Combined amino acid PET-MRI for identifying recurrence in post-treatment gliomas: together we grow

ABSTRACT: PURPOSE: The aim of this study is to compare the diagnostic accuracy of amino acid PET, MR perfusion and diffusion as stand-alone modalities and in combination in identifying recurrence in post-treatment gliomas and to qualitatively assess spatial concordance between the three modalities using simultaneous PET-MR acquisition. METHODS: A retrospective review of 48 cases of post-treatment gliomas who underwent simultaneous PET-MRI using C11 methionine as radiotracer was performed. MR perfusion and diffusion sequences were acquired during the PET study. The following parameters were obtained: TBR(max), TBR(mean), SUV(max), and SUV(mean) from the PET images; rCBV from perfusion; and ADC(mean) and ADC(ratio) from the diffusion images. The final diagnosis was based on clinical/imaging follow-up and histopathology when available. ROC curve analysis in combination with logistic regression analysis was used to compare the diagnostic performance. Spatial concordance between modalities was graded as 0, 1, and 2 representing discordance, < 50% and > 50% concordance respectively. RESULTS: There were 35 cases of recurrence and 13 cases of post-treatment changes without recurrence. The highest area under curve (AUC) was obtained for TBR(max) followed by rCBV and ADC(ratio). The AUC increased significantly with a combination of rCBV and TBR(max). Amino acid PET showed the highest diagnostic accuracy and maximum agreement with the final diagnosis. There was discordance between ADC and PET in 22.9%, between rCBV and PET in 16.7% and between PET and contrast enhancement in 14.6% cases. CONCLUSION: Amino acid PET had the highest diagnostic accuracy in identifying recurrence in post-treatment gliomas. Combination of PET with MRI further increased the AUC thus improving the diagnostic performance.