Three dimensional visualization of thoracic tumors from MRI images

Aim: In cooperation with the thoraxsurgical department 3-D visualizations of tumors were generated to support the surgeons in preoperative planning. As opposed to 3-D reconstructions of CT data, those representations based on MRI images remain an exception. In this paper different methods of three dimensional visualization of lung tumors are presented and compared to each other. Design: These methods are 1.contour based surface models, 2.threshold based surface models and 3.rendered scenes of segmented volume data with transparent, color-coded display. Results: Combinations of all three methods in one single image are possible as well with the software we developed. Furthermore, cut planes through the original data can be integrated into the 3-D scene. The segmentation of anatomical objects is performed either manually or automatically with various procedures. Conclusions: Adequate possibilities of manipulation and archiving allow a fast handling and processing even of large data sets. The calculated volumes of the anatomical objects can be used for quantitative studies. Arbitrary views of the three dimensional reconstructions can be generated within a few seconds. Even animations can easily be calculated and displayed with 30 frames/secend.