DESCRIPTION (provided by applicant): Neo-adjuvant chemotherapy with temozolomide, along with radiation therapy has become the standard of care for the treatment of gliomas as this regimen improves survival of newly diagnosed glioblastoma patients. However, in about a third of the patients, progressive and enhancing lesions on MRI occur within six months after end of chemo-radiation therapy, which resolve without further interventions. These lesions are termed as pseudo-progression (PsP) as they mimic the appearance of a tumor on standard clinical MRI. Patients with PsP have a longer survival indicating that the progressive lesions were due to treatment effect rather than due to true tumor progression (TP). Accurate identification of PsP is critical for patient management as PsP patients can avoid unnecessary repeat surgery and continue effective chemotherapy, while alternative treatments, including resection or immunotherapy can be offered for TP. Functional MRI tools, such as, diffusion tensor imaging (DTI), perfusion imaging and magnetic resonance spectroscopy (MRS), probe changes in water mobility, blood flow and tissue metabolism, which are more sensitive in detecting changes than anatomical changes observed by standard MRI. Furthermore, in a clinical setting, it is unlikely that a single imaging parameter or modality will suffice in accurae detection of PsP. We thus hypothesize that the combination of whole brain MRS, perfusion imaging and DTI can provide a more accurate detection of PsP. In order to test this hypothesis, we will develop a novel decision support system that uses important features from all the imaging modalities in a completely non-biased way. Specifically, we will achieve the following aims: SA1: To assess the utility of 3D EPSI in differentiation of PsP patients from TP. SA2: To determine the role of dynamic susceptibility weighted contrast imaging in differentiation of PsP from TP. SA3: To evaluate the accuracy of DTI parameters, such as mean diffusivity (MD), fractional anisotropy (FA) and tensor shape measures including linear anisotropy coefficient (CL), planar anisotropy coefficient (CP) and spherical anisotropy coefficient (CS) in the identification of PsP. SA4: To determine the best combination of imaging parameters for accurate diagnosis of PsP. The utility of these parameters will be tested both at first clinical presentation (baseline) as well as change in these parameters on one-month follow up scan. A Bayesian network based decision support system will be used to detect PsP which will include advanced imaging parameters, as well as change in these parameters from baseline. The novelty of this approach is that it will provide a multi-dimensional analysis of brain tumors, whic will facilitate an accurate diagnosis of PsP and avoid unnecessary second-look surgery in these patients. |