Accuracy is a quantity, not a quality: Reflections on required and achievable accuracy in external-beam photon and particle therapy.

This review provides a succinct up-to-date discussion of required and achievable accuracy in MV photon beam and particle therapy, focusing on technology and techniques reflecting today's general clinical practice for radical radiotherapy. The major advances in recent decades have enabled a new degree of geometric and dosimetric precision to be achieved. These are related to dosimetry standards, imaging for treatment planning, dose calculation algorithms and approaches, fluence-modulated techniques facilitating conformality optimisation, image guidance (IG), IG-enabled adaptive and motion management methods and volume-based dose prescription concepts. In addition, evolving quality assurance methodologies have contributed substantially to improve the dosimetric accuracy. Consequently, in state-of-the-art radiotherapy the emphasis has shifted from accuracy of reference (prescription) dose delivery to accuracy of spatial dose distribution delivery, with greater focus on geometric uncertainties. The relationship between delivered dose and the resulting biological effect is examined, to consider the clinical impact of radiotherapy accuracy. Developments in required and achievable accuracy for photon beams are discussed, including insights from audits and patient specific quality assurance. Based on the steeper dose-response relations, the dosimetric accuracy requirements are still similar to those in the late 1980's, but the geometric accuracy demands have become significantly more stringent. In optimal conditions, radiotherapy practice can achieve these standards, but this necessitates a continued strong focus on treatment precision and quality. Robust quality assurance programmes, evolving in step with conceptual and methodological radiotherapy developments, remain essential to meet the accuracy requirements. As regards particle therapy, there are some significant uncertainty differences to photon therapy, but the models and clinical data are not yet sufficiently detailed to confidently predict differences in accuracy requirements compared to photon beam therapy.