90Y-PET for the assessment of microsphere biodistribution after selective internal radiotherapy.

Objectives: To demonstrate the feasibility of 90Y-PET imaging for biodistribution assessment after selective internal radiotherapy treatments with 90Y-microspheres, comparing the results with 99mTc-macroaggregated albumin (MAA) images obtained with single-photon emission computed tomography. Methods: Preliminary studies were performed with the aim of evaluating the imaging system spatial resolution and scanner sensitivity for detecting annihilation photons. Subsequently, microsphere distribution was evaluated in 10 patients who underwent liver selective internal radiotherapy treatment. 99mTc-MAA and 90Y-microsphere were simultaneously injected for immediate monitoring after treatment. For each patient, the metastases detected with 90Y-PET and 99mTc-MAA were assessed and compared with 18F-fluorodeoxyglucose-PET (18F-FDG-PET) obtained before treatment and used as an imaging benchmark procedure. The correlation between these techniques was thus investigated in terms of matching lesions. Lesions were considered true positive in the case of matching with 18F-FDG-PET. The sensitivity of both techniques was evaluated as the true-positive fraction of detected spots in the treated liver sectors. Results: With our experimental setup, a maximum scanner sensitivity of 0.577 and 0.077 cps/MBq was obtained for three-dimensional and two-dimensional acquisitions, respectively. A good correlation was obtained between images obtained before and after treatment, with 90Y-PET being by far the most accurate technique in detecting microsphere distribution and tumor nonhomogeneity areas. A sensitivity as high as 0.91 was obtained with 90Y-PET, whereas 99mTc-MAA imaging showed a SE of 0.75. Conclusion: 90Y-PET is a promising and reliable technique for microsphere biodistribution evaluation after liver selective internal radiotherapy treatment. Because of the better resolution and the possibility to perform computed tomography fusion, 90Y-PET images are more accurate than 99mTc-MAA single-photon emission computed tomography, which is now considered the gold standard for biodistribution assessment.