Predicting tumour and organ dose in α-radiopharmaceutical therapy: a ⁸⁹Zr-based dosimetric approach for ²²⁵Ac.

Abstract

The aim of this study was to assess whether zirconium-89 (⁸⁹Zr)-labelled antibodies can act as a surrogate for actinium-225 (²²⁵Ac)-labelled antibodies to predict tumour- and organ dose. To this end, we performed ex vivo biodistribution and dosimetry studies with two anti-CAIX antibodies (hG250 and FcRn(-)hG250), radiolabelled with ⁸⁹Zr and ²²⁵Ac, in a renal cell carcinoma mouse model. BALB/c nude mice bearing subcutaneous SK-RC-52 tumours were injected intravenously with 25-50 kBq ²²⁵Ac-labelled or 1 MBq ⁸⁹Zr-labelled anti-CAIX antibody, at equal protein dose. Organ and tumour absorbed doses were calculated based on the uptake data at 4, 24, 72, and 168 h for FcRn(-)hG250 and 24, 72, and 168 h for hG250. The predictive value of ⁸⁹Zr for ²²⁵Ac was evaluated in both groups, by comparing the predicted ²²⁵Ac absorbed dose (based on ⁸⁹Zr-antibody) and the measured ²²⁵Ac absorbed dose. The biodistribution profile of the theranostic pair ⁸⁹Zr/²²⁵Ac was comparable for FcRn(-)hG250. For hG250, significant discrepancies between tumour, spleen and liver uptake were observed. The predicted ²²⁵Ac tumour dose was overestimated by 3.9 ± 2.0 Gy (6 ± 50.6%) and underestimated by 1413.7 ± 629.1 Gy (69 ± 44.5%) in the FcRn(-)hG250 and hG250 groups, respectively. The deviation of the predicted versus measured ²²⁵Ac-dose in off-target organs was most notable in liver and spleen for both antibodies. This study highlights variability in the predictive accuracy of the ⁸⁹Zr/²²⁵Ac theranostic pair for tumour and organ dosimetry in a ccRCC mouse model, depending on the antibody used. FcRn(-)hG250 enabled accurate tumour dose estimation, whereas hG250 showed mismatched uptake limiting prediction, possibly due to inherent differences in the properties and biological effects of both ²²⁵Ac and ⁸⁹Zr. These findings underscore the need to further investigate how ⁸⁹Zr-PET can be used for ²²⁵Ac-TRT dose prediction.