Stereotactic radiotherapy (SRT) of brain metastases is considered effective when long-term local control is obtained. However, dose-effect data are scarce. We, therefore, performed a systematic literature search to assess the evidence concerning the relation of SRT dose and local control probability.

A search was performed for papers describing patients treated with SRT for brain metastases, published from 1990 through 2009, in the electronic databases Medline (Pubmed) and Embase. We selected only papers reporting actuarial local control probability, in which a fixed dose had been prescribed and in which the size of the metastases was given. Series with SRT as a boost after whole brain irradiation (WBI) or with SRT after surgery were excluded. From the selected papers we extracted data on dose, local control rates and necrosis rates. Biological effective doses of the linear-quadratic-cubic model, using an α/β of 12Gy (BED(12)), were calculated and a dose-response curve was constructed.

Eleven papers fulfilled the selection criteria for further analysis. Six-month local control rates were higher than 80% in almost all the series irrespective of dose. Twelve-month local control rates, however, varied and were higher than 80%, higher than 60% and lower than 50% with single doses of ≥21Gy, ≥18Gy and ≤15Gy, respectively, and 70% or higher with fractionated SRT (FSRT). A BED(12) of at least 40Gy was associated with a twelve-month local control rate of 70% or more.

Local control after single fraction SRT is highly dependent upon dose and is high (>80%) after 21Gy or more, but low (<50%) after 15Gy or less. We conclude that SRT for brain metastases should preferably be applied with a BED(12) of at least 40Gy corresponding with a single fraction of 20Gy, two fractions of 11.6Gy or three fractions of 8.5Gy.