Analysis of 3D and 4D proton treatment planning for hepatic tumors

Abstract

The purpose of this study is to assess the difference between 4D liver dose calculations versus standard 3D treatment planning and to investigate the dosimetric gain of gating on radiation dose to normal tissue. 4DCT scans are collected for 25 patients with hepatic tumors treated by proton radiotherapy. The 4D treatment planning process explicitly takes into account respiratory motion of abdominal organs. A 4DCT scan consists of 10 3D anatomical states, each at an instant of time in the respiratory cycle. 4D treatment planning includes 1) propagating the target contours, drawn by a physician on one phase, to all breathing phases using deformable registration, 2) calculating the compensating bolus for proton therapy, and then 3) calculating 4D dose distributions. Dose volume histograms are used to compute the effective uniform dose (EUD) delivered to normal liver. We found that 4DCT planning always results in a larger EUD to normal liver when compared with dose from a 3DCT plan. The mean EUD difference between 4D and 3D planning is 3.8Gy ([sigma]= 1.9Gy, p<0.000 1). Gated 4D treatment planning results in a lower EUD to normal liver compared to ungated planning, with a mean difference of 2.9 Gy ([sigma]=1.9Gy, p<0.0001). The EUD difference is only weakly correlated with the magnitude of the superior-inferior (S-I) tumor motion ([tau]=0.59 for 4D/3D, [tau]=0.48 for ungated/gated). The [Delta]EUD correlation with clinical target volume (CTV) (as fraction of liver volume) is much weaker ([tau]-0.31 for 4D/3D, [tau]=0.26 for ungated/gated). There was no evidence that the tumor position within the liver influenced the [Delta]EUD. This study suggests that physicians should consider 4D treatment planning if the risk of normal tissue complications is high. Normal tissues may also be significantly spared by gated treatment as a motion management strategy.