Inter-center variability of dosimetry methodology and the impact on reported absorbed dose for [¹⁷⁷Lu]Lu-PSMA-617 therapy.

Abstract

[¹⁷⁷Lu]Lu-PSMA-617 radioligand therapy improves overall survival in advanced prostate cancer and is under evaluation in early-stage disease. Dosimetry can guide personalized [¹⁷⁷Lu]Lu-PSMA-617 therapy by determining tumor and organ absorbed doses (ADs), but workflow variability and lack of standardization limit clinical use. To identify factors most influencing AD, we evaluated inter-center variability using a single-center dataset analyzed by three centers, focusing on the effects of volume delineation and time-activity integration. Dosimetry was performed at three experienced centers using local workflows. Inter-center comparisons were conducted in three steps: (1) comparing absolute ADs, (2) comparing inter-center variability using the coefficient of variation (CV) per patient and anatomical region, and (3) evaluating the methodological steps by homogenizing data across centers. Homogenizations included organ and tumor volumes for S-factors (volume-homogenization effect), activity within volumes of interest (activity-homogenization effect), and time-activity curve fitting (TAC-fitting effect), allowing assessment of the impact of individual methodological AD differences. Significant inter-center differences in median ADs were observed for salivary glands (<i>p</i> = 0.010), kidneys (<i>p</i> = 0.002), liver (<i>p</i> = 0.001), and lesions (<i>p</i> = 0.001). Variability was highest for the lesions (median CV 71.5%). In organs, variability was highest in salivary glands (median CV 35.7%), followed by liver (median CV 21.2%) and kidneys (median CV 16.8%). Volume-homogenization had the biggest impact on inter-center variability for lesions, reducing median CV to 32.3%; this reduction in variability was less pronounced for organs. Activity-homogenization proved more influential for organs, reducing median CV in salivary glands, kidneys and liver to 12.9%, 11.8% and 11.5%, respectively. The TAC-fitting effect was limited. Substantial inter-center variability in AD was observed, this was primarily reduced by volume- and activity-homogenization, emphasizing the need for standardized delineation protocols and consensus guidelines to ensure reproducible and comparable dosimetry results. The online version contains supplementary material available at 10.1007/s00259-026-07832-3.