Fractionated Stereotactic Radiosurgery for Brainstem Metastasis and Brainstem Tolerance

Metastases to the brainstem are typically not amenable to surgical resection, and have traditionally been treated with whole brain radiotherapy (WBRT). Due to concern about the relative radiosensitivity of the brainstem, stereotactic radiosurgery (SRS) is not widely used for this indication, and its oncologic and neurologic outcomes are not well studied. The goal of this retrospective study is to review the safety and efficacy of hypofractionated SRS for brainstem metastasis. Patients with brainstem metastases treated with SRS between 2009 and 2018 at our institution were retrospectively reviewed. Grade 2 complications were defined as worsening neurologic symptoms requiring new or increased use of steroids after SRS. A probit normal tissue complication probability (NTCP) model for brainstem tolerance was constructed assuming α/β of 2. A total of 47 individual brainstem lesions were treated over 1-5 fractions in 37 SRS treatments. Ten lesions were in the midbrain (21%), 30 in the pons (64%), and 7 in the medulla (15%). Nine (24%) and fourteen (38%) patients received WBRT and/or SRS to other brain lesions prior to brainstem SRS. Seven patients (19%) had more than one brainstem lesion (range 2-4 lesions), and 28 patients (76%) had more than one brain metastasis treated during the same treatment. Twenty four patients (65%) were treated with systemic therapy within 1 month of brainstem SRS, eight of whom received immunotherapy (22%). Median imaging follow up duration was 8 months, and the overall local control rate was 83%. The median total dose for 1, 3, and 5-fraction treatment was 15, 21, and 25 Gy respectively, and the median prescription isodose line was 85%. The mean volume of the planning target volume (PTV) was 258, 1751 and 1808 mm3 for 1, 3 and 5-fraction treatments. The median maximum point dose to the brainstem (Dmax) was 18.9, 26.0, and 27.4 Gy for 1, 3, and 5-fraction treatments. Twenty one patients (57%) were not on any steroid at 3-month follow up. Seven patients (19%) were able to continue or decrease their steroid dose after SRS. Nine patients (27%) required starting or escalating the dose of steroid, and only one (3%) was briefly hospitalized for worsening neurologic symptoms (grade 3). No grade 4 or higher toxicity was observed. The only non-dosimetric predictor of steroid use after brainstem SRS is pre-SRS steroid use (p = 0.002). NTCP modeling showed a steep dose response between post-SRS toxicity with radiation dose to 1cm3, 5cm3 and 10% of brainstem volume (D1cc, D5cc, D10%), but not Dmax. In 3 fractions, the 10% and 20% risk of grade 2-3 complications was 7 and 11 Gy for D10%, and was 8 and 14 Gy for 1cc. Stereotactic radiosurgery for metastasis in the brainstem may be appropriately safe and effective. There is a dose-volume relationship between irradiated brainstem and the risk of post-SRS complications. Further research is needed to confirm these dose constraints for optimal implementation.