Maximum-likelihood quantification of low-activity 225 Ac and 134 Ce/La samples using NaI(Tl) gamma spectroscopy
Year: 2026
Type: article
Abstract: Accurate activity quantification is essential for all radionuclide studies. Gamma spectroscopy with well-type NaI(Tl) detectors offers a sensitive method for activity determination, but composite samples pose a challenge because integration of counts in photopeak energy windows is sensitive to spectral crosstalk. Using the theranostic pair of positron-emitting 134 Ce/La and therapeutic 225 Ac as a focus case, we implemented a maximum-likelihood estimation (MLE) framework to estimate 225 Ac and ^134Ce/La activities in pure and mixed samples, and compared its performance against weighted least squares (WLS), spectrum stripping, and energy window counting (EWC). For MLE and WLS, pure-isotope spectra were energy-calibrated and normalized over a defined energy window to form activity-independent spectral templates. The counts in a measured spectrum were modeled as a linear combination of isotope-specific spectral templates, which capture the measured spectral shape under fixed acquisition conditions. These were scaled by isotope-specific calibration response curves to account for detector sensitivity and geometric efficiency. For MLE, activity estimates were obtained by maximizing the Poisson likelihood of the observed spectrum over the analysis window. When optimal 225 Ac and 134 Ce templates were chosen, MLE estimated the 225 Ac activities of pure 225 Ac standards to within 8 ± 3%, compared to 7 ± 7% for EWC and 7 ± 4% for WLS. For pure 134 Ce/La standards, 134 Ce/La activities were estimated to within 6 ± 2% using MLE, compared to 5 ± 9% for EWC and 6 ± 2% for WLS. In mixed 225 Ac and 134 Ce/La samples, estimated 225 Ac activity from the MLE method agreed with known calibration values to within 0 ± 4%, and 134 Ce activities agreed within 7 ± 2%. The corresponding values were 2 ± 5% and 2 ± 2% for WLS, and 6 ± 9% and 0 ± 5% for spectrum stripping.
Source: Journal of Instrumentation
Language: English
Subfield: Radiation
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,290
OpenAlex ID: W7197003482