Single Fraction LINAC-Based SRS Re-Treatment of Brain Metastases Using Conservative Doses

Abstract

Listen

Translate article

We sought to evaluate the efficacy and safety of single fraction stereotactic radiosurgery (SRS), using conservative doses (relative to maximum tolerated doses from RTOG 9005), for brain metastases that recurred after initial SRS also with conservative doses.Under an IRB approved protocol, brain metastases that underwent retreatment with single-fraction, LINAC-based SRS were identified from 2014-2018. For each treated brain metastasis, histology, prescribed dose, lesion size, local control, and symptomatic radiation necrosis were recorded. Control was analyzed by Kaplan-Meier and Cox models.From an initial database containing 597 individual lesions210, 52 lesions in 22 were identified that recurred after initial SRS, for a total of 59 courses of retreatment (with some lesions being treated 3 times). Of the 52 initial lesions, 16 were from breast, 23 non-small cell lung, 9 melanoma, and 4 colorectal primaries. The median size and volume at first treatment were 0.9 mm and 1.47 cc respectively (ranges 0.3-31 mm and 0.2-13.4 cc respectively). Median dose covering 99% of the PTV was 14.6 Gy on initial treatment (10.7-18.2 Gy). Median time to re-treatment was 6.0 months (1-31). The median size, volume, and 99% PTV dose at re-treatment was 1.4 cm, 2.76 cc, and 15 Gy respectively. Both size (P = 0.006) and volume (P = 0.04), but not prescribed dose (P = 0.06) were significantly different when comparing initial and re-treatment lesion and treatment characteristics. The crude local control at 6 months was 58% after initial treatment (with 42% of recurrences occurring at > 6 months), and 70% after re-treatment (P = 0.02). Seven lesions in 6 patients developed symptomatic radiation necrosis after retreatment, 3 confirmed pathologically, 2 by magnetic resonance spectroscopy, and the remainder by imaging characteristics in absence of CNS progression. The mean normal brain V10 for the re-treatment course was 6.4 cc (range 3.8-18.4 cc) in these lesions. In a Cox multivariate model including: initial versus retreatment course, lesion volume, and PTV, retreatment was associated with improved local control (0.49, HR 0.29-0.83 P = 0.008).In this series of single fraction SRS using conservative doses, re-treated lesions showed improved control relative to initial treatment without significant increases in prescription dose in subsequent courses. These findings suggest that a threshold effective dose was not met on initial course but was able to be successfully made up later. Symptomatic necrosis after retreatment was relatively high, despite relatively low V10.M.A. Cummings: None. K.Y. Usuki: None. S. Hardy: None. M.T. Milano: Honoraria; UpToDate, self.