MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

CT-Derived Pulmonary Perfusion

E Castillo1*, R Castillo2, Y Vinogradskiy3, G Nair1, I Grills1, T Guerrero1, C Stevens1, (1) William Beaumont Hospital, Royal Oak, MI, (2) Emory Univ, Atlanta, GA, (3) University of Colorado Denver, Aurora, CO

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: We evaluate a novel method for computing the variations in pulmonary blood mass apparent on non-contrast dynamic CT as a surrogate for pulmonary perfusion. Similar to CT-ventilation methods, our CT-perfusion method employs deformable image registration (DIR) to determine a spatial transformation that characterizes respiratory motion in an inhale/exhale CT image pair. Whereas CT-ventilation estimates volume change, our CT-perfusion method uses the DIR solution and Hounsfield Unit (HU)-defined material density to compute pulmonary blood mass changes, which have recently been shown to be a potential surrogate for perfusion. We assess the physiological fidelity of CT-perfusion via comparison to single photon emission CT perfusion (SPECT-P).

Methods: CT-Perfusion requires first estimating volume change using the robust Integrated Jacobian Formulation algorithm. Using the volume change information, the magnitude mass difference between spatially corresponding inhale/exhale subregional volumes is computed from the corresponding HU-density values. A full volumetric mass change image, which denotes the magnitude mass change for each lung voxel, is computed from the subregional estimates by solving a constrained linear least squares problem. The SPECT-P images and radiotherapy planning 4DCTs, acquired prior to treatment, for ten patients with non-small cell lung cancer were used to assess CT-perfusion. For all cases, CT-Perfusion was computed from 4DCT. After an affine registration, voxel-wise Spearman correlation between the SPECT-P and the spatially aligned CT-perfusion was computed.

Results: The mean (std) voxel-wise Spearman correlation between CT-perfusion and SPECT-P across the ten test cases was 0.455 (0.142). The values ranged between [0.231, 0.730], indicating good agreement between the two modalities.

Conclusion: The proposed CT-Perfusion method demonstrates a high correlation with SPECT-P, indicating that the modality is an accurate surrogate for pulmonary perfusion on the ten-patient cohort. Our approach represents the first method capable of quantifying pulmonary perfusion from non-contrast CT imaging.

Keywords

Ventilation/perfusion, Image Processing, CT

Taxonomy

IM/TH- Image Analysis (Single Modality or Multi-Modality): Image processing

Contact Email