**1. Introduction**

An important challenge facing the neuro-oncological treatment of gliomas is discriminating between tumor recurrence and treatment-related effects using non-invasive diagnostic imaging [1]. Not only do tissue types appear similar on standard magnetic resonance imaging (MRI), but new lesions are often a composite of tumor cells, gliosis, necrosis, inflammatory cells, and neovascularity, which confounds characterization [2]. Moreover, targeted therapies like bevacizumab complicate follow-up imaging even further by modifying vascular endothelial growth factor (VEGF), often causing a "pseudoresponse" with vascular changes resulting in a subsequent decrease in contrast enhancement [3]. Similarly, changes related to radiation or immunotherapy can mimic tumor progression, including changes in T1-weighted (T1w) contrast enhancement and T2-weighted (T2w) hyperintensity, once again complicating imaging-based tissue discrimination [4]. Etiological characterization of lesions observed on longitudinal follow-up scans factors into the clinical decision-making in the course of treatment and prognostic decisions.

While histopathology remains the gold standard for tissue type identification, it is not without its problems, such as the need for additional surgery, sampling bias, and risks of neurological complication [2,5]. Thus, a non-invasive method capable of distinguishing recurrence from treatment effects must be established in order to reduce the dependency on biopsy and improve the efficacy of patient follow-up with noninvasive imaging. Advanced MRI methods such as magnetic resonance (MR) spectroscopy, MR perfusion, positron emission tomography (PET), single photon emission CT (SPECT), diffusion weighted imaging (DWI), and diffusion tensor imaging (DTI) have been used to explore the feasibility of differentiating tumor recurrence and treatment effects with varying success [6–11]. PET-based methods, which measure glucose metabolism, demonstrate some ability in distinguishing glioma recurrence from radiation-induced necrosis. For example, increased fludeoxyglucose (FDG) tracer activity, corresponding to enhanced uptake on post-contrast T1 imaging, is consistent with tumor recurrence, while decreased FDG tracer activity is less specific, typically denoting vasogenic edema, stemming from recurrence and treatment effects [12,13]. Amino acid transport PET-based imaging, especially the use of tyrosine or tryptophan-based tracers, has also been studied to improve the ability to distinguish tumor recurrence from treatment-related changes. O-(2-[18F]fluoroethyl)-L-tyrosine (FET) has been studied since the 1990s and is believed to be more specific for tumor recurrence given the enhanced uptake of glucose in all brain (FDG) versus less amino acids uptake [14]. This should make FET PET more specific than FDG, and there have been a number of cases showing increased uptake of FET in tumors, and it may also be useful at assessing pseudoprogression from true recurrence in glioma [15–20]. However, several other tissue types can also have increased uptake, including brain abscesses, demyelinating processes, epilepsy, and in tissue adjacent to cerebral ischemia or hematomas, making some interpretation of results challenging [14,15,17]. MR perfusion techniques, like dynamic contrast-enhanced (DCE) MRI and dynamic susceptibility contrast (DSC) MRI, yield estimates of relative cerebral blood volume (rCBV) and vascular permeability (ktrans), reflecting underlying microvasculature and angiogenesis [21–23]. Studies have indicated MR perfusion's utility in differentiating tumor progression from treatment effects and pseudoprogression [24–26]. However, these techniques are hindered by mixed results [27], model complexity [28], and sensitivity to thresholds [29]. MR spectroscopy, estimating biomarkers like lactate and choline to creatinine ratios, has demonstrated higher diagnostic accuracy than conventional MRI in detecting tumor progression as well, reaching a sensitivity and specificity as high as 91% and 95%, respectively [30]. The diffusion metrics fractional anisotropy (FA) and mean diffusivity (MD) have been useful in differentiating between tissues types as well [31–33]. Recent research on glioblastoma demonstrated that MD can help differentiate between tumor recurrence and radiation-induced necrosis, as it is known that more free water lies within necrotic tissue than enhancing solid tumor [34]. Also, Apparent Diffusion Coefficient (ADC) ratios and mean ADC of tumor recurrence are significantly lower than those of radionecrosis, since higher cellularity (tumor recurrence) contributes to more restricted diffusion [35]. Verma et al. (2013) suggests the combination of low ADC values and high FA values help define the presence of tumor recurrence [2].

High grade gliomas, the most prevalent intracranial neoplasm, are highly heterogenous in the lesion area, have an invasive nature, and often require additional multimodality treatment later in the course of the disease. For these reasons, noninvasive diagnosis, monitoring, and prognosis strategies, such as MRI, must be sought and refined. With the goal to improve the noninvasive diagnostic utility of advanced MRI for gliomas, we studied a group of patients who had imaging localized histopathology. Through the combination of both conventional and advanced MRI modalities, we demonstrate improved efficacy in diagnosing recurrent tumor versus imaging effects related to treatment. These results demonstrate the potential for refining multi-modal MRI assessment of glioma tissue classification, thereby facilitating the clinical decision-making process.
