Share

Researchers at the University of Connecticut Health Center (UCHC) are exploring applications of Maple in the interpretation of magnetic resonance imaging (MRI) information for the early detection and characterization of small tumors.

Vascular networks required for nurturing growth in tumor microenvironments can be much denser than in normal tissue. Vascular density is an important parameter in assessing tumor activity but it is universally appreciated that accurate determination is a difficult task. Determination of this parameter is based on manual counting procedures and is observer subjective. At best, one has to be satisfied with relative measurements.

Moreover, these measurements are very labor intensive and time consuming. For example, quantification along a major axis of a 1-cm tumor using several histological sections can take up to 200 hours. This experience prompted the UCHC team to develop a box-counting routine to interrogate digitized microscope images of well-resolved vascular networks.

A Maple program generates a binary map of the microscope images adjusted by selected thresholds. A second algorithm creates an adjustable-size square box that peruses the entire binary image and tests each pixel site for a value (1 or 0) to provide the basis for counting intersections (1's) and non-intersecting spaces (0's). In the small limit of s, a relative measure of the vascular density can thus be achieved in a matter of seconds. The measurements are used to validate the MR images.

The UCHC team is demonstrating new ways that MRI can be used as a superior way to detect and characterize small (2-4 mm), rapidly growing tumors. This research will enable doctors to use non-invasive MRI to identify and characterize human cancer tumors in their early stages of development and quantitatively follow the course of therapies. To date preliminary results have been done on mice. Clinical trials with human patients are anticipated in the near future.