Evaluation of dosimetric misrepresentations from 3D conventional planning of liver SBRT using 4D deformable dose integration

The purpose of this study is to evaluate dosimetric errors in 3D conventional plan- ning of stereotactic body radiotherapy (SBRT) by using a 4D deformable image registration (DIR)-based dose-warping and integration technique. Respiratory- correlated 4D CT image sets with 10 phases were acquired for four consecutive patients with five liver tumors. Average intensity projection (AIP) images were used to generate 3D conventional plans of SBRT. Quasi-4D path-integrated dose accumulation was performed over all 10 phases using dose-warping techniques based on DIR. This result was compared to the conventional plan in order to evalu- ate the appropriateness of 3D (static) dose calculations. In addition, we consider whether organ dose metrics derived from contours defined on the average intensity projection (AIP), or on a reference phase, provide the better approximation of the 4D values. The impact of using fewer (< 10) phases was also explored. The AIP- based 3D planning approach overestimated doses to targets by 1.4% to 8.7% (mean 4.2%) and underestimated dose to normal liver by up to 8% (mean -5.5%; range -2.3% to -8.0%), compared to the 4D methodology. The homogeneity of the dose distribution was overestimated when using conventional 3D calculations by up to 24%. OAR doses estimated by 3D planning were, on average, within 10% of the 4D calculations; however, differences of up to 100% were observed. Four-dimensional dose calculation using 3 phases gave a reasonable approximation of that calculated from the full 10 phases for all patients, which is potentially useful from a workload perspective. 4D evaluation showed that conventional 3D planning on an AIP can significantly overestimate target dose (ITV and GTV+5mm), underestimate normal liver dose, and overestimate dose homogeneity. Implementing nonadaptive quasi- 4D dose calculation can highlight the potential limitation of 3D conventional SBRT planning and the resultant misrepresentations of dose in some regions affected by motion and deformation. Where the 4D approach is unavailable, contouring on the full expiration phase may yield more accurate dose calculations, most relevant in the case of the healthy liver, but the absolute dose differences are in general small for the other healthy organs. The technique has the potential to quantify under- and over-dosage and improve treatment plan evaluation, retrospective plan analysis, and clinical outcome correlation.