An objective method for combining multi-parametric MRI datasets to characterize malignant tumors.

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Medical imaging is crucial for characterizing malignant tumors.
• Complete tumor mapping often requires integrating data from multiple imaging modalities.
• Existing methods may lack objectivity in combining diverse imaging datasets.

Purpose of the Study:

• To propose and validate an objective method, the percent overlap method (POM), for combining multiple imaging datasets.
• To enhance the characterization of malignant tumors by creating a single composite map.
• To demonstrate the utility of POM in analyzing recurrent glioblastoma multiforme.

Main Methods:

• Developed the percent overlap method (POM) for objective integration of multimodal imaging data.
• Acquired seven parameter maps from magnetic resonance imaging (MRI) for four patients with recurrent glioblastoma multiforme.
• Included data from chemical shift imaging, perfusion scans, diffusion, and hypoxia mapping.

Main Results:

• Generated composite maps for each patient using the POM.
• Compared POM-derived maps with results from the ISODATA clustering technique.
• Found POM maps of likely recurrent tumor regions to be consistent with ISODATA clustering outcomes.

Conclusions:

• The percent overlap method (POM) provides an objective approach to combine parameters from multiple physiologic imaging techniques into a single composite map.
• The accuracy of POM maps is dependent on the sensitivity of the chosen imaging parameters to the disease.
• Further validation of POM through correlation with histological data is recommended.