Clinical evidence for dose tolerance of the central nervous system in hypofractionated radiotherapy

Background Stereotactic radiosurgery (SRS), stereotactic body radiotherapy (SBRT), and stereotactic ablative body radiotherapy (SABR) are commonly used in the treatment of central nervous system (CNS) disease. This study has refined the radiation toxicity estimates for some normal tissues of the CNS based on review and analysis of the clinical evidence for single fraction radiosurgery, hypofractionated SBRT, and conventionally fractionated radiation therapy. Methods Published guidelines and protocols are reviewed. In the past, many normal tissue tolerances were compiled based on the experience of the investigators and publications in the literature. Some tolerances were determined by modeling or calculation using the existing biological formulas, in particular the linear quadratic (LQ) model. In the present study, the estimate of risk for each dose tolerance limit in some CNS tissues is provided exclusively with normal tissue complication probability (NTCP). The clinical outcomes are compared to understand the difference in biological effect between radiosurgery and radiotherapy. Results Normal tissue dose tolerances and the corresponding complication rates are provided for brainstem, optic nerves, cochlea, and spinal cord, including single fraction SRS, five-fraction SBRT, and conventional radiation therapy. Calculation of biologically effective dose (BED) or single fraction equivalent dose (SFED) alone using the LQ model conveys no consensus on the biological effect across different fractionations. Comparison of conventional radiation therapy to brain and spinal cord with single fraction equivalent dose leads to even conflicting clinical outcomes. Conclusions Effective differences between single fraction SRS and conventional radiotherapy need to be better understood. The existing biological model might not be valid to predict the radiosurgical outcomes based on conventionally fractionated radiotherapy. However, application of the statistical dose response models of clinical SRS and SBRT outcomes data to selected current dose tolerance guidelines into simple tables can be a clinically useful resource.