The new software combines data from different medical imaging techniques (MRI, PET, and other scans) using a non-supervised statistical analysis to characterize various types of tumor tissue, providing oncologists and radiotherapists with valuable assistance diagnosing tumors and planning radiotherapy treatments.
The tumor and organ edges are automatically located and active areas of the tumor are distinguished from necrotic tumor tissue and zones of edema. “The software uses statistical analysis methods to perform a 3D segmentation of the tissue, giving each type of tissue its own unique signature,” explained a List researcher. “This gives practitioners a basis for identifying tissue. The practitioner can then either approve or correct the results as needed. This saves time and provides a more objective assessment than manually segmenting the tumor slice by slice.”
The software also delivers benefits for patient monitoring by providing quantitative data on tumor growth and identifying new tumor markers. For example, the software can be used to determine tissue signatures for the early detection of recidivism in glioblastoma patients, as well as for any other type of pathology.