A validation protocol for177Lu-SPECT image quantification as a basis for multi-centre kidney dosimetry.

Objective. Multicentre studies in molecular radiotherapy face challenges in ensuring the comparability of dosimetry results. The aim of this study was to develop a validation protocol to achieve equivalent¹⁷⁷Lu-SPECT quantification across centres, consisting of a test exercise in combination with an action level.Approach.¹⁷⁷Lu-SPECT/CT imaging of 3D-printed kidney phantoms with non-uniform activity distributions was conducted at five hospitals in Sweden and the UK. In total, 12 quantification systems (QSs) were included, where a QS was defined as the combination of a given camera system, reconstruction method and image-analysis method. The accuracy of image-based estimates of the activity concentration was investigated for three approaches to multicentre dosimetry: one site-specific, and two variants of centralised approaches. For the site-specific approach, the reference value was blinded to the evaluating operators. Prior to the evaluation, the participants agreed on an action level, specified as an obtained deviation larger than ±10% from the reference value. Sites responsible for a QS that yielded larger deviations, and therefore did not comply, were required to present an action plan, i.e., a plan for further method optimisation.Main results. For the site-specific approach, deviations larger than 10% were obtained for six out of 12 QS (range -11% to +27%). Action plans were presented for these QS and was implemented by one site (two QS). For the two variants of centralised approaches, where nine QS were included in total, deviations of within 10% were obtained for nine and seven QS, respectively.Significance. A quantification test combined with an action level can serve as an initial validation to enable further development and to achieve equivalent¹⁷⁷Lu-SPECT quantification across sites.