VolMoVis: Real-Time Volumetric Motion Visualization with 4D Tomographic Reconstruction

Source: Mark Kostich, stock.adobe.com
Background
Cone beam computed tomography (CBCT) is an essential imaging modality use in radiation therapy procedures. CBCT works by taking 2D x‑ray images from multiple angles, which are reconstructed into a 3D volume. CBCT is used for patient setup and verification, to ensure that the dose is delivered to the tumor region precisely. However, during the procedure, the tumor region may move, especially in the thoracic and abdominal regions due to breathing. Conventional CBCT reconstruction methods are too slow or lack sufficient temporal resolution to enable visualization in real‑time during the procedure.
Technology
VolMoVis is a cutting-edge technique that utilizes neural networks to overcome the temporal limitations of conventional 4D CBCT reconstruction. By utilizing a continuous spatiotemporal neural representation, it decomposes dynamic volumetric data into a static reference volume and a continuous deformation field, which allows rapid visualization of patient movements. This innovative approach allows for real-time volume generation and segmentation, enabling motion visualization with CBCT for medical imaging.
Advantages
Enables real‑time reconstruction of 4D CBCT images - Improves image quality and signal to noise ratio over existing methods - Allows for real‑time visualization of target areas in motion during radiation therapy treatment - Compatible with existing medical linear accelerator devices with built‑in CBCT imaging
Application
Image‑guided radiation therapy - Real‑time motion visualization for treatment planning and medical research
Inventors
Arie Kaufman, Distinguished Prof. & Chair, Computer Science
Gaofeng Deng, Graduate Research Assistant, Computer Science
Licensing Potential
Development partner - Commercial partner - Licensing
Licensing Status
Available
Licensing Contact
Donna Tumminello, Assistant Director, Intellectual Property Partners, donna.tumminello@stonybrook.edu, 6316324163
Patent Status
Provisional patent
Tech ID
050-9516
