Auto-Segmentation Model for PSMA PET-Delineated Tubarial Glands

By Emily Menendez - December 20, 2022

The development of a computed tomography (CT)-based auto-segmentation model for prostate-specific membrane antigen (PSMA) positron emission tomography (PET)-delineated tubarial glands was examined by Sample et al, along with how to carry out clinical studies that involve the tubarial glands.

Bilateral tubarial salivary glands were recently discovered in the nasopharynx using PSMA PET, which raises concerns about how doses to this region may affect patient outcomes. The dose response of the major salivary glands varies, and the dose in tubarial glands may constitute a missing variable in the optimization of head and neck (HN) radiotherapy plans.

Developing methods to delineate the tubarial glands is a necessary first step to conducting clinical studies without the use of PSMA-PET images. An open-source program, Organogenesis, was developed by the research team for the auto-segmentation of tubarial glands using CT images.

Through Organogenesis, a predictive model is trained using contours from PSMA-PET images, which allows for accurate delineation of tubarial glands. Organogenesis provides a predictive model for tubarial glands that can create a path to clinical studies for determining the importance of incorporating tubarial glands into HN radiotherapy plan optimization.