Functions using dual-energy CT (DECT) for bone imaging:

Two major functions:

1. Optimized 3-material decomposition for edema imaging and edema segmentation
2. DECT based bone density calibration

All functions are based on simulated monoenergetic images at 2 energy levels (e.g., 40 and 90 keV)

Also includes functions for image registration:

1. MRI to DECT
2. HR-pQCT to DECT
3. Contralateral knee to injured knee

For in-depth instructions:

• Bone Lab members: Please see BOYD drive: /BOYD/04 - Resources/05 - Tutorials & Training/DECT
• Others: Please contact the Bone Imaging Lab

Quick Start

To run 3-material decomposition:

1. Generate simulated monoenergetic images at 40 and 90 keV
2. Convert DICOMs to MHA: dicom_Converter_batch.py
3. Run basic material decomposition, optimized for edema imaging: MatDecomp_EMSI_final.py
4. Segment marrow space:
   1. Generate periosteal segmentation (e.g., using FemurSegmentation repository)
   2. Calibrate CT images: Standard_SEQCT_PhantomCalibration_GSI.py
   3. Run EndostealSegmentation_LANG.py
5. Filter edema image, then apply bone marrow mask and threshold to segment edema region: Segment_Edema.py

For DECT-based bone density calibration:

• Scan using density calibration phantom
• Create mask for density rods
• Run DECT_PhantomCalibration.py

For image registrations (Note: Tibia and Femur should always be registered separately):

• MRI to DECT registrations:
  • Flip sagittal MRI into axial coordinate system to match DECT: Flip_Coordinates_SagittalToAxial.py
  • Run the appropriate registration script:
    • To register PD or T1 MRI to DECT: Register_PDMRI-DECT_LandmarkInitialized.py
    • To register T2 MRI to DECT:
      1. First register PD/T1 to DECT
      2. Registser T2 to PD/T1: Register_T2-PDMRI.py
      3. Combine results of steps 1 and 2 above: Transform_T2MRI-DECT.py
• HR-pQCT to DECT registrations:
  1. Convert HR-pQCT aims to mha: Aim2Mha_py3.py
  2. Downsample HR-pQCT images by a factor of 4: Downsample_XCT.py
  3. Run registration: Register_XCT-DECT_LandmarkInitialized.py
• Contralateral registrations: run Register_Injured-Contralateral.py