Abstract
PURPOSE: The aim of this paper was to investigate correlations between pre- therapeutic [(68)Ga]Ga-DOTA-TOC uptake and absorbed dose to tumours from therapy with [(177)Lu]Lu-DOTA-TATE. METHODS: This retrospective study included 301 tumours from 54 GEP-NET patients. The tumours were segmented on pre-therapeutic [(68)Ga]Ga-DOTA-TOC PET/CT, and post-therapy [(177)Lu]Lu-DOTA-TATE SPECT/CT images, using a fixed 40% threshold. The SPECT/CT images were used for absorbed dose calculations by assuming a linear build-up from time zero to day one, and mono-exponential wash-out after that. Both SUV(mean) and SUV(max) were measured from the PET images. A linear absorbed-dose prediction model was formed with SUV(mean) as the independent variable, and the accuracy was tested with a split 70-30 training-test set. RESULTS: Mean SUV(mean) and SUV(max) from [(68)Ga]Ga-DOTA-TOC PET was 24.0 (3.6-84.4) and 41.0 (6.7-146.5), and the mean absorbed dose from [(177)Lu]Lu-DOTA-TATE was 26.9 Gy (2.4-101.9). A linear relationship between SUV(mean) and [(177)Lu]Lu-DOTA-TATE activity concentration at 24 h post injection was found (R(2) = 0.44, p < 0.05). In the prediction model, a root mean squared error and a mean absolute error of 1.77 and 1.33 Gy/GBq, respectively, were found for the test set. CONCLUSIONS: There was a high inter- and intra-patient variability in tumour measurements, both for [(68)Ga]Ga-DOTA-TOC SUVs and absorbed doses from [(177)Lu]Lu-DOTA-TATE. Depending on the required accuracy, [(68)Ga]Ga-DOTA-TOC PET imaging may estimate the [(177)Lu]Lu-DOTA-TATE uptake. However, there could be a high variance between predicted and actual absorbed doses.