Data Collection and Definitions
Medical records of these patients were reviewed. Patients with missing data and without abdominal CT were not included in the study. This study screening, inclusion and exclusion flow chart is presented in Figure 1.
Radiological evaluation of the patients with abdominal CT was performed by the same experienced radiologist using standard anatomical landmarks. During abdominal CT the muscles at the level of L3 vertebra were determined and the psoas, paraspinal muscles (erector spinae, quadratus lumborum) and abdominal wall muscles (transverse abdominus, external and internal obliques, rectus abdominus) were marked. The muscle cross-sectional area at the level of L3 was used as it was linearly associated with whole-body muscle mass, and by dividing this value by the square of the height the skeletal muscle index was calculated for each patient in cm2/ m2. CT scan with 64-detector (Aquilion 64, Toshiba Medical Systems, Tochigi, Japan, 2011) was used in abdominal CT. Two mm section thickness, 64 x 0.5 collimation, 0.5 seconds rotation time, 120 kV and 300 mA were used in CT investigations. The evaluation of the images was done on the OsiriX (10.0, 64 bit, Switzerland) workstation. The evaluation was performed at L3 vertebral level, in the axial section in which both transverse projections were observed. Skeletal muscle area was measured by 2D/3D segmentation tool at this level. In skeletal muscle area measurements, pixels in the range of -30, +150 Hounsfield Unit (HU) density were marked by automatic segmentation [14,15]. The necessary adjustments in the marked contours of the fields have been made manually.