Using Semi-Automated CT Analysis to Determine Cyst Burden in Diffuse Cystic Lung Diseases

Diffuse cystic lung diseases (DCLD) are pathophysiologically distinct entities causing parenchymal cysts. Lymphangioleiomyomatosis (LAM) and Birt-Hogg Dube Syndrome (BHD) are examples of such DCLD. Patients have increased risk of pneumothorax and progression to end-stage respiratory failure. <bold>Methods:</bold> Using Siemens SyngoVia post-processing software, low attenuation volumes (LAV%) were calculated to determine cyst burden score in patients with LAM and BHD. Slices of 3mm at -950 Hounsfield Units were used. Cysts were clustered by volume (LAV%) and divided into 4 classes. Total cyst volume was calculated. Lung function,VEGF-D and pneumothorax history was compared to cyst scores. <bold>Results:</bold> Seventy-seven CTs from 32 patients with LAM and 19 CTs from 9 patients with BHD were interrogated. Mean total LAV% in the LAM and BHD group were 6.10% (SD+/−8.499%)and 6.43%(SD +/− 9.54%)respectively.28% of patients with LAM and 78% of BHD had a history of pneumothorax. There was no difference in mean cyst burden between those with LAM and BHD(p = 0.8866). Combining groups, there no difference between those with and without a history of pneumothorax(R2 = 0.01468). In BHD,cyst score did not correlate with lung function (FEV1:R2=0.0035, FVC:R2=0.0239 or DLCO:R2=0.0035). In LAM there was no between cyst score and VEGF-D level (R2 = 0.06570) or lung function(FEV1:R2=0.0923,FVC:R2=0.03762,DLCO:R2=0.02828). <bold>Discussion:</bold> We found no difference in total LAV% between BHD and LAM or between cyst burden, pneumothorax history or lung function. These results indicate that cyst burden alone does not seem to be a useful biomarker, however it may be that longitudinal cyst volume progression over time may correlate with progression of disease.