A Dose-Response Model of Local Tumor Control Probability After Stereotactic Radiosurgery for Brain Metastases Resection Cavities

Abstract

Purpose: Recent randomized controlled trials evaluating stereotactic surgery (SRS) for resected brain metastases question the high rates of local control previously reported in retrospective studies. Tumor control probability (TCP) models were developed to quantify the relationship between radiation dose and local control after SRS for resected brain metastases. Methods and Materials: Patients with resected brain metastases treated with SRS were evaluated retrospectively. Melanoma, sarcoma, and renal cell carcinoma were considered radio-resistant histologies. The planning target volume (PTV) was the region of enhancement on T1 post-gadolinium magnetic resonance imaging plus a 2-mm uniform margin. The primary outcome was local recurrence, defined as tumor progression within the resection cavity. Cox regression evaluated predictors of local recurrence. Dose-volume histograms for the PTV were obtained from treatment plans and converted to 3-fraction equivalent doses (α/β = 12 Gy). TCP models evaluated local control at 1-year follow-up as a logistic function of dose-volume histogram data. Results: Among 150 cavities, 41 (27.3%) were radio-resistant. The median PTV volume was 14.6 mL (range, 1.3-65.3). The median prescription was 21 Gy (range, 15-25) in 3 fractions (range, 1-5). Local control rates at 12 and 24 months were 86% and 82%. On Cox regression, larger cavities (PTV > 12 cm3) predicted increased risk of local recurrence (P = .03). TCP modeling demonstrated relationships between improved 1-year local control and higher radiation doses delivered to radio-resistant cavities. Maximum PTV doses of 30, 35, and 40 Gy predicted 78%, 89%, and 94% local control among all radio-resistant cavities, versus 69%, 79%, and 86% among larger radio-resistant cavities. Conclusions: After SRS for resected brain metastases, larger cavities are at greater risk of local recurrence. TCP models suggests that higher radiation doses may improve local control among cavities of radio-resistant histology. Given maximum tolerated doses established for single-fraction SRS, fractionated regimens may be required to optimize local control in large radio-resistant cavities.