Image Registration and Segmentation in Radiation Therapy

Computed tomography (CT) has become the standard for establishing a primary model of the patient for treatment planning. CT imaging provides a geometrically accurate and robust static representation of the patient’s anatomy and estimates the distribution of tissue density, which is an important radiological property for computing dose distributions [1-4]. Classifying tissues by density alone is not always sucient and other imaging modalities provide complimentary information to enhance the delineation of so tissues. In some cases, tissue characterization includes assessments of organ movement or the distribution of physiological and metabolic activity within an organ. For example, respiratory-correlated CT imaging, or "4DCT,” has also emerged as an innovative means of characterizing organ movement related to breathing and estimating its contribution for determining planning target volume (PTV) margins [5-8]. Other imaging modalities, such as contrast-enhanced CT, magnetic resonance imaging (MR), and positron-emission tomography (PET), are used increasingly to augment the collective understanding of the disease and surrounding vasculature, organ motion, and normal organ function [9-17]. MRI and PET oer a wealth of options (higher eld strength, pulse sequences, novel agents) to aid in the discrimination of tumors and normal tissues for radiation therapy.