Challenges with Quantitative Analysis of Whole Lung Images

Objectives Concentration within a lung voxel is not representative of the concentration within the lung tissue due to the significant pulmonary air and blood components. If this ‘Tissue Fraction Effect’ (TFE) is not accounted for, large PET quantitative errors will be introduced in both SUV and kinetic parameters. This work aims to determine methods of correcting for the TFE and to assess the errors introduced if the effect is ignored.

Methods Dynamic FDG PET/CT images were acquired from a patient suffering from idiopathic pulmonary fibrosis (IPF). Uptake within this patient was used to create simulations representing voxels with different lung parenchyma, blood and air components. Kinetic and SUV parameters were compared to the true values to determine which parameters were affected by the TFE and the methods with which to correct for this affect. The corrections were then applied to the patient data to determine the extent of the errors introduced if the TFE is not accounted for.

Results Both the air and blood components were found to be significant in the patient with respect to the quantity of parenchyma in both regions showing fibrosis and healthy appearing tissue (table 1). SUV and the kinetic parameters K₁ (Flow-Extraction Product) and K_i (influx rate) were found to be affected by the TFE. Determination of the air and blood correction factors was implemented through use of the CT image and compartment modelling respectively. After correction, the ratio of fibrosis to normal appearing tissue in the patient was found to decrease from around 1.7 to approximately 0.4 for all parameters, showing an inversion of the initial interpretation of results (i.e. from hot to cold).

Conclusions Blood and air are significant components of the lung and cannot be neglected for both SUV and kinetic parameter accuracy. These results suggest that patient images could be misinterpreted if the TFE is not appropriately corrected for.

Research Support This work is supported by funding from GSK, Fibrosis DPU, GlaxoSmithKline R&D and EPSRC. This project is supported by researchers at the National Institute for Health Research, University College London Hospitals Biomedical Research Centre.