Advances in Radiologic Image Analysis From MICCAI 2006.

Medical Image Computing and Computer-Assisted Intervention (MICCAI) is the premier international conference for research in medical image processing and analysis, computer assisted intervention, and medical robotics. MICCAI brings together engineers, clinicians, computer scientists, biologists, physicists, and other researchers and offers them a forum to exchange ideas and present in-depth papers in these exciting and rapidly growing interdisciplinary fields. MIC-CAI is a single-track conference and has been held in North America, Europe, and Asia. Selected papers from MICCAI 2003 (1–5) and MICCAI 2004 (6 –12) were published in previous issues of Academic Radiology and were very well received by the readers. A selection of papers derived from work first presented at the MICCAI 2005 conference is contained in this special issue of Academic Radiology. Of 632 submissions to MICCAI 2005, 7% were accepted for oral presentation. The authors of 11 of these papers, which have both clinical and technical significance, were invited to submit extended and enhanced versions of their papers for this special issue. Of those, eight papers were finally accepted for publication in this issue, after a standard peer-review process. The papers here cover an array of topics of interest to Academic Radiology readers, aimed at improving image quality and quantification using the most advanced MICCAI technologies, including registration, acquisition, segmenta-tion, detection, motion, and morphometry. Additional articles from the conference have also been selected, based on their relevance in the respective fields, for special issues of Medical Image Analysis and Computer-Aided Surgery. The methodologies discussed in these papers give an indication of how far these techniques have progressed and some hints as to where they are going. Image analysis techniques have begun to infiltrate beyond their original spheres. For example, image registration techniques are moving beyond the basic development stage and have begun to be le-veraged into improving acquisition (13), facilitating deformation analysis (14), enhancing tracer kinetic model fits (15), and simultaneous segmentation (16). There is also work integrating image analysis into image acquisition (13, 17), improving computer-aided detection strategies (18), analyzing the sources of error in morphometry (19) and modeling cardiac motion (20). Three-dimensional fetal magnetic resonance imaging (MRI) is extremely difficult due to motion during image acquisition. Motion corrupts the three-dimensional positions and orientations of individual slices during the acquisition. Rousseau et al. (13) present a registration and intensity correction scheme in order to correct for this motion between two-dimensional slice acquisitions in MRI. By iteratively determining …