), chi-square tests, t-tests, correlation coefficients, and Kaplan−Meier log-rank tests were utilized as appropriate.

Results When evaluating the entire cohort of 292 cavernous meningiomas undergoing first-time radiosurgery, 33 tumors recurred during follow-up (11.3%). We found that progressive tumors had a significantly larger volume when compared with the ones that did not (mean tumor volume of 8.2 mL for progressive tumors vs. 4.7 mL for controlled tumors, p = 0.0004). To effectively eliminate volume as the main determinant of recurrences, we proceeded with an in-depth analysis of the volumetrically matched cohort of 66 patients. Our matching process was successful, resulting in comparable treatment volumes between tumors that recurred and those that did not (mean: 8.2 vs. 7.9 mL, respectively, p = 0.815). The average Kaplan−Meier follow-up was also comparable between the two groups (73.1 months for the tumors that progressed vs. 81.7 months for those that did not, p = 0.86, range for the whole group: 6−206 months). A major predictor of recurrence in the volumetrically matched cohort was prior surgery (HR = 4.3, p = 0.001). Delayed treatment for progressive tumors after surgery (HR = 2.3, p = 0.018) was also predictive of progression after radiosurgery, whereas early radiosurgery for residual tumors was not (p = 0.1). In addition, the presence of other meningiomas (HR = 2.7, p = 0.022), and increasing K_i-67s correlated with recurrence (r = 0.59, p = 0.05). Presentation with trigeminal (HR = 2.4, p = 0.035) or auditory dysfunction (HR = 3.2, p = 0.004) was also associated with an increased risk for recurrence after radiosurgery. Interestingly, the group that recurred had slightly higher margin doses than the group that did not (mean: 12.8 vs. 11.8 Gy, p = 0.43). We did not find any significant differences with regard to maximum doses, or the number of isocenters used. Furthermore, there were no significant differences with regard to the grade of the tumors, age, sex, presenting Karnofsky scores, or other presenting signs or symptoms (including all other cranial neuropathies, motor or sensory deficits, or ataxia).

Conclusion Although tumor control after radiosurgery is increased with smaller tumors, our results suggest that prior surgery maybe associated with an increased risk of recurrence after radiosurgery despite adequate margin doses. The risk, however, was increased only when radiosurgery was performed in a delayed fashion after surgery for recurrent/progressive tumors, while residual tumors treated early after surgery were better controlled. A possible explanation could be related to difficulties in targeting tumors after surgery or tumor recurrences. As such, the results of this study suggest a benefit of early radiosurgery for residual tumors.