Small Animal Cone beam CT

1
Small Animal Cone beam CT

• Owen Gray
• Sungjun Kim
• Pinak Pujari

2
Presentation Outline

• Definition
• Relevance of the Project
• Deliverables
• Summary of Approach
• Dependencies
• Background Reading
• Future Extensions

3
Definition

• This project is a part of a large research
  effort to develop an image guided small animal
  radiation research platform (SARRP) that will
  accurately deliver complex ionizing radiation
  dose distributions in small animal tumor model
  systems, mice, rats and rabbits.

4
Relevance of the Project

• Currently, no facilities exist for validating
  advanced radiation therapy techniques in animal
  models
• Systemic interventions are widely tested in mice
  and other models
• Image guided radiation therapy techniques are
  typically tested on human subjects

5
Relevance of the Project

• The ability to accurately deliver dose
  distributions in standard laboratory models would
  provide substantial benefits
• Preclinical demonstration of efficacy of new
  techniques
• Faster validation and deployment of novel
  treatments

6
Deliverables

• We are going to build a software tool
• for importing, processing, reconstructing and
  exporting images acquired by CBCT as a
  deliverable portion of the project.

7
Accounting for Intensity Variation

• Beam intensity varies across the phosphor
  screen/flat panel detector
• Need to characterize the intensity variation over
  the image area
• Need to correct for any disparities
• Acquire images with no object between X-ray
  source and detector at varies beam intensities
• Calculate a correction function to normalize
  intensity

8
Accounting for Intensity Variation

• Beam intensity falls of as an inverse square
• Beam intensity is relatively stable close to the
  image center, but falls off towards the edges

9
Dewarping

• Cone beam images will be subject to some level of
  distortion due to external fields (as in 445
  assignment)
• These distortions may not be sufficiently large
  to impact reconstruction
• Data must be collected to characterize
  distortion, and correct if necessary

10
Attenuation

• In traditional CT, images are acquired from
  points around the major axis of the patient
• The patient may modeled as a cylinder with
  reasonable accuracy
• In the cone beam setup, the patient rotates
  relative to the scanner
• The difference in attenuation between transverse
  and axial images must be characterized and
  corrected

11
Attenuation

• An axial image will require higher beam
  intensities, or longer beam duration
• The attenuation must be corrected to allow
  uniform mapping of image intensity to tissue
  density across multiple aspects
• Need to determine upper limit of exposure before
  detector/camera saturation occurs

12
Attenuation

• Attenuation must be approximated at each gantry
  position
• A lookup table or high-order polynomial must be
  developed to account for the attenuation
  variation at each image angle through 180 degrees

13
Attenuation

• Attenuation varies substantially with rotation
• With typical scanner geometry, the amount of
  tissue traversed from each image angle varies
  relatively little
• In the small animal apparatus, attenuation may
  vary by a factor of four between axial and
  transverse images

14
System Overview
15
Lab Setting
16
(No Transcript)
17
Dependancies

• Experiment data
• Software (MS C, Matlab)
• Existing algorithms

18
Background Reading
19
Future Extensions

• Summer research and masters thesis
• 1) A mechanical shutter will be developed and
  integrated with the semester software to
  facilitate real CBCT acquisition.
• 2) Pencil Beam Dose Engine for focused beam.