Radiosurgery fractionation and post-treatment hemorrhage development for intact melanoma brain metastases

Purpose To assess, for intact melanoma brain metastases (MBM), whether single-fraction stereotactic radiosurgery (SRS) versus fractionated stereotactic radiotherapy (fSRT) is associated with a differential risk of post-treatment lesion hemorrhage (HA) development. Methods A single institution retrospective database review identified consecutive patients with previously unresected MBM treated with robotic SRS/fSRT between 2013 and 2021. The presence of lesion HA was determined by multi-disciplinary imaging review. Dosimetric variables were reported as biologically effective doses using an alpha/beta ratio of 2.5 (BED2.5). Statistical analysis was performed using mixed effect logistic regression for post-treatment HA and Cox frailty modeling for local control (LC). Results The cohort included 48 patients with 226 intact MBM treated with SRS/fSRT. Of lesions without prior HA, 63 of 133 lesions (47.4%) receiving SRS demonstrated evidence of post-treatment HA versus 2 of 24 lesions (8.3%) treated with fSRT (p = 0.01). A larger maximum BED2.5 was observed in lesions developing HA compared to no HA (238.3 Gy vs. 211.4 Gy; p = 0.022). 12-month LC was 65.7% (95% CI 37.2-87.3%) and 77.5% (95% CI 58.5-91.2%) for lesions demonstrating pre-treatment and post-treatment HA, respectively, with no local failure events observed within 12 months for non-hemorrhagic lesions (p < 0.001). Conclusion We found an increased incidence of post-treatment HA for intact MBM receiving a larger maximum BED2.5, which was significantly higher for single fraction treatments within our cohort. The presence of lesion HA, either pre- or post-treatment, was indicative of inferior LC. Further investigations of optimal dose and fractionation schedules for treatment of MBM in the era of immunotherapy are warranted.