Magnetic resonance imaging (MRI) plays an integral role in the diagnosis and monitoring of gliomas. One means by which MRI has been used to assess treatment efficacy has been measuring the volumes of contrast-enhancing lesions on post-contrast T1-weighted images. Clinically, an increase of the lesion volume by a certain percentage compared to a previous baseline scan warrants a change in treatment. However, this type of imaging has its limitations, as exemplified by false positive radiographic determination of tumor progression (“pseudoprogression”) and false positive radiographic determination of treatment response (“pseudoresponse”).

Given the vascular nature of gliomas, perfusion-weighted dynamic susceptibility contrast (DSC)-MRI has been studied in efforts to improve the detection, characterization, and monitoring of gliomas after treatment. However, applying DSC-MRI biomarkers is not necessarily straightforward. One of the biggest problems with the calculation of relative cerebral blood volume (rCBV) is that it is compromised by artifacts created by the extravasation of contrast agent from the vasculature. This can be a particular challenge in the neuro-oncology field since blood brain barrier disruption is a common feature of gliomas.

This work attempts to improve estimates of rCBV in gliomas by incorporating a two-compartment pharmacokinetic model into the indicator dilution theory, which we term the “bidirectional” leakage correction. In Chapter II, we use simulation methods to demonstrate improved accuracy gained by the bidirectional leakage correction, as compared to a current, popular leakage correction (“unidirectional” leakage correction). In Chapter III, we demonstrate that the bidirectional model-generated permeability curves have better correlation with DCE-MRI permeability curves than those generated by the unidirectional model. We also demonstrate that the rCBV is more similar for the bidirectional model between two separate pre-treatment scans from the same patient. In Chapter IV, we demonstrate that the change in bidirectional rCBV can stratify glioblastoma patients treated with bevacizumab according to long- or short-term survival. In all, the above works demonstrate that the new technique better combats leakage effects, thereby improving the clinical utility of rCBV for human gliomas.

Degenerative cervical myelopathy (DCM) is a chronic, progressive disorder characterized by the age-related degeneration of osseocartilaginous structures within the cervical spine resulting in narrowing of the spinal canal and chronic compression of the spinal cord. Chronic spinal cord compression can result in persisting neck pain and neurological deficits including loss of fine motor skills, weakness or numbness in the upper limbs, and gait abnormalities and imbalance, ultimately requiring surgical intervention to relieve cord compression. DCM is the most common form of spinal cord injury in adults and as the elderly population continues to grow, incidence of DCM will rise alongside an increased demand on healthcare resources. Further investigation into the neural response to chronic spinal cord compression may not only inform disease progression and prognosis but may benefit patient monitoring and treatment planning.This dissertation aims to elucidate how symptom presentation, degree of spinal compression, microstructural and cellular integrity of the affected cord, and sex impact supraspinal structure and function in patients with DCM. To address the goals of the dissertation, we implemented a multimodal neuroimaging approach including anatomical, functional, and diffusion imaging of the brain and T2-weighted, diffusion, and metabolic imaging of the spine. First, we characterized and compared spinal cord compression induced alterations in cerebral morphometry and functional connectivity between symptomatic DCM and asymptomatic spinal cord compression (ASCC) patients to further uncover potential compensatory neural mechanisms driving symptom presentation and disease progression. Because the degree of cervical cord compression is not strongly linked to symptom severity, we investigated whether macrostructural, microstructural, and metabolic properties of the cervical spinal cord result in conventional anatomical and functional alterations within the brains of patients with DCM. Lastly, we identified sex-specific differences on cerebral structure and functional connectivity in patients with DCM. In summary, the dissertation revealed unique cerebral signatures between symptomatic and asymptomatic patients, novel insights into the interrelationship between spinal and supraspinal alterations, and sex-specific supraspinal reorganization in patients with DCM. Findings from this work contribute to our knowledge of disease characteristics and compensatory neural mechanisms; and may benefit future development of non-invasive imaging biomarkers, more precise predicative models to inform disease progression, and novel pharmacological strategies to enhance neuroprotective mechanisms and functional recovery in patients with DCM.

Magnetic resonance imaging (MRI) is a valuable tool for the clinical management and scientific investigation of patients diagnosed with brain tumors. Continued efforts in brain tumor imaging research are critical because of the ever-evolving landscape of MRI technological advancements and new insights into brain tumor biology. For example, while historically, tumor tissue diagnoses relied on histopathological features, today, tumor molecular genetic features are the foundational component of tumor classification and patient management. Isocitrate dehydrogenase (IDH)-mutant gliomas are a specific brain tumor molecular subtype that particularly impact younger adults and remain incurable. Within IDH-mutant gliomas, there are also now molecularly defined IDH-mutant astrocytomas (1p/19q intact) and IDH-mutant oligodendrogliomas (1p/19q-codeleted). Although IDH-mutant gliomas often present as indolent, low-grade tumors, these tumors eventually become more aggressive in a process clinically described as “malignant transformation” into high-grade tumors, which are more resistant to therapy and have worse prognosis. This dissertation focuses on developing novel MRI tools to characterize molecular subtypes of IDH-mutant gliomas and malignant transformation. Specifically, this dissertation introduces four new tools for IDH-mutant glioma imaging research and clinical care: (1) optimal normal appearing white matter-normalization for diffusion and perfusion MRI analyses, (2) T2-FLAIR subtraction maps for quantitative T2-FLAIR mismatch analyses, (3) pseudo-resting-state functional MRI for functional connectivity analyses using DSC perfusion MRI (provisional patent filed), and (4) digital flipbooks of patient MRI scans to visually assess brain tumors. This dissertation utilized these tools to study IDH-mutant gliomas, particularly focusing on classifying glioma molecular subtypes, assessing cognitive impairment in patients, characterizing IDH inhibitor targeted therapy treatment response, and identifying IDH-mutant glioma tumor progression, including malignant transformation.

Cervical stenosis (CS), or tightening of the spinal canal, that causes long-term neurological impairment including neck pain and motor weakness, is termed cervical spondylotic myelopathy (CSM). However, substantial compression of the spinal cord does not necessarily entail more severe symptomatology and evaluations of clinical magnetic resonance imaging (MRI) scans are only weakly associated with patient symptom severity and have shown limited value in predicting patient disease progression or response to treatment. Thus, non-invasive imaging biomarkers that better reflect patient symptoms are in great need. While the spinal cord is the primary site of injury in CSM, the brain can also present with degenerative or compensatory changes in relation to patient symptoms. The purpose of this dissertation was to evaluate, via advanced MRI scans, the associations between structure, function, and microstructure, of the spinal cord and brain, with neurological symptoms and pain severity in patients with CS.

High-resolution structural, resting-state functional, and diffusion MRI scans of the brain and spinal cord were acquired in 26 CS patients and 17 healthy control subjects. Measures of brain and spinal cord structure, functional connectivity, and microstrucutre, were compared to the modified Japanese Orthopedic (mJOA) and the Neck Disability Index (NDI) scores. Similar to previous findings, structural measures of the spinal cord were not associated with symptom severity. However, patients with CS presented decreased brain structure with worsening neurological symptoms in subregions of the primary sensorimotor cortex, the superior frontal gyrus, and precuneus, and in the anterior cingulate and putamen in relation to both neurological and pain symptoms. These regions also demonstrated altered functional connectivity, presenting mostly with increasing connectivity with worsening symptoms. Diffusion MRI revealed altered microstructure consistent with degenerative changes in the spinal cord, both at the site of injury and at uncompressed sites, and further rostrally in the corticospinal tracts in the brain.

This cross-sectional study brings the brain and spinal cord together using advanced MRI to paint a more holistic picture of patients with CSM, and provides a groundwork for future studies to evaluate the brain and spinal cord longitudinally and after surgery to help determine optimal treatment strategies.

Magnetic resonance imaging is an integral part of medical diagnoses, treatment, and evaluation of patients with brain tumors. While standard anatomical imaging is useful, it does not provide information about molecular-level tumor characteristics that may spatially and temporally vary throughout the tumor. As such, there remains a need for the development of novel MRI techniques that can be used for evaluation of tumor growth and treatment response in patients with glioma undergoing radiochemotherapy. Extracellular acidosis is a hallmark of cancer and is intertwined with other common characteristics of the tumor microenvironment including hypoxia and angiogenesis. Therefore, the central objective of this dissertation was to develop a non-invasive imaging technique for identifying regions of acidosis within glioma and surrounding tissues using MRI. The most common types of glioma are highly aggressive and often require radiochemotherapy, which can result in variable responses across the patient population. Information about the acidity characteristics of these gliomas and the surrounding tissue may allow us to more accurately select targets for biopsy and radiation therapy, identify which patients are responding well to treatment, and predict prognosis.

Chemical exchange saturation transfer (CEST) MRI is a molecular imaging technique that generates contrast indirectly from protons on labile functional groups such as amines, amides, and hydroxyls. CEST image contrast is dependent on the exchange rate between bulk water protons and these functional groups, which is in turn dependent upon local pH. Because of this, we hypothesized that we could utilize CEST MRI for pH-weighted imaging in human tissues. By developing simulations of the Bloch-McConnell equations governing chemical exchange, we have shown that the CEST contrast generated by fast-exchanging amino acid amine protons increases with decreasing pH within a physiologically relevant range (6.0-7.4). We have also incorporated experimental scan parameters into these simulations to more accurately simulate the CEST contrast obtained during clinical data acquisition. Data were acquired in amino acid phantoms at varying pH and concentration, verifying our image contrast was dependent on pH. Our pH-weighted MRI sequence was also applied in animal models of glioma, providing evidence it can be used to generate unique contrast within tumors and can serve as a potential biomarker for response to treatment.

Our CEST MRI method was then applied serially in a cohort of glioblastoma patients undergoing treatment with standard radiochemotherapy, along with select cases of patients undergoing targeted biopsy. Results showed that tumor acidity characteristics were predictive of progression-free survival in the glioblastoma patient cohort. Acidity of targets selected for biopsy on pH-weighted images was indicative of tumor within those biopsy samples. To improve the imaging time of our sequence, we then upgraded the readout to utilize echo-planar imaging (EPI) rather than the standard gradient echo method. This allowed for whole brain coverage and multiple averages within a reduced scan time. The pH-weighted CEST-EPI sequence was applied in healthy volunteers and in a cohort of glioma patients prior to biopsy, in order to select targets for biopsy in regions of acidic and non-acidic tumor tissue. A subset of these patients also underwent PET imaging using 18F-FDOPA, an amino acid analog, near the time of their their pH-weighted scan. 18F-FDOPA uptake was shown to correlate quantitatively and qualitatively with regions of acidity, although pH-weighted imaging provided unique contrast in some cases. pH-weighted MRI was also acquired in recurrent glioblastoma patients before and after the start of treatment with bevacizumab. Acidity was shown to decrease after bevacizumab treatment, and in some cases acidic regions with no apparent contrast enhancement were shown to develop contrast enhancement on follow-up images, indicating that acidic lesions on pH-weighted MRI may be predictive of further tumor growth. Two additional advanced pH-weighted CEST MRI techniques were also implemented. CEST-EPI with a multi-echo readout was developed and acquired in a small cohort of glioma patients. The short and long echoes can provide sensitivity to more and less restricted water molecules, respectively. Separately, our CEST simulation incorporating T1 and T2 maps was used to quantitatively calculate estimates of pH in each image voxel for a subset of patients and animal models. This allows us to correct for T1 and T2 effects and generate numerical estimates of pH rather than pH-weighted images.

Together, these experiments and results present a comprehensive description of pH-weighted molecular MRI in gliomas. This technique has the potential to be implemented clinically for detection of acidosis in gliomas and other brain pathologies.

Purpose

Methods

Results

Conclusion

The reprogrammed energy metabolism and the dysfunctional vascular network of tumors create a hypoxic and acidic microenvironment, which is related to various malignant properties of cancer and poor patient prognosis. We have developed an amine chemical exchange saturation transfer (CEST) sequence with spin-and-gradient echo (SAGE) echo-planar imaging (EPI) readout to evaluate tumor acidity and hypoxia in human gliomas simultaneously. Amine CEST provided pH-sensitivity through labeling the endogenous amine protons that undergo chemical exchange with water protons, with a pH-dependent exchange rate. On the other hand, the reversible transverse relaxation rate quantified using the multi-echo EPI readout reflects oxygen extraction through sensitivity to paramagnetic deoxyhemoglobin.

This dissertation focused on developing and validating this novel dual-function imaging technique, mainly from three aspects: the technical development and validation, the biological validation, and the clinical validation of the proposed pH- and oxygen-sensitive CEST-SAGE-EPI technique in human gliomas. We have developed a new post-processing method for improved $B_0$ correction. A customized CEST physical phantom was designed and developed with validated temporal stability. We also evaluated the CEST contrast variability in healthy volunteers and the normal-appearing contralateral brain regions in glioma patients. The proposed pH- and oxygen-sensitive imaging biomarkers showed significant correlations with the tumor cell metabolomics features and MRI-guided biopsy tissue biomarkers, which validated the biological bases of the imaging biomarkers. Additionally, we have examined the association between tumor acidity with tumor vascularity, as measured by perfusion MRI. Lastly, we investigated the clinical usefulness of the biomarkers to characterize different glioma genotypes, predict patient prognosis, and monitor treatment responses.

In summary, this dissertation demonstrated that the novel dual-function pH- and oxygen-sensitive imaging technique reflects the abnormal metabolism in glioma patients and has the potential to provide clinical values for patient diagnosis, prognosis, and treatment efficacy assessment.