Towards a generic respiratory motion model for 4D CT imaging of the thorax

One of the most important parameters reducing the sensitivity and specificity in the thoracic and abdominal areas is respiratory motion and associated deformations which represent today an important challenge in medical imaging. The negative impact of respiratory motion is not only evident in the field of cancer diagnostic imaging but also in therapy applications, particularly in the field of radiation therapy. Modelling of the respiratory motion is therefore very important for the efficacy of both diagnostic and therapeutic imaging applications. All respiratory motion models presented to date are patient specific and require a 4D CT acquisition for every patient implying an associated increased dose which cannot be easily justified for all patients. The objective of this study is the description of a framework for the creation of a global generic model from 4D CT images of different patients. This global model is based on principal component analysis (PCA) and can be adapted to a given patient anatomy needing only one static CT image in combination with respiratory synchronised images of the patient surface. In this work 4D CT images of three patients were used to create the model which was tested by subsequently adapting it on the static CT images of two different patients. Our comparative results between the model derived and the corresponding acquired 4D CT images demonstrate the efficiency of the developed generic motion model.