Multislice ct scanner

All subjects underwent HRCT (high resolution computed tomography) of the chest. Scans were performed on a Siemens multislice CT scanner (16-slice, 20-slice open CT, or 64-slice), without IV contrast. Inspiratory and expiratory images were performed. All scans were read by an experienced chest radiologist. The presence or absence of emphysema (bullae, thin-walled cystic spaces, or abnormal decreases in attenuation accompanied by vascular disruption) was recorded, as was the presence of bronchial dilatation, bronchial wall thickening, and air trapping as previously described [12 (link)].

A postoperative axial CT scan (resolution 3 mm, multi-slice CT scanner, Siemens Medical GmbH) was performed for each patient within 24 h after surgery to exclude any operative complication. Postoperative pneumocephalus is recognized as intracranial air-isodense voxel (Hounsfield value − 1000) mainly occurring in the prefrontal region (Mount Fuji sign, arrow) [18 ] and occasionally in the cisterns (air bubble sign, asterisk) [12 ] (Fig. 1c). VBV of postoperative supratentorial intracranial air volume was assessed by custom-made Matlab scripts (Version R2019a, MathWorks, Natick, MA, USA). The most challenging situation for automated detection was the discrimination between intracranial air and the frontal sinus. To overcome this, bone-weighted CT images were imported into the Matlab software, and all voxels with extracranial air were automatically removed based on their Hounsfield value (− 1000). Subsequently, manual removal of all slices below the frontal base was performed in order to avoid bias from the ethmoidal cells. Finally, the frontal sinus was denoted and removed manually (Fig. 1c).

Visceral fat (VF) and subcutaneous fat (SC) area were measured at the level of umbilicus (L4 vertebrae level) with the patient in supine position by multislice CT scanner (Siemens, Erlangen, Germany) in all the patients. The technique used for fat tissue measurements on CT has been standardized and validated [23 (link)]. The VF area was defined as all the pixels with adipose tissue attenuation coefficients. The tomographic attenuation of adipose tissue was defined to be between −150 and −50 Hounsfield units. The CT measurement was done by a radiologist blinded to the clinical data.