Research


Ablation Planning

Thermal ablation treatment is an important minimally invasive alternative to surgical resection of tumor in cancer treatment. However, the planning tools currently available are of limited practical use in real clinical decision making. We are developing and validating a framework for patient-specific thermal ablation simulations that is much more precise than existing tools. The model is being validated against 4D thermographic images of in ex vivo experiments and an extensive database of 3D ablation images from actual clinical treatments. Such a model will be a critical component in developing an evidence based suite of ablation treatment planning and guidance tools, as well as future technologies for accurate treatment delivery.

Collaborators

Dr. Damian Dupuy

  • Diagnostic Imaging
  • Director of Tumor Ablation at RIH

Prof. Benjamin Kimia

Prof. Edward Walsh

Quick project links (access restricted): APS Planpy Basecamp


Ultrasound Texture Analysis

We are developing an ultrasound texture analysis and machine learning framework and identifying image markers that indicate the presence of disease. This provides clinical decision making support that is expected to be helpful both in identifying diseases that are difficult to detect from ultrasound alone and in sparing invasive procedures when disease is unlikely.

This framework can be trained from extant clinical images and is robust to differences in devices and operators. It can be tuned for sensitivity and specificity for different screening procedures. Additional data sources can be incorporated, for example, elastography imaging or clinical findings of fibrosis can be used to improve classification accuracy. Alternate machine learning paradigms can also be easily integrated and evaluated.

Collaborators

Dr. Michael Beland (Diagnostic Imaging)

Dr. David Grand (Diagnostic Imaging)

Quick project links (access restricted): MUSTACHE Basecamp


ED Hydrocephalus Assessment

Collaborators

Dr. Lisa Merck (RIH Emergency Medicine/Neuroemergencies)

Prof. Stephen Pizer (UNC Computer Science)

Quick project links (access restricted): Basecamp


ED Smart Alarms

Collaborators

Dr. Leo Kobayashi (RIH Emergency Medicine/Sim Center)

Quick project links: PERSEUS