Table 1.
Summary of the studies included in the review.
| Authors | Year | Study Design | Number of Patients | Tumor Histotype/ Glioma Grade | PET Scanner Type | MRI Technique | Main Findings |
|---|---|---|---|---|---|---|---|
| Diagnosis and Differential Diagnosis | |||||||
| Valentini et al. [20] | 2017 | R | 12 (48 biopsy specimens) |
GBM | PET/CT | DWI, DTI, DSC-PWI, MRSI |
Highest values of rCBV, Cho/Cr, Cho/NAA, proportional decrease of SUVmax with increasing distance from the CE region. At histological examination, the CE region showed maximum tumor histological malignancy and presented the maximum values of rCBV, Cho/Cr, Cho/NAA, LL and SUVmax. |
| Yamashita et al. [47] | 2016 | R | 50 | GBM = 33 PCNSL = 17 |
PET/CT | DWI, IVIM |
Significantly higher fmax (p < 0.001) and Dmin (p < 0.0001) and significantly lower SUVmax (p < 0.0005) in GBM than in PCNSL. |
| Nakajima et al. [38] | 2015 | R | 34 | GBM = 23 PCNSL = 11 |
PET/CT | DWI, DSC-PWI |
High SS (100%) and SP (73.9%) of 18F-FDG PET in differentiating GBM from PCNSL. Good accuracy of ADC5% and uncorr |
| Grading | |||||||
| Shaw et al. [44] | 2019 | R | 33 | 36 histology samples: II = 11 III = 17 IV = 4 metastases = 1 benign = 3 |
PET/CT | Gd MRI | Combination of PET and MRI imaging enhances AC in identifying high-grade regions of glioma. PET: SS = 59%, SP = 79%, PPV = 89%, NPV = 55%. MRI: SS = 77%, SP = 86%, PPV = 89%, NPV = 71%. Combined PET and MRI: SS = 79%, SP = 100%, PPV = 100%, NPV = 75%. |
| Sakata et al. [41] | 2018 | R | 49 | II = 15 III-IV = 34 |
PET/CT | DWI, APT | Comparable AC of T/N and ADCmin and amide proton transfer in the discrimination of HGGs from LGGs. A larger increase for the diagnosis of HGGs with the combination APT + T/N compared to ADCmin + T/N. |
| Takano et al. [46] | 2016 | R | 35 | II = 23 III = 12 |
PET/CT | DTI, DWI | No satisfactory performance for average fractional anisotropy, and maximum fractional anisotropy, minimum ADC, T/Nmax and T/Nave in discriminating III from II grade. |
| Song et al. [45] | 2016 | R | 70 | LGG and HGG | PET/CT | Gd MRI | 18F-FDG PET/CT performs better (in terms of SS, SP and AC) than MRI (p < 0.05) for identifying different grades of glioma. |
| Sacconi et al. [40] | 2016 | R | 20 | II = 6 III = 3 IV = 6 metastases = 2 meningioma = 2 lymphoma = 1 |
PET/MR | PWI | Utility of rCBVmean and SUVmean in discriminating HGGs from LGGS. rCBVmean (optimal cut-off value = 1.74): SS = 100%, SP = 74%. SUVmean, (optimal cut-off value = 4.0): SS = 50%, SP = 79.5%. |
| Prognosis | |||||||
| Lundemann et al. [37] | 2019 | P | 16 | GBM | PET/CT (18F-FET) PET/MR (18F-FDG) |
DWI, DCE |
18F-FDG and 18F-FET uptake demonstrate the highest mutual correlation in CELs and NELs, with 18F-FET being the most important to predict recurrence. Fractional anisotropy resulted in the second most important parameter for recurrence probability in apparently healthy tissue. |
| Chiang et al. [29] | 2017 | R | 44 | GBM | PET/CT | ADC | Metabolic tumor volume and tumor cross products on 18F-FDG PET and on MRI may serve as prognostic variables. Combining the cross products of both PET and MRI, the AC in predicting poor survival increased to 74% from 58% using MRI alone. |
| Leiva-Salinas et al. [36] | 2017 | R | 56 | GBM | PET/CT | Gd MRI | SUVr may be a useful imaging marker to identify patients’ decreased survival after standard therapy. SUVr was not influenced by tumor size and location on MRI images at diagnosis. |
| Assessment of Recurrence | |||||||
| Seligman et al. [42] | 2019 | R | 41 | III = 21 IV = 20 |
PET/MRI | DCE | 18F-FDG PET and DCE-MRI hold comparable AC (80% vs. 83%) in identifying tumor recurrence. |
| Hojjati et al. [32] | 2018 | R | 24 (28 lesions) | GBM | PET/MRI PET/CT |
DCE, DSC-PWI, DWI | The authors documented an AUC of 1.0 in a joint predictive model including r-mean ≥ 1.31 and a CBV ≥ 3.32. By contrast, a model encompassing only CBV ≥ 3.32 demonstrated a lower AUC (0.94). |
| Arora et al. [27] | 2018 | P | 29 | LGG = 15, HGG = 14 | PET/CT | Gd MRI | On per-patient analysis, no significance difference was found between the performance of 18F-FDG PET/CT and MRI (AC = 82.8% vs. 76.6%) in detecting recurrence. MRI did not detect significantly more lesions than 18F-FDG PET/CT (p = 0.14). |
| Jena et al. [35] | 2017 | R | 35 | II = 9 III = 13 IV = 19 |
PET/MR | DWI, PWI, MRS | PET provides complementary information to MRI. The AUC obtained combining MRI metrics (rCBV, mean ADC, Cho/Cr) and the PET parameter (mean T/N) was higher (0.935 ± 0.046) than the curve that resulted only from the three MRI parameters (0.913 ± 0.053). |
| Hatzoglou et al. [30] | 2016 | P | 29 | II = 7 III = 8 IV = 18 |
PET/CT | DCE | The combination of a plasma volume ratio ≥ 2.1 and a SUVratio ≥ 1.2 improve the performance in distinguishing progression from radiation injury compared to individual PET and DCE metrics. |
| Sharma et al. [43] | 2016 | R | 64 | Low-grade astrocytoma = 22 High-grade astrocytoma = 16 Medulloblastoma = 10 Other miscellaneous brain tumors = 6 |
PET/CT | NR | Good performance of PET and MRI in detecting recurrence in oligodendroglioma. In low-grade astrocytomas, a high rate of false positive cases (10/22 patients) were documented for PET. Nevertheless, PET was helpful in all cases reported as equivocal (n = 5) by MRI. |
| Iagaru et al. [33] | 2015 | P | 17 | GBM | PET/CT | Gd MRI | Similar diagnostic performance of the two modalities for recurrent GBM (13/15 detected recurrences for PET vs. 14/15 MR). |
| Treatment Planning and Evaluation of Response to Therapy | |||||||
| Idegushi et al. [34] | 2018 | P | 16 | II = 8 III = 8 |
PET/CT | Gd MRI, T2-w, FLAIR | 18F-FDG PET may also help in planning surgical resection. Only partial overlap between 18F-FDG uptake and the contrast-enhancement area. Tissue extracted from the 18F-FDG and Gd MRI positive areas presented anaplastic features. Tissue extracted from 18F-FDG and Gd MRI negative areas resulted in grade II glioma at pathological examination. |
| Hirata et al. [31] | 2019 | P | 25 | III = 10 IV = 15 |
PET | Gd MRI, T2-w | Tumor delineation is underestimated by Gd MRI. High overlap of DS and T1-Gd positively influenced survival. |
| Back et al. [28] | 2017 | P | 10 | III | PET/CT | T1-w, Gd MRI, T2-w | 18F-FDG PET guided integrated boost intensity-modulated RT (b-IMRT) that may result in a reduced dose to the normal brain when compared to standard IMRT (s-IMRT). |
| O’Neill et al. [39] | 2016 | P | 12 | III | PET/CT | DCE, DWI | The MRI-derived metrics (ADCmean, Ktrans, Ve) demonstrated significant variation in the patients (median difference of Ktrans = −41.8%, p < 0.02, median difference of Ve = −42.6%, p < 0.04), possibly reflecting the early effects of VEGF trap on tumour vasculature. No systematic changes were observed for SUVmax (median difference = −7.8%, p > 0.67). |
R: retrospective; P: prospective; N: number; GBM: glioblastoma multiforme; DWI: diffusion-weighted imaging; DSC-PWI: dynamic susceptibility-contrast perfusion-weighted imaging; MRSI: MR spectroscopic imaging; CBV: cerebral blood volume; Cho/Cr: Choline/Creatine; Cho/NAA: Choline/N-acetylaspartate; LL: Lipids/Lactate; IVIM: intravoxel incoherent motion; f: perfusion fraction; D: diffusion coefficient; SS: sensitivity; SP: specificity; AC: accuracy; ADC: apparent diffusion coefficient; Gd: gadolinium; DTI: diffusion tensor imaging; CE: contrast-enhancing; CEL: contrast-enhancing lesion; NE: non-enhancing; NEL: non-enhancing lesion; APT: amide proton transfer; T/N: tumor-to-normal tissue ratio; SUVr: standardized uptake value ratio (calculated as the SUVmax in the tumor relative to that in healthy white matter); AUC: area under the curve; r-mean: SUVmean of the lesion/ SUVmean of the contralateral background; FLAIR: Fluid Attenuated Inversion Recovery; T1-w: T1-weighted; T2-w: T2-weighted; DS: decoupling score (magnitude of the disrupted correlation of 11C-methionine and 18F-FDG, reflecting glioma cell invasion); Ktrans: transfer constant; Ve: extravascular extracellular volume fraction; VEGF: vascular endothelial growth factor; VEGF Trap: a soluble recombinant decoy receptor inactivating extravascular and circulating VEGF); NR: not reported.