Methods for image-based dosimetry and radiobiological modelling in radionuclide therapy

In radionuclide therapy (RNT), radioactive pharmaceuticals are administered for therapeutic purposes. Dosimetry, i.e. the estimation of the absorbed dose to various targets, can potentially lead to a better understanding of the therapeutic effects in RNT and allow for individualization of the therapy. However, dosimetry is seldom used in RNT clinical practice, and several problems degrade the information obtained from the measurements. This thesis comprises studies investigating the accuracy of gamma-camera based dosimetry and methods for subsequent radiobiological modelling through the calculation of the biologically effective dose (BED).
   Image-based activity quantification and absorbed-dose estimation are discussed in Papers III, IV, and VI. In Paper III, a pharmacokinetic (PK) model of the radiopharmaceutical ¹⁷⁷Lu-DOTATATE is used in combination with anthropomorphic computer phantoms to generate realistic spatial and temporal activity distributions that can be used as input for Monte Carlo simulations of gamma-camera images. The technique is demonstrated to be useful for accuracy assessment of image-based dosimetry. In Paper IV, the phantom-PK-model complexes are used to study uncertainty propagation in SPECT/CT-based renal dosimetry in ¹⁷⁷Lu-DOTATATE therapy. A range of uncertainty sources are modelled in the processing of the images and the combined uncertainty in renal absorbed dose is estimated to be approximately 6 %. In Paper VI, the accuracy of estimated tumour volumes and activity concentrations are considered for three automatic delineation methods. The accuracy is demonstrated to vary as a function of tumour size and reconstruction settings.
   Radiobiological modelling is considered in three papers. In Paper I a consistent theoretical framework for BED calculation including repair modelling is constructed. In Paper II, the equations are given a formulation suitable for numerical implementation, thus providing a flexible tool for use with non-standard absorbed-dose rate or repair functions. The agreement with analytical expressions is excellent. In Paper V, a formula is derived for BED in fractionated RNT with overlapping absorbed-dose rate functions. This is applied to a case where the originally prescribed activity would violate the local activity-handling regulations. Fractionation is demonstrated to result in a higher tumour BED than that resulting from a reduction of the total administered activity.