Quantitative Imaging Techniques to Guide Clinical Decision Making and Treatment of Glioblastoma
Author(s)
Ramesh, Karthik
Advisor(s)
Shim, Hyunsuk
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Treatment of glioblastoma (GBM), a grade IV brain tumor, is complex and challenging. Despite an aggressive treatment protocol of surgery, radiation treatment, and chemotherapy, patient outcomes are poor. Throughout a patient’s treatment timeline, imaging is central to clinical decision making. The gold standard imaging for treatment of GBM is magnetic resonance imaging (MRI). While clinical MRIs show some extent of tumor infiltration, they do not capture the entire extent and further, can struggle to define tumor margins. Spectroscopic MRI (sMRI) measures the endogenous metabolites in the brain and in particular, a ratio of two metabolites (choline and N-acetyl aspartate) shows tumor margins that extend well past what’s seen in the clinic. To that end, we first present our efforts to guide GBM treatment with sMRI. Next, we analyze whether pre-treatment insights from sMRI correlate with patient survival and outcomes. Finally, GBM patients are tracked with clinical imaging after completing surgery and chemoradiation. To improve objectivity and robustness in post-treatment patient tracking, we have developed deep learning algorithms to quantify patient lesions and provide assistive suggestions to physicians for whether tumor recurrence has occurred. By utilizing sMRI during treatment and the latest deep learning algorithms after treatment, we have improved GBM patient survival and provided insights that can help physicians treat patients more proactively. Further, all our algorithms are housed in software that fit seamlessly into clinician workflows as the integration of quantitative tools and imaging into the clinic is vital towards affecting patient lives directly. Future plans involve translating sMRI and our quantitative software into the clinic to assist clinicians and improve patient livelihood.
Sponsor
Date
2024-04-17
Extent
Resource Type
Text
Resource Subtype
Dissertation