Hybrid method for estimating lung ventilation from CT by combining intensity and motion information.

Area of Science:

• Medical Imaging
• Radiology
• Computational Biology

Background:

• Functional lung imaging captures regional ventilation, but CT-based ventilation imaging (CTVI) faces challenges in deriving functional data from anatomical observations.
• Accurate ventilation maps are crucial for clinical applications, yet solely relying on CT data presents limitations.

Purpose of the Study:

• Introduce the hybrid estimation of computed tomography obtained respiratory function (HECTOR) method for generating accurate CTVI.
• Integrate a deformable image registration (DIR)-based volume estimation with an established air-tissue density model.
• Develop a novel approach for combining these two complementary ventilation estimation components.

Main Methods:

• Employ a tailored DIR algorithm to quantify respiratory motion between inhale and exhale CT phases.
• Generate volumetric change maps using the Jacobian determinant method from motion data.
• Calculate an HU-based air-tissue density estimation and combine it with volumetric data using a smooth minimum function.
• Validate the HECTOR method against reference ventilation images (RefVIs) using Spearman correlation and Dice similarity coefficients (DSC) on public datasets (VAMPIRE, TCIA).

Main Results:

• The HECTOR method achieved superior performance compared to existing CTVI methods, outperforming the best reported results on the VAMPIRE dataset.
• Achieved mean Spearman correlation coefficients of 0.62 (Galligas PET), 0.49 (DTPA-SPECT), and 0.66 (TCIA).
• Demonstrated higher correlation scores than individual volume- and density-based methods, highlighting the benefit of the hybrid approach.

Conclusions:

• The HECTOR method provides a novel, high-fidelity workflow for CTVI, generating accurate ventilation maps.
• Combining diverse data types (volume and density) effectively models complex respiratory dynamics.
• This approach holds potential for improving radiation therapy planning and thoracic dose-response assessment.