Abstract
303
Objectives 90Y microspheres are increasingly used for the treatment of inoperable hepatocellular carcinoma (HCC). Currently, the confirmation of 90Y biodistribution after treatment is assessed indirectly by Bremsstrahlung-SPECT, which is known to have poor statistical counting and inaccurate quantification of actual delivered dose to the tumor (TD) and nontumorous liver (NTD). TD and NTD are important prognostic and radiobiological factors in HCC patients; therefore, we aim to evaluate the efficacy of 90Y PET/CT post radioembolization for dosimetry measurement and for quantitative optimization of injected activity (IA) of 90Y glass microspheres individualized to HCC patients.
Methods From April 2014 to Oct 2015, 34 patients (M: 28, F: 6; age range: 34~82y, mean=60.1±13.9y; HBV: 24, HCV: 1) treated by 90Y glass microspheres for intermediate or advanced stage HCC were recruited. All patients underwent pretreatment 99mTc-MAA planar and SPECT/CT to estimate the lung shunting (S: % of IA) and tumor-to-nontumorous liver ratio (TNR), respectively. The dose to the injected liver (ILD) was determined by adjustment of the fixed radiation dose (120Gy recommended by pharmaceutical company) according to patient’s tumor extent, liver size and liver function. The IA was then calculated by a simplified MIRD equation: IA(GBq)= ILD(Gy)/[(1-S)×49.67/W(kg)] (W=ILvolume×1.03, where 1.03kg/L is the liver density). Regional 90Y PET/CT (Biograph mCT) was performed 17~22 hours post radioembolization with 15-min acquisition/bed and “time-of-flight” reconstruction (2 iterations, 21 subsets, 10-mm at FWHM, matrix size 200×200). For each patient, right & left nontumorous liver (NL), lung and all HCC lesions were contoured by automatic algorithms on PET/CT with mean activity concentration (Bq/ml) and volume measured for each VOI. They are then corrected for the 90Y positron branching ratio (3.186×10−5) and time decay factors to obtain the actual PET-measured activities (PA) in tumor, NL and lung. The TD, NTD and lung dose were calculated by D(Gy)=PA(GBq)×49.67/W(kg). Partial volume correction was not considered due to voluminous lesions (>6cm) in all patients. The accuracy of 90Y PET/CT for dosimetry measurement was checked by comparing the summation of PA in tumor, NL and lung with the injected activity.
Results 33/34 patients received reduced IA with ILD ranging from 40 to 100Gy (median: 60Gy). The IA was 2.45±1.11GBq. 90Y PET/CT post radioembolization showed high TD (>120Gy) in 30/34 patients (mean: 215±72Gy, range: 139~363Gy) and 73~117Gy in the other 4 patients. The NTD in all patients was less than the limiting dose of 70Gy recommended by the literature (mean: 37±11Gy, range: 16~55Gy). The lung shunting on PET/CT was minimal in 33/34 patients (mean: 1.8±1.2%, range: 0.4~4.6%) and mild in 1 patient (14.6%), lower than that on 99mTc-MAA planar (mean: 6.8±3.8%, range: 2.9~15.4%) in all patients. The summation of PA in tumor, NL and lung was comparable to the IA (91.2~99.3% with a median of 96.4% of IA), indicating that quantification by 90Y PET/CT for dosimetry measurement is feasible and accurate. The TNR on pretreatmet 99mTc-MAA SPECT/CT strongly correlated with 90Y PET/CT post radioembolization (mean=5.4±3.0 vs 5.6±3.6, linear curve: TNRY90=1.044×TNRMAA-0.033, P<<0.05).
Conclusions PET/CT post 90Y-radioembolization could provide accurate dosimetry measurement and confirmation of biodistribution in HCC patients, thus obviating the conventional use of Bremsstrahlung-SPECT. The TNR predicted by 99mTc-MAA SPECT/CT was accurate but lung shunting in 90Y treatment was overestimated by 99mTc-MAA planar imaging, therefore, the threshold limit might possibly be given greater allowance. Quantitative optimization of IA of 90Y glass microspheres could be individualized and guided by target TD or limiting NTD instead of a routinely fixed ILD.