SU‐FF‐T‐474: QA of Dose‐Volume Histogram

Abstract

Purpose: Among the various tools used in evaluating the quality of a treatment plan, dose‐volume histogram (DVH) has become the most important plan evaluation tool. DVH has also been used in TCP and NTCP calculations to aid in plan selection. The accuracy of a dose‐volume relationship of a particular structure depends on the size of the calculation dose grid, but more critically on the contouring of a structure and the dose algorithm used by a treatment planning system (TPS), especially in the presence of heterogeneities. This paper presents a comprehensive review on the calculation of DVH, and attempts to answer the following questions: (1) How do we verify that a DVH calculation is correct? and (2) What is the optimal grid size used in DVH calculation? Materials & Methods: DVHs of rectangular and cylindrical structures of known volumes are calculated for a single 6 MV beam setup. Various grid sizes (1.5mm–10mm) and calculation box sizes are studied. In addition, DVHs of various structures uploaded from VoxelQ software are calculated for 3 disease sites (Prostate, Lung and H&N) with different treatment planning systems (TPS). The volume information from the TPS is then compared with that from VoxelQ. Results: All grid sizes give the same total volume of a structure, even for such small structures as the lens (<0.5 cm3). However, the shape of the DVH is grid size dependent. Grid sizes 2–4 mm yield practically identical DVHs, even for small structures and with the presence of heterogeneity. Largest volume difference (>50%) between TPS with pencil beam algorithm and VoxelQ is observed for small structures. VoxelQ provides a good estimation of the volume of a structure, whereas a TPS may not necessarily provide adequate volume information. Conclusions: DVH QA should be performed with known structures from CT‐simulation as part of TPS commissioning.