Somatom sensation 64

Patients generally underwent contrast-enhanced CT of the entire abdomen, which is from the dome of the diaphragm to the pubic symphysis according to the standard clinical scanning protocols (tube voltage of 120 kV, tube current of 180–200 mAs, slice thickness of 3 mm, slice interval of 3 mm, field view of 35–40 cm, matrix of 256 × 256, and pitch of 0.6–1.0) on CT machines (Somatom Definition AS 128; Somatom Sensation 64; Light speed 128, GE Healthcare). All CT scans were reconstructed into slices of 3-mm thickness and interval. After fasting from food for at least 6 h, each patient ingested about 600–1000 ml of water for more than 15 min to achieve gastric distension before scanning. All patients were trained to hold their breath in the supine position during CT scanning. After the plain CT, a 1.5 ml/kg iodinated contrast agent (ioversol 320 mgI/mL, Jiangsu hengrui) was injected with a pump injector (OptiVantage DH, Tyco) at a flow rate of 3 ml/s into the dorsal hand vein. The CT images were obtained during the arterial phase (25–30 s after the initiation of the injection), the portal venous phase (60–70 s), and the delay phase (180 s). The acquired images were analyzed on a PACS station (GE Healthcare).

Preoperative chest CT images of 263 NSCLC patients who underwent curative surgical resection were acquired from the Veterans Health Service Medical Center, Seoul, South Korea. The CT images were obtained from three different CT scanners (SIEMENS SOMATOM Definition AS+, SIEMENS SOMATOM Sensation 64, GE Healthcare LightSpeed ULTRA) using the following scan parameters: 99~129 kVp at 60~534 mAs. Each image had a matrix size of 512 × 512 pixels with in-plane resolutions ranging from 0.54 to 0.83 mm. The slice thickness ranged from 1.0 to 7.5 mm. This study was approved by the Institutional Review Board of the Veterans Health Service Medical Center (IRB File Number: BOHUN 2018-07-009-005). In the acquired data, patient selection was applied, as shown in Figure 3. This study excluded patients without information pertaining to recurrence, those at the T stage and N stage, and patients whose tumors are incorrectly labeled as too large. In addition, according to the opinion of clinicians that it is better to exclude data with large deviations, patients with a tumor larger than 19 cm based on the representative cross-section were also excluded. Finally, the remaining 217 NSCLC patients served as the study group. The patient characteristics are shown in Table 1.