A total of 219 meningiomas diagnosed by imaging criteria underwent gamma knife radiosurgery to a median marginal tumor dose of 14 Gy (range 8.9–20), a median treatment volume of 5.0 cm3 (range 0.47–56.5), and a median maximal dose of 28 Gy (range 22–50). The median follow-up was 29 months (range 2–164).Tumor progression developed in 7 cases, 2 of which turned out to be different tumors (metastatic nasopharyngeal adenoid cystic carcinoma and chondrosarcoma). One tumor was controlled, but the development of other brain metastases suggested a different diagnosis. The actuarial tumor control rate was 93.2% ± 2.7% at 5 and 10 years. The actuarial rate of identifying a diagnosis other than meningioma was 2.3% ± 1.4% at 5 and 10 years. The actuarial rate of developing any postradiosurgical injury reaction was 8.8% ± 3.0% at 5 and 10 years. No pretreatment variables correlated with tumor control in univariate or multivariate analysis. The risk of postradiosurgery sequelae was lower (5.3% ± 2.3%) in patients treated after 1991 (with stereotactic MRI and lower doses; p = 0.0104) and tended to increase with treatment volume (p = 0.0537).Radiosurgery of meningioma diagnosed by imaging without tissue confirmation is associated with a high rate of tumor control and acceptable morbidity but carries a small risk (2.3%) of an incorrect diagnosis.

Discussion

Stereotactic radiosurgery is an effective alternative to surgical resection or fractionated large-field conventional radiotherapy for the treatment of meningiomas . Radiosurgery seems to be an excellent way of managing well-circumscribed, small, benign, intracranial meningiomas. Tumor control rates varied in different series from 60% to 100%, depending on the proportions of atypical or malignant meningiomas, the proportion of postoperative patients treated, and the length of follow-up . A 93% tumor control rate was reported for the University of Pittsburgh series of 99 consecutive meningioma patients treated by radiosurgery between 1987 and 1992 with 5–10 years of follow-up. other representative series was the multicenter analysis by Kondziolka et al. of 203 benign parasagittal meningiomas treated by radiosurgery. The 5-year actuarial tumor control rate for primary radiosurgery (usually without tissue diagnosis) was 93% ± 4% compared with 60% ± 11% for postoperative radiosurgery (85% in-field tumor control with a 25% rate of out-of-field failure). Multivariate analysis correlated tumor progression with increasing tumor volume (greater than the median volume of 7.5 cm3 in the series) and with prior neurologic deficit. Trends were noted for increased tumor progression with prior surgery (p = 0.08) and lower marginal tumor dose (p = 0.06, less than the median of 15 Gy).

We chose to review patients who underwent radiosurgery with only an imaging diagnosis of meningioma. By excluding patients with prior surgery, who have a higher rate of marginal failure, we could more reliably assess the factors influencing local control.. Local failure can be difficult to separate from marginal failure. With marginal failure, it is difficult to identify precisely both where the tumor started to regrow and the dose in that location for a dose–response analysis. The second reason we chose to review patients receiving radiosurgery for an imaging-only diagnosis of meningioma was to define the chance of a misdiagnosis.. In particular, we were interested in the chance of a misdiagnosis that might have changed the treatment plan. Tumors or conditions that can be mistaken for benign meningioma include schwannomas, malignant meningioma, brain metastases (sometimes into a preexisting meningioma), glioblastoma, hemangiopericytoma, chordoma, chondrosarcoma, and central nervous system sarcoid ). Many of these tumors are reasonable to treat with radiosurgery to doses similar to those used for meningioma. By analyzing the tumors that failed after radiosurgery and were initially misdiagnosed as meningiomas, we can determine how often establishing a tissue diagnosis before treatment might be clinically important.

Tumor progression developed in 7 cases, 2 of which turned out to be different tumors (metastatic adenoid cystic carcinoma of the nasopharynx and chondrosarcoma). One additional tumor was controlled locally but the development of other brain metastases suggested a different diagnosis. Four patients were found to have 1-mm increases in tumor diameter that were not counted as recurrences because the tumors subsequently regressed (n = 2) or remained stable in size with protracted follow-up (n = 2). The actuarial tumor control rate was 93.2% ± 2.7% at 5 and 10 years. The actuarial rate of identifying a diagnosis other than meningioma was 2.3% ± 1.4% at 5 and 10 years

Twelve patients developed postradiosurgery neurologic sequelae, consisting of edema with headaches in four, worsening hemiparesis in two, mental status changes in two (one requiring steroids, one requiring a ventriculoperitoneal shunt), trigeminal nerve problems in three (numbness in two, tic pain in one), and a temporary visual field deficit in one. The actuarial rate of developing any postradiosurgical injury reaction was 8.8% ± 3.0% at 5 and 10 years. The development of postradiosurgery sequelae was significantly greater in the 28 patients treated between 1987 and 1991 (with CT targeting and higher doses; hazard rate ratio 4.45, 95% confidence interval 1.4–14.4), compared with patients treated later in the series with stereotactic MRI and lower doses ( Table 3). Between 1987 and 1991, the median marginal dose was 17 Gy (range 10–20), and between 1991 and 2000, the median marginal dose was 14 Gy (range 8.9–20). The actuarial rate of any symptomatic postradiosurgical sequelae was 5.3% ± 2.3% from 1991 to 2000 compared with 22.9% ± 9.3% from 1987 to 1991  Complications marginally correlated with treatment volume (p = 0.0537) and the volume of tissue receiving ge12 Gy (p = 0.0634).