Quantifying pulmonary inflammation in cystic fibrosis with positron emission tomography

Abstract

Rationale: Although infection contributes to morbidity in patients with cystic fibrosis (CF), the host inflammatory response is also an important cause of progressive pulmonary function deterioration. Quantifying the inflammatory burden in these patients is challenging and often requires invasive procedures. Positron emission tomographic imaging with [18F]fluorodeoxyglucose ([18FDG]) could be used as a noninvasive alternative to quantify lung inflammation.

Objective: To determine the relationships among lung [18F]FDG uptake, bronchoalveolar lavage (BAL) neutrophil concentrations, and pulmonary function in patients with CF.

Methods: Twenty patients and seven healthy volunteers were studied. A subset of seven patients also consented to undergo BAL. The uptake of [18F]FDG by the lungs was measured as the net influx rate constant Ki. Patients were stratified by rate of decline in pulmonary function into stable, intermediate, and rapidly declining groups. Ki was compared among groups and was correlated against neutrophil concentrations in BAL fluid.

Results: Ki was significantly elevated (p<0.05) among patients with CF as a whole compared with healthy control subjects (0.0015+/-0.0009 versus 0.0007+/-0.0002 ml blood/ml lung/min) but especially in patients with rapidly declining pulmonary function (0.0022+/-0.0011 ml blood/ml lung/min). Ki correlated positively with the number of neutrophils present in BAL fluid.

Conclusion: Imaging with [18F]fluorodeoxyglucose and positron emission tomography can be used to assess inflammatory burden in patients with CF. Elevations in Ki may be able to identify patients with more aggressive disease and may be useful in monitoring changes in inflammatory burden in response to novel treatments.

Figure 1.

Examples of PET images obtained from a patient with cystic fibrosis (CF) (rapidly declining group, rate of decline = 7.8% decrease in FEV₁/year, K_i = 32. 4 × 10⁻⁴ ml blood/ml lung/min) and a healthy volunteer. All images are in a transverse orientation at the midventricular level. [¹⁸F]fluorodeoxyglucose uptake by the myocardium (e.g., horseshoe-shaped left ventricle) is seen. Changes in uptake by the lungs over time are more subtle, with increased washout of tracer from the lungs (consistent with less “trapping” of the radiopharmaceutical within lung tissue) in the healthy volunteer when compared with the patient with CF.

Figure 2.

Time–activity curves obtained from analysis of the images shown in Figure 1 (the analysis shown here includes multiple image time-points not included in Figure 1). Lung tissue radioactivity remains elevated by the end of the imaging session in the patient with cystic fibrosis (CF), whereas activity is declining in the healthy volunteer, consistent with increased trapping (i.e., uptake) of the radiotracer [¹⁸F]fluorodeoxyglucose.

Figure 3.

Patlak plots for the same two subjects shown in Figure 1. The slope of the plot for the patient with cystic fibrosis (CF) is noticeably steeper than that of the healthy volunteer, consistent with a greater rate of [¹⁸F]fluorodeoxyglucose uptake by the lungs in the patient with CF.

Figure 4.

Two separate microscopic views (A and B) of autoradiograms of cells harvested from bronchoalveolar lavage fluid from one patient with cystic fibrosis (intermediate group, rate of decline = 1.9% decline in FEV₁/year, K_i = 11. 4 × 10⁻⁴ ml blood/ml lung/min) after incubation with [³H]deoxyglucose and subsequent 8.5 wk-exposure of slides to autoradiographic film. Despite a relatively high background signal, grains are concentrated over neutrophils (enlargements of two example cells, indicated by arrows, shown in insets), indicating specific uptake of the radiotracer by neutrophils.

Figure 5.

(A) Correlation between neutrophil numbers measured in bronchoalveolar lavage (BAL) fluid and K_i, as a measure of in vivo uptake of [¹⁸F]fluorodeoxyglucose by positron emission tomography. The R² value without the “outlier” point representing the highest cell concentration was still high at 0.95. (B) Correlation between neutrophil numbers collected by BAL and lung function, expressed as the rate of decline in % predicted FEV₁. In this case, R² decreased to 0.44 if the outlier point is eliminated. PMN = polymorphonuclear leukocytes.

Figure 6.

Means and SDs of regional K_i in right and left upper lung zones (RUZ and LUZ) and right and left lower lung zones (RLZ and LLZ) in patients with cystic fibrosis (CF) and healthy volunteers, using standard ANOVA to compare healthy volunteers and patients with CF and repeated measures ANOVA on ranks to assess differences within groups. *p < 0.05 compared with healthy volunteers and compared with the RLZ regions in the patients with CF. +p < 0.05 compared with healthy volunteers.

Figure 7.

Point plots of K_i (A) and corrected K_i (B), as measures of the rate of [¹⁸F]fluorodeoxyglucose uptake by the lungs, for healthy volunteers (Nl) and patients with cystic fibrosis, divided into stable (S), intermediate (I), or rapidly (R) declining groups. Also shown are the mean and SD for each group and the average % predicted FEV₁ in each group (A). The dashed line indicates 2 SD above the mean in the healthy volunteers. In the patients with cystic fibrosis, most values are > 2 SD above the normal mean.