Report of first recurrent glioma patients examined with PET-MRI prior to re-irradiation

BACKGROUND AND PURPOSE: The advantage of combined PET-MRI over sequential PET and MRI is the high spatial conformity and the absence of time delay between the examinations. The benefit of this technique for planning of re-irradiation (re-RT) treatment is unkown yet. Imaging data from a phase 1 trial of re-RT for recurrent glioma was analysed to assess whether planning target volumes and treatment margins in glioma re-RT can be adjusted by PET-MRI with rater independent PET based biological tumour volumes (BTVs). PATIENTS AND METHODS: Combined PET-MRI with the tracer O-(2-(18)F-fluoroethyl)-l-tyrosine ((18)F-FET) prior to re-RT was performed in recurrent glioma patients in a phase I trial. GTVs including all regions suspicious of tumour on contrast enhanced MRI were delineated by three experienced radiation oncologists and included into MRI based consensus GTVs ((MR)GTVs). BTVs were semi-automatically delineated with a fixed threshold of 1.6 x background activity. Corresponding BTVs and (MR)GTVs were fused into union volume (PET-MR)GTVs. The Sørensen–Dice coefficient and the conformity index were used to assess the geometric overlap of the BTVs with the (MR)GTVs. A recurrence pattern analysis was performed based on the original planning target volumes (PTVs = GTV + 10 mm margin or 5 mm in one case) and the (PET-MR)GTVs with margins of 10, 8, 5 and 3 mm. RESULTS: Seven recurrent glioma patients, who received PET-MRI prior to re-RT, were included into the present planning study. At the time of re-RT, patients were in median 54 years old and had a median Karnofsky Performance Status (KPS) score of 80. Median post-recurrence survival after the beginning of re-RT was 13 months. Concomitant bevacizumab therapy was applied in six patients and one patient received chemoradiation with temozolomide. Median GTV volumes of the three radiation oncologists were 35.0, 37.5 and 40.5 cubic centimeters (cc) and median (MR)GTV volume 41.8 cc. Median BTV volume was 36.6 cc and median (PET-MR)GTV volume 59.3 cc. The median Sørensen–Dice coefficient for the comparison between (MR)GTV and BTV was 0.61 and the median conformity index 0.44. Recurrence pattern analysis revealed two central, two in-field and one distant recurrence within both, the original PTV, as well as the (PET-MR)GTV with a reduced margin of 3 mm. CONCLUSION: PET-MRI provides radiation treatment planning imaging with high spatial and timely conformity for high-grade glioma patients treated with re-RT with potential advancements for target volume delineation. Prospective randomised trials are warranted to further investigate the treatment benefits of PET-MRI based re-RT planning.