11 C-methionine- 18 F-FDG dual-PET-tracer-based target delineation of malignant glioma: evaluation of its geometrical and clinical features for planning radiation therapy.

OBJECTIVE: It is important to correctly and precisely define the target volume for radiotherapy (RT) of malignant glioma. 11 C-methionine (MET) positron emission tomography (PET) holds promise for detecting areas of glioma cell infiltration: the authors' previous research showed that the magnitude of disruption of MET and 18 F-fluorodeoxyglucose (FDG) uptake correlation (decoupling score [DS]) precisely reflects glioma cell invasion. The purpose of the present study was to analyze volumetric and geometrical properties of RT target delineation based on DS and compare them with those based on MRI.

METHODS: Twenty-five patients with a diagnosis of malignant glioma were included in this study. Three target volumes were compared: 1) contrast-enhancing core lesions identified by contrast-enhanced T1-weighted images (T1Gd), 2) high-intensity lesions on T2-weighted images, and 3) lesions showing high DS (DS ≥ 3; hDS). The geometrical differences of these target volumes were assessed by calculating the probabilities of overlap and one encompassing the other. The correlation of geometrical features of RT planning and recurrence patterns was further analyzed.

RESULTS: The analysis revealed that T1Gd with a 2.0-cm margin was able to cover the entire high DS area only in 6 (24%) patients, which indicates that microscopic invasion of glioma cells often extended more than 2.0 cm beyond a Gd-enhanced core lesion. Insufficient coverage of high DS regions with RT target volumes was suggested to be a risk for out-of-field recurrence. Higher coverage of hDS by T1Gd with a 2-cm margin (i.e., higher values of "[T1Gd + 2 cm]/hDS") had a trend to positively impact overall and progression-free survival. Cox regression analysis demonstrated that low coverage of hDS by T1Gd with a 2-cm margin was predictive of disease recurrence outside the Gd-enhanced core lesion, indicative of out-of-field reoccurrence.

CONCLUSIONS: The findings of this study indicate that MRI is inadequate for target delineation for RT in malignant glioma treatment. Expanding the treated margins substantially beyond the MRI-based target volume may reduce the risk of undertreatment, but it may also result in unnecessary irradiation of uninvolved regions. As MET/FDG PET-DS seems to provide more accurate information for target delineation than MRI in malignant glioma treatment, this method should be further evaluated on a larger scale.