• Sharada Kadaba Sridhar

Computed Tomography (CT) is a timely, affordable, and accessible first line of screening for falls or altered mental status, especially among older adults. This introductory session will primarily motivate the objective evaluation of radiological markers that aid in the diagnosis and prognosis of dementia and traumatic brain injury and also serve as biomarkers for decision making in therapeutic intervention. We will also install software required for a hands-on experience of fundamental image processing and deep learning methods for head CT analysis, that will be presented in the upcoming parts of this Special Session.

• Alyssa Eastman

Navigating regulatory pathways can be challenging at different institutions. This session is designed to introduce terminology you may encounter and standard operating procedures for data acquisitions. We will review Institutional Review Board processes, data use agreements or certificates, NIH/DoD data access channels, and some tips for success.

• Sharada Kadaba Sridhar

Domain knowledge is indispensable in data-driven projects, and especially so for head Computed Tomography (CT) analysis using computational tools. This session will include a brief introduction to how head CT is acquired, and how to interpret it, followed by a tour of different intracranial anatomical regions and how they appear on multiplanar head CT, using 3D Slicer. Pathological hallmarks of neurodegeneration in dementia and traumatic brain injury (TBI) will be presented in comparison to healthy brains, to better appreciate the utility of head CT in neurological healthcare and research.

• Scarlett Cheong

This hands-on session introduces participants to CT neuroimaging data formats, metadata, de-identification practices, and common imaging artifacts. Using ImageJ and 3D Slicer, attendees will explore basic image inspection, intensity profiling, and format conversion. The session aims to build foundational skills in structural neuroimaging analysis for healthcare research applications.

• Sharada Kadaba Sridhar

Manipulating spatial and intensity properties of CT images are a prerequisite for all research pipelines, and highly translatable to different research applications. This session will introduce the fundamentals of such manipulations using Python and Simple Insights Toolkit (SITK).

• Charles Broadbent

Imaging features can be extracted from CTs that provide crucial information on local and global structures that can help understand the morphology and identify anomalies in scanned tissues. In this portion, we will walk through using Pyradiomics, a Python library that allows for streamlined extraction of radiomics features in CT images. We will look at real examples in brain CTs that demonstrate the utility of Pyradiomics in finding clinically important features that can be used in the diagnosis and progression of brain disease and injury.

• Sakshi Rathi

This tutorial provides an introduction to medical image segmentation, covering its types, key applications—especially in healthcare—and common methods and models. Focusing on deep learning approaches, it demonstrates how to leverage Python and Medical Open Network for AI (MONAI) within a PyTorch framework to develop effective 3D medical segmentation pipelines. Attendees will learn to set up a development environment using Google Colab, install essential libraries, utilize GPU acceleration, and access public medical imaging datasets. The hands-on session integrates MONAI into an existing PyTorch program, showcasing features like dictionary-based data transforms, NIfTI image loading with metadata, intensity scaling, channel manipulation, label-balanced cropping, and caching for speed optimization. A 3D U-Net architecture is employed with Dice loss and Mean Dice metric, along with sliding window inference and deterministic training for reproducibility. The tutorial concludes with a walkthrough of MONAI’s 3D segmentation pipeline using Jupyter or Google Colab notebooks.

• Alyssa Eastman, Charles Broadbent, Sakshi Rathi, Scarlett Cheong

We invite the attendees to discuss the tutorials presented during the sessions, and take this opportunity to network with other participants and presenters. We plan to split-up into groups based on specific interests and exchange ideas related to how the methods presented for structural CT image analysis can be extended to other areas of healthcare research and imaging modalities.