Room: AAPM ePoster Library
Purpose: To quantify: 1) the accuracy of a new pulmonary functional imaging approach, 68Ga-Galligas/68Ga-macroaggregated albumin (MAA) PET/CT, to assess regional ventilation/perfusion (V/Q) matching by comparison with pulmonary function test (PFT) results; and 2) correlations between V and Q images, in patients with lung cancer.
Methods: Thirty patients who previously underwent 68Ga-PET/CT-V/Q and PFT in a clinical trial were analyzed. Patients with missing 4-dimensional (4D)-PET/CT, missing PFT, or substantially low 68Ga signal in the lungs were excluded. 4D-PET/CT images at the end-exhalation phase were used. V and Q defects were segmented by thresholding. For the rest of the lungs, a base-10 logarithm of the V/Q ratio was calculated for each voxel using the normalized V and Q values, with zero log(V/Q) representing perfect matching. We calculated the proportions of voxels with log(V/Q) values within eight different cutoff deviations from zero (between 0.05-0.4) to the total lung to quantify V/Q matching, and correlated with diffusing capacity of the lung for carbon monoxide (DLCO) using the Spearman’s correlation coefficient. We also calculated the voxel-wise Spearman’s correlation between V and Q images for each patient.
Results: For 19 evaluable patients, the proportions of voxels with log(V/Q) values within different cutoff deviations were all significantly correlated with DLCO, with the cutoff of 0.025 showing the highest correlation (0.64, P<0.01). The correlations between V and Q images ranged from 0.41 to 0.86. Four patients showed a strong correlation (>0.8) with homogeneous or heterogeneous distributions (including defects). Two patients showed a weak correlation (<0.5) with differences in the location of highly functional regions.
Conclusion: 68Ga PET/CT-measured V/Q matching significantly correlates with PFT results, indicating its potential to provide reasonable estimates of regional gas exchange efficiency in lung cancer. The voxel-wise correlation between V and Q varied widely between patients, suggesting that V and may provide complementary information.
Funding Support, Disclosures, and Conflict of Interest: This study is supported by the Cancer Australia Priority-driven Collaborative Cancer Research Scheme (Project No. 1060919). Dr. Hofman is supported by a Clinical Research Fellowship from the Peter MacCallum Foundation.