Abstract:        

Dental enamel, the most mineralized tissue in the body, protects the tooth and is vital to our overall health and wellbeing. Enamel has evolved remarkable material properties to withstand chewing forces; it is characterized by high hardness and stiffness yet possesses a relatively high toughness. Defects in enamel can compromise its performance and represent a significant cost to society; the U.S. alone spends over $135 billion annually on prevention and treatment. Congenital defects in particular are highly diverse in their etiology and prevalence, although they share a commonality in that they must occur during enamel formation, or amelogenesis. There is thus a need for a better fundamental understanding of amelogenesis to better delineate mechanisms of disease and eventually develop new approaches for intervention and treatment.

In this presentation, we will describe our work in two sections; first, we report on the creation and application of a deep learning tool for rapid, quantitative measurements in synchrotron μCT images of mouse jaws. Specifically, we used semantic segmentation using convolutional neural networks for automatic segmentation of dental tissues followed by an analysis pipeline to output metrics for the whole jaw. μCT is often the first tool used to visualize changes in the mineral in amelogenesis models and is widely available to researchers, so we anticipate that this tool will be broadly useful for the enamel community. To that end, we demonstrate its accuracy and flexibility in segmenting a variety of mutant phenotypes. We then use this tool in combination with other characterization techniques to study two mouse lines mimicking a genetic disorder in humans that leads to misassembled enamel protein matrix. We combine synchrotron μCT, electron microscopy, and synchrotron μXRD mapping for a thorough analysis of the enamel layer that revealed defects across length scales. Although significant in vitro work has been done to understand protein assembly in amelogenesis, our work is the first to study it in vivo, with important implications for similar mutations in humans.

Location: 

431/C010 or Teams

Microsoft Teams Information
Click here to join the meeting
Meeting ID: 241 289 087 666
Passcode: xuL2aH