Patient dosimetry for [Formula: see text] selective internal radiation treatment based on [Formula: see text] PET imaging

Until recently, the radiation dose to patients undergoing the [Formula: see text] selective internal radiation treatment (SIRT) procedure is determined by applying the partition model to [Formula: see text] MAA pretreatment scan. There can be great uncertainty in radiation dose calculated from this approach and we presented a method to compute the 3D dose distributions resulting from [Formula: see text] SIRT based on [Formula: see text] positron emission tomography (PET) imaging. Five [Formula: see text] SIRT treatments were retrospectively analyzed. After [Formula: see text] SIRT, patients had [Formula: see text] PET/CT imaging within 6 hours of the procedure. To obtain the 3D dose distribution of the patients, their respective [Formula: see text] PET images were convolved with a Monte Carlo generated voxel dose kernel. The sensitivity of the PET/CT scanner for [Formula: see text] was determined through phantom studies. The 3D dose distributions were then presented in DICOM RT dose format. By applying the linear quadratic model to the dose data, we derived the biologically effective dose and dose equivalent to 2 Gy/fraction delivery, taking into account the spatial and temporal dose rate variations specific for SIRT. Based on this data, we intend to infer tumor control probability and risk of radiation induced liver injury from SIRT by comparison with established dose limits. For the five cases, the mean dose to target ranged from [Formula: see text] to [Formula: see text]. Due to the inhomogeneous nature of the dose distribution, the GTVs were not covered adequately, leading to very low values of tumor control probability. The mean dose to the normal liver ranged from [Formula: see text] to [Formula: see text]. According to QUANTEC recommendation, a patient with primary liver cancer and a patient with metastatic liver cancer has more than 5% risk of radiotherapy‐induced liver disease (RILD). PACS number: 87.53.Bn