Patient anatomy, biochemical response, as well functional evaluation at organ level, are key fields
that produce a significant amount of multi modal information during medical diagnosis. Visualization,
processing, and storage of the acquired data sets are essential tasks in everyday medical practice. In
order to perform complex processing that involves or rely on image data a robust as well versatile data
structure was used as extension of the Visualization Toolkit (VTK). The proposed structure serves as a
universal registration container for acquired information and post processed resulted data. The structure
is a dynamic multidimensional data holder to host several modalities and/or Meta data like fused image
sets, extracted features (volumetric, surfaces, edges) providing a universal coordinate system used for
calculations and geometric processes. A case study of Treatment Planning System (TPS) in the stereotactic radiotherapy (RT) based on the proposed structure is discussed as an efficient medical application.
TopIntroduction
Computer aided medical applications for diagnosis, therapy, simulation or training proliferate gradually in everyday practice heavily relying on image data (Dawson and Kaufmann, 1998; Spitzer & Whitlock, 1998). Radio therapy planning, surgery as well medical simulation requires anatomic and physical modeling of the whole or part of the human body. Further on in silico functional study of the human body physiology requires the interoperation of several models to approximate an as real as possible behavior (Noble, 2002; Gavaghan, Garny, Maini, & Kohl, 2006; Seemann, Hoeper, Doessel, Holden, & Zhang, 2006). Major sources of real world data related to anatomical details are tomographic image sets. Tomographic image sets of 3D solid objects are in general stacked 2D cross-sectional images of the inspected object that contain geometric information mixed with material properties of the solid in terms of radiation absorbance.
The modality used to obtain the tomographic set determines the geometric accuracy of anatomical regions, the resolution of material properties as well functional characteristics. The almost annual doubling in computer power permits today real time manipulation of simultaneous multimodal datasets representation also know as image fusion and 3D image registration. Sometimes modalities act complementary in cases of sparse acquired datasets (Shim, Pitto, Streicher, Hunter P. J. & Anderson I. A. 2007). In order to study physiological function of internal organs several models have been proposed ranging from simple calculation diagrams to complex animated 3D solid models (France et al. 2005; Noble 2002; Seemann, 2007; Selberg & Vanderploeg, 1994; Spirka & Damasa, 2007). Almost all models involve dynamics and geometry. Physiological functions rule dynamics, dynamics produce data and data become finally visualized (Freudenberg, Schiemann, Tiede, & Hoehne, 2000). Soft tissue simulation used extensively in computer assisted surgery planning or training is an example. Tissue is modeled as a deformable object while collision detection between the virtual surgery instruments or even neighbor organs is used. Deformation is modeled according to physiological data while collision detection queries the geometric models to undertake the desired action for example simulated tissue ablation, rapture (Nealen et al. 2005; Teschner et al., 2005).