Session Item

Voxel-based methods have already been applied for unveiling spatially variable dose–effect patterns in radiotherapy, allowing the identification of sub-regions at risk in several anatomical locations such as the lungs, the heart, head and neck (H&N), the rectum or the bladder. Overall, the principles of voxel-based methods rely on the analysis of the local dose–toxicity relationship at fine spatial scales, through (i) non-rigid registration , (ii) dose resampling to a common space, and (iii) voxel-wise comparisons between patients with and without toxicity. The located regions have been shown to be more predictive for toxicity than the whole organs.  These steps are well established from the methodological point of view, yet the translation of voxel-based models into clinical practice it is still fragmented, lacking research in several areas:

-    Widespread validation of spatial suborgan models on external data.
-    Physiological interpretation of subregions resulting from these models.
-    Transfer to the patient-specific planning in a personalized treatment context.
-    Translation into planning constraints to be transferred to the TPS.
-    Evaluation of the dosimetric and clinical benefit in different RT scenarios including alternative techniques: VMAT vs SBRT, Conventional fractionation vs hypofractionation, photons vs protons, etc.
-    Automation in segmentation and planning to make it available to the everyday clinical practice.

This presentation will describe The methodology and pitfalls of voxel-wise analysis.  We discuss ways on how Contemporary/Current RT protocols can incorporated the findings in order to move RT from the one--fits-all concept towards tailored RT protocols.