3D voxel fusion of multi-modality medical images in a clinical treatment planning system

We present a novel 3D voxel fusion approach to achieve multi-modality image fusion in our clinical treatment planning system. This approach shows its superior visual representation and straightforward operation in the image fusion process and in verification of the fusion result. With RGB color bands to represent each of the fusing imaging modalities (voxel volumes), the homogeneity of the color distribution in a given tissue or anatomic structure has demonstrated to be a good indicator for high quality fusion. With a pre-defined lookup table library for different imaging technologies and different body segments, tissue of interest can be selectively viewed and individually fused, especially for those served as fusion landmarks. With control of voxel transparency with its geometrical location, an "onion pealing" effect can be established for further examination of the interior of the voxel volumes. It will be shown that the 3D voxel fusion can resolve the ambiguity resulted from 3D pixel fusion, because the 3D voxel fusion can distinguish a systematic positioning error from the random error present in the fusion result. It will also be shown that the 3D voxel fusion is straightforward, in the form of real-time "virtual reality". It can be applied for setting initial fusion position, as well as fine-tuning final fusion result with an improved accuracy (single voxel, or sub mm). Its simplicity and accuracy has made this technique a practical tool for multi-modality image fusion and for quality assurance. Additional clinical applications can be developed based on this technique.