Software and hardware fusion with respect to PET/CT and PET/MRI - Research Paper Example

In addition to providing a relatively little information about the viability or metabolic activity of organs and lesions, the image interpretation in case of CT and MRI, can become complicated when normal and pathologic tissues with similar densities are considered or when normal perfusion patterns get disrupted by prior surgery or radiotherapy, thus resulting in tissue damage or necrosis where contrast patterns resemble those associated with neoplasia (eg. the detection of metastases in normal-size lymph nodes or isodense metastases in a deformed liver after surgery or liver cirrhosis) (Zaidi, Montandon, Alavi, 2008; Vogel, Oyen, Barentsz, Kaanders, Corstens, 2004).

Apart from, lacking sufficient sensitivity and specificity, these techniques pose a challenge when defining the anatomic extent of disease, in case of planning highly conformal radiation treatment or highly targeted therapeutic regimens (Zaidi et al., 2008; Vogel et al., 2004). Functional imaging with 18-fluorodeoxyglucose, (18F-FDG) PET, despite a high detection sensitivity reaching below the picomolar range, produces a relatively low spatial resolution of images (range: 4–6 mm and physically limited to about 2 mm) along with poor recognition capacity and delineation of anatomic structures, which may lead to uncertainty or failure in correctly localizing detected abnormalities.

A few examples include lesions in the upper abdomen, situated near the border of organs, or between adjacent organs (Vogel et al., 2004). In order to address these shortcomings of stand-alone CT and PET imaging modalities, whilst taking advantage of the less redundant and highly complimentary information from different imaging techniques, the present technology explores software- and/ or hardware-based correlative functional-anatomic imaging combining PET with CT or MR imaging modalities (PET/CT and PET/MRI) (Judenhofer et al., 2008). In view of the above and the emerging data, the present paper discusses software and hardware fusion with respect to the two dual-modality imaging technologies (PET/CT and PET/MRI), their role in a clinical setting, possibilities, limitations and technological obstacles.

Software-based image registration and fusion with respect to PET/CT and PET/MRI Software-based image registration and fusion techniques basically involve the integration of two or more image sets from two or more different imaging modalities following their separate acquisition on stand-alone imaging systems. When a colorful image of functional feature is superimposed over a gray-scale CT or MR image of the corresponding anatomic region the resulting fused image has a small rotation along 3D axes and displacement, which could be corrected manually (Vogel et al.

, 2004, Zaidi et al., 2008). A study by Vogel et al. (2004), focusing on PET/CT commented that although numerous algorithms including, count difference, shape difference, mutual information, normalized mutual information, have been developed, the resultant fused images, vary in accuracy (p. 18S). Software-based image registration techniques employ two approaches namely, rigid and non-rigid registration of images. Research by Slomka & Baum (2009) highlights the accuracy of rigid image registration of brain PET with CT and MRI in case of software techniques.

Non-rigid registration approaches are usually combined with linear registration tech