A population-based method to determine the time-integrated activity in molecular radiotherapy

Area of Science:

• Nuclear Medicine
• Medical Physics
• Radiotherapy Dosimetry

Background:

• Accurate dosimetry in molecular radiotherapy relies on calculating time-integrated activities (TIAs) for tumors and organs.
• The precision of TIAs is critically dependent on the selection of an appropriate mathematical fit function.
• Current model selection is limited by the scarcity of biokinetic data typically available from individual patients.

Purpose of the Study:

• To develop and demonstrate a population-based model selection method for determining individual TIAs.
• To enhance the accuracy of dosimetry calculations, particularly when limited patient biokinetic data is available.
• To optimize the selection of fit functions for analyzing [177Lu]Lu-PSMA-I&T kidney biokinetics.

Main Methods:

• Retrospective analysis of renal biokinetics from thirteen patients undergoing [177Lu]Lu-PSMA-I&T therapy for metastatic castration-resistant prostate cancer.
• Fitting twenty different mono- and bi-exponential functions to patient data, with varying shared and individual parameters.
• Utilizing Akaike weights, derived from the corrected Akaike Information Criterion, to select the best-fit function among those with acceptable goodness of fit (CV < 50%).

Main Results:

• A specific exponential function ([Formula: see text]) with a shared parameter ([Formula: see text]) was identified as the most supported model, achieving an Akaike weight of 97%.
• Individual parameters ([Formula: see text] and [Formula: see text]) were fitted per patient, while the parameter [Formula: see text] was shared across the population, yielding a value of 0.9632 ± 0.0037.
• The population-based approach allowed for the investigation of functions with a higher number of parameters, leading to improved fits.

Conclusions:

• The population-based model selection method enables the use of more complex fit functions, resulting in superior data fitting for TIAs.
• This approach reduces uncertainty in Akaike weights and the selection of the optimal best-fit function.
• Shared parameters determined from population data facilitate the fitting of more appropriate functions for individual patients, even with limited biokinetic data.