Abstract
INTRODUCTION: Comparison of local control rates at 12 months (LC) for brain metastases (BM) treated with stereotactic radiosurgery (SRS) is hindered by the dose-volume paradox wherein larger metastases may receive lower doses. This phenomenon will have a significant effect on differences in LC between centres, some treating small asymptomatic metastases. Linear quadratic (LQ) or linear quadratic cubic (LQC) tumour control probability (TCP) models are proposed as audit tools to assist such comparisons. METHOD: Series were identified reporting LC by volume of BM. The parameters of the LQ and LQC models α/β, α and DL were derived by adjusting these variables to achieve the highest correlation between the observed rate of LC and the modelled TCP for variable clonogen density (CD) 107-109 cm-1. The models obtained were then tested on a series of 48 brain metastases treated with single fraction SRS in a single centre using the surface dose and volume of individual metastases. RESULTS: 8 single fraction series with sufficient LC data were identified. Parameters derived for the models for CD 107 and 109 respectively were LQ α/β = 16.7,15.7 Gray (Gy); α = 0.49, 0.59 Gy-1; LQC α/β = 10.3, 9.6 Gy; α = 0.43, 0.52 Gy-1 and DL 26.2, 26.3 Gy. All 4 models gave a similar TCP at 12 months for the test population, e.g. LQ CD 107 gives TCP of 0.58 (95% CI 0.47-0.69) predicting LC in the test population between 47 and 69% compared to the observed LC of 60% (95% CI 46-74%). CONCLUSION: The models presented may assist in the comparison of LC between series with variable volumetric case mix. However, modification to correct for histological heterogeneity would help to refine these models.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call