Surgically targeted radiation therapy versus stereotactic radiation therapy: A dosimetric comparison for brain metastasis resection cavities.

Surgically targeted radiation therapy (STaRT) with Cesium-131 seeds embedded in a collagen tile is a promising treatment for recurrent brain metastasis. In this study, the biological effective doses (BED) for normal and target tissues from STaRT plans were compared with those of external beam radiotherapy (EBRT) modalities. Nine patients (n = 9) with 12 resection cavities (RCs) who underwent STaRT (cumulative physical dose of 60 Gy to a depth of 5 mm from the RC edge) were replanned with CyberKnifeⓇ (CK), Gamma KnifeⓇ (GK), and intensity modulated proton therapy (IMPT) using an SRT approach (30 Gy in 5 fractions). Statistical significance comparing D95% and D90% in BED 10Gy (BED 10Gy 95% and BED 10Gy 90%) and to RC + 0 to + 5 mm expansion margins, and parameters associated with radiation necrosis risk (V8 3Gy , V10 3Gy , V12 3Gy and V24 3Gy) to the normal brain were evaluated by a Wilcoxon-signed rank test. For RC + 0 mm, median BED 10Gy 90% for STaRT (90.1 Gy 10 , range: 64.1–140.9 Gy 10) was significantly higher than CK (74.3 Gy 10 , range:59.3–80.4 Gy 10 , p = 0.04), GK (69.4 Gy 10 , range: 59.8–77.1 Gy 10 , p = 0.005), and IMPT (49.3 Gy 10 , range: 49.0–49.7 Gy 10 , p = 0.003), respectively. However, for the RC + 5 mm, the median BED 10Gy 90% for STaRT (34.1 Gy 10 , range: 22.2–59.7 Gy 10) was significantly lower than CK (44.3 Gy 10 , range: 37.8–52.4 Gy 10), and IMPT (46.6 Gy 10 , range: 45.1–48.5 Gy 10), respectively, but not significantly different from GK (34.1 Gy 10 , range: 22.8–47.0 Gy 10). The median V24 3Gy was significantly higher in CK (11.7 cc, range: 4.7–20.1 cc), GK(6.2 cc, range: 2.3–11.9 cc) and IMPT (19.9 cc, range: 11.1–36.6 cc) compared to STaRT (1.1 cc, range: 0.0–7.8 cc) (p < 0.01). This comparative analysis suggests a STaRT approach may treat recurrent brain tumors effectively via delivery of higher radiation doses with equivalent or greater BED up to at least 3 mm from the RC edge as compared to EBRT approaches. [ABSTRACT FROM AUTHOR]