The aim of this work was evaluating the diagnostic accuracy of PET-MRI in difficult cases of differentiating between tumor progression and radionecrosis in neuro-oncology. For each lesion, PET (SUVmax, SUV mean, SUVpeak) and MRI (ADC, CBV, CBF, pCASL CBF) biomarkers were extracted. The combination of PET and MRI biomarkers allowed to improve the diagnostic accuracy. The logistic regression model has shown that 94% cases were correctly classified using the combination of SUVpeak and pCASL rCBF. Excellent diagnostic accuracy was achieved for both qualitative and quantitative evaluation by means of combined analysis of morphological, functional and metabolic imaging markers.