Generalized Student's t-distribution mixtures for autoradiographic image spread modelling

Area of Science:

• Nuclear imaging and autoradiography
• Statistical modeling
• Data analysis

Background:

• Current line-spread models in autoradiography have limitations in accurately fitting experimental data.
• There is a need for more robust and flexible models to analyze autoradiographic hot-line data.

Purpose of the Study:

• To propose a new class of models for autoradiographic hot-line data analysis.
• To demonstrate the superiority of these new models compared to existing ones.
• To introduce a specific, simple yet effective model from this new class.

Main Methods:

• Development of a new class of models as a linear combination of generalized Student's t-distributions.
• Application and comparison of these models to experimental iodine-125 labeled hot-line data in a resin section.
• Evaluation of model fit using goodness-of-fit metrics.

Main Results:

• The proposed models, derived from generalized Student's t-distributions, encompass all current line-spread models as special cases.
• A significant improvement in goodness of fit was achieved with the new models compared to previous ones.
• A specific two-component model from this class demonstrated superior performance in fitting experimental data.

Conclusions:

• The new class of models provides a significant advancement in analyzing autoradiographic hot-line data.
• The proposed modeling approach is versatile and applicable to various mixture modeling scenarios beyond autoradiography.
• Reliable estimation is indicated by sensitivity analysis, supporting the practical utility of these models.