Plain CXR: The plain CXRs were performed at maximal inspiration, with each patient standing upright with the chest against a wall and arms away from the sides. Patients were asked to maintain an erect position. The following measurements were taken (Fig 1): (1) transverse diameter (on AP view) at the levels of the manubrium sterni and the 7th and 11th thoracic vertebral bodies (T7, T11); (2) AP diameter (on lateral view) using a vertical reference line drawn tangentially to the posterior aspect of vertebral bodies T6 through T9. A parallel vertical line was drawn at the most posterior aspect of patient’s back (thoracic soft tissue). Three horizontal AP diameter measurements were taken from this line to the posterior aspect of the sternum, at the same levels as the transverse diameter measurements (as modified from Walsh et al).
CT Scan: The CT scans were obtained with the GE Highlight Advantage scanner (General Electric Corp; Milwaukee, WI) or the Picker PQ 5000 scanner (Picker Corp; Cleveland, OH). Scan thickness was 10 mm and scanning time was 1 s. Scans were obtained at the level of the pedicles of the 4th, 6th, 8th, and 10th thoracic vertebrae. All scans were obtained at full inspiration, with patients holding their breath in supine position, and arms positioned parallel to the longitudinal axis of the body. At each aforementioned vertebral level, the AP diameter was measured from the anterior aspect of the vertebral body to the posterior aspect of the sternum (AP midsagittal). The maximal AP diameter of the right and left hemithoraces (AP right and left) and the maximal transverse diameter were measured as well (Fig 2) (taken from Cassart et al). canadianfamilypharmacy

Data Analysis
All the measurements were made in centimeters and were rounded off to the nearest 0.1 cm. Each diameter represents the maximal internal thoracic diameter at a particular level. Every measurement was obtained independently by two investigators and the variability in each measurement between the two investigators was < 0.1 cm. Values reported represent the average of two independently obtained measurements at each level. Both investigators were blinded to the physiologic data.

Figure 1. Thoracic diameter measurements on the AP and lateral CXRs. A to B, C to D, and E to F represent the transverse diameters at the levels of the manubrium sterni, T7, and T11, respectively. G to H, I to J, and K to L represent the AP diameters at the levels of the manubrium sterni, T7, and T11, respectively.Figure 1. Thoracic diameter measurements on the AP and lateral CXRs. A to B, C to D, and E to F represent the transverse diameters at the levels of the manubrium sterni, T7, and T11, respectively. G to H, I to J, and K to L represent the AP diameters at the levels of the manubrium sterni, T7, and T11, respectively.

Figure 2. Thoracic diameter measurements on the chest CT scan. Lines 1, 2, and 3 represent midsagittal, left, and right AP diameters, respectively. Line 4 represents the transverse diameter.Figure 2. Thoracic diameter measurements on the chest CT scan. Lines 1, 2, and 3 represent midsagittal, left, and right AP diameters, respectively. Line 4 represents the transverse diameter.