Software automatically contours tumor volumes and activity

For six years, MIMVista has been all about software that displays, registers, and fuses medical images. The focus has been on the development of practical tools, ones that have real-life consequences. Now the Cleveland-based company is ratcheting those basic tools up a notch, automating the steps needed to contour and segment structures to extract as much detail as quickly and easily as possible for cancer treatment planning.

MIM 4.0, shown at the 2006 RSNA meeting, was designed with features that take some of the guesswork out of using data from fused PET/CT or PET/MR imaging studies for planning radiotherapy, said Adam Neff, applications coordinator.

A technique known as absolute threshold provides contours based on standardized uptake values. After a physician sets a baseline SUV, the software contours every voxel that has the same or a higher value, applying the contour to any type of image fusion - ones combining PET and CT, MR and PET, or even an old CT with a new one - Neff said.

A spatial derivative technique called autocontour outlines the boundary of a lesion, defining the volume not only for solid tumors but also for those with a necrotic center or with a fingerlike configuration. Autocontour has 2D and 3D editing options, so physicians can edit on sagittal and coronal planes and have their editing apply to subsequent axial slices. This is a time-saving feature for dosimetrists who have had to go through and edit each axial slide separately when using traditional radiation planning systems, Neff said.

A 2D constant threshold technique is also a time saver.

"With feedback from dosimetrists, we have heard that it usually takes around 20 minutes to contour a lung by going through each axial slice," Neff said. "With the constant threshold technique, we can do it for any structure in the body in about 10 to 20 seconds."