- [NUCLEAR MEDICINE]
- [RADIOTHERAPEUTIC ONCOLOGY]
- [PATHOLOGICAL ANATOMY]
- [RADIOPHYSICS AND RADIOLOGICAL PROTECTION]
Quantitative volumetric analysis of gliomas with sequential MRI and (11)C-methionine PET assessment: patterns of integration in therapy planning
Arbizu J, Tejada S, Marti-Climent JM, Diez-Valle R, Prieto E, Quincoces G, Vigil C, Idoate MA, Zubieta JL, Peñuelas I, Richter JA.
The aim of the study was to evaluate the volumetric integration patterns of standard MRI and (11)C-methionine positron emission tomography (PET) images in the surgery planning of gliomas and their relationship to the histological grade.
We studied 23 patients with suspected or previously treated glioma who underwent preoperative (11)C-methionine PET because MRI was imprecise in defining the surgical target contour. Images were transferred to the treatment planning system, coregistered and fused (BrainLAB). Tumour delineation was performed by (11)C-methionine PET thresholding (vPET) and manual segmentation over MRI (vMRI). A 3-D volumetric study was conducted to evaluate the contribution of each modality to tumour target volume. All cases were surgically treated and histological classification was performed according to WHO grades. Additionally, several biopsy samples were taken according to the results derived either from PET or from MRI and analysed separately.
Fifteen patients had high-grade tumours [ten glioblastoma multiforme (GBM) and five anaplastic), whereas eight patients had low-grade tumours. Biopsies from areas with high (11)C-methionine uptake without correspondence in MRI showed tumour proliferation, including infiltrative zones, distinguishing them from dysplasia and radionecrosis. Two main PET/MRI integration patterns emerged after analysis of volumetric data: pattern vMRI-in-vPET (11/23) and pattern vPET-in-vMRI (9/23). Besides, a possible third pattern with differences in both directions (vMRI-diff-vPET) could also be observed (3/23). There was a statistically significant association between the tumour classification and integration patterns described above (p < 0.001, κ = 0.72). GBM was associated with pattern vMRI-in-vPET (9/10), low-grade with pattern vPET-in-vMRI (7/8) and anaplastic with pattern vMRI-diff-vPET (3/5).
The metabolically active tumour volume observed in (11)C-methionine PET differs from the volume of MRI by showing areas of infiltrative tumour and distinguishing from non-tumour lesions. Differences in (11)C-methionine PET/MRI integration patterns can be assigned to tumour grades according to the WHO classification. This finding may improve tumour delineation and therapy planning for gliomas.