Lung diseases are one of the leading causes of death in the UK; responsible for 20% of all deaths each year. Inflammation is thought to be an important driver of the pathogenesis and progression of several lung diseases. Positron emission tomography/computed tomography (PET/CT) is an imaging modality capable of providing functional and molecular imaging through detection of trace quantities of a radioactive tracer. 18F-FDG is the most widely available tracer and in several small studies has been used to investigate diffuse lung diseases such as chronic obstructive pulmonary disease (COPD). These studies rely on quantification of the PET image.

Static acquisitions provide information on the biodistribution of the tracer at a single time point after administration, the standardised uptake value (SUV) is the most widely used measure of 18F-FDG uptake. Dynamic acquisitions provide information on the spatial and temporal distribution of the tracer; linear and non-linear modelling techniques allow estimation of the metabolic rate of 18F-FDG in the lung. Previous studies have used a linear method called Patlak graphical analysis, whilst non-linear compartmental models have been used more recently to estimate metabolism. However, these contending measures of pulmonary 18F-FDG uptake, which are putative biomarkers of lung inflammation, have so far been disparately applied. Further, there is nascent understanding that these imaging endpoints are affected by pulmonary air and blood volume; the importance of this effect will likely depend on the disease and its severity. These issues are exacerbated by the presence of respiratory motion and the low signal- to-noise ratio achieved in PET studies. This has led to questions regarding the utility of quantitative 18F-FDG PET to assess lung inflammation.

In this work, prospective and retrospective clinical studies were used to assess the clinical, biological and technical validity of 18F-FDG PET imaging endpoints in several clinically relevant diseases. The central hypothesis was that pulmonary inflammation can be assessed using quantitative 18F-FDG.

Using retrospective data from two complementary imaging studies, pulmonary 18F- FDG uptake was investigated in COPD patients, α1ATD patients, smokers without COPD and heathy non-smokers. The results demonstrate that the different 18F-FDG imaging endpoints produce disparate findings, and this is exacerbated by the presence of varying blood and air volumes due to emphysema. Nevertheless, measures derived from Patlak analysis revealed elevated uptake consistent with the pathophysiologi- cal understanding of the disease process and further demonstrated correlation with other putative markers of inflammation hence, one could speculate that it relates to inflammation. Further, 18F-FDG imaging outcomes assessed using Patlak analysis were shown to be more reliable than compartmental modelling outcomes. However, the Patlak outcomes are composite measures, not only driven by inflammation but also by pulmonary blood and air. In circumstances where differences in pulmonary blood and air volume between subjects are substantial, it may not be a suitable biomarker of inflammation, but it could be a useful marker of disease activity. Further tissue validation and independent measures of pulmonary blood are required to support its role as a marker of inflammation.

In a prospective study, dynamic 18F-FDG PET scans were used to evaluate pulmonary inflammation in sarcoidosis patients and healthy controls. The results show that 18F-FDG uptake was increased in sarcoidosis using Patlak analysis, whilst no difference was found using SUV or compartmental models. Preliminary findings suggest that the signal relates to inflammatory cell counts (macrophages and lymphocytes were most numerous) rather than any one specific cell line; however, further evidence is required to determine if 18F-FDG uptake is driven by underlying inflammation. Consistent with previous findings, mismatch in CT and PET lung images has substantial effect on the quantitative 18F-FDG outcomes; notably, Patlak outcomes were less influenced than compartmental modelling.

In summary, the observations made in this work demonstrate the substantial challenges of using 18F-FDG PET/CT to assess diffuse lung disease. Given the incongruity between the different imaging outcomes, these data highlight that future studies should be carefully planned with particular justification of the acquisition and analysis methods. The results of this work suggest that Patlak measures may have the most utility in diffuse lung disease. However, differences in Patlak measures should be interpreted judiciously, as they may be driven by differences in pulmonary blood and air along with inflammation; further study is required to determine if it may be a useful marker of disease activity. In contrast, no differences in pulmonary 18F-FDG uptake between patients and controls were found using compartmental modelling across all studies. Equally, the SUV was found to have poor utility across studies. Future efforts to expedite the use of novel tracers that are more specific to inflammation combined with the development of improved noise reduction techniques may improve the utility of quantitative PET/CT in the context of lung inflammation.