Evaluation of optimal monoenergetic images acquired by dual-energy CT in the diagnosis of T staging of thoracic esophageal cancer

Patient inclusion

This retrospective study was approved by the ethics committee of our hospital. The need for written informed consent was waived. All consecutive patients who had chest DECT and endoscopy were recruited from September 2019 to December 2020. Inclusion criteria were as follows: (i) single lesions located in the thoracic portion; (ii) all patients diagnosed with EC by endoscopy and biopsy; and (iii) contrast-enhanced ultrasound including arterial phase and venous phase DECT of the chest. Exclusion criteria were as follows: (i) clinical data missing or incomplete; (ii) second tumor besides EC; (iii) radiotherapy or chemotherapy treatment before DECT; and (iv) poor image quality on DECT. After applying these inclusion and exclusion criteria, 68 patients with EC were analyzed for tumor visualization on MEI (+) and PEI objectively and subjectively; of these 49 patients received radical surgery and obtained pathologically confirmed T staging. The workflow chart is shown in Fig. 1.

Fig. 1figure 1

Flowchart showing overview of patients inclusion and exclusion

Patient clinicopathological data were collected, which included age, sex, tumor location, tumor histology, pathological grading, clinical TN staging, and pathological TN staging. The clinical T and N staging was established by DECT. T staging was performed according to the International Union Against Cancer/American Joint Committee on Cancer (UICC/AJCC) TNM classification for EC (7th edition, 2011), details as follows: T1-tumor invades lamina propria, muscularis mucosae or submucosa; T2-tumor invades muscularis propria but without breaking through muscularis propria; T3-tumor invades adventitia; T4a-tumor invades pleura, pericardium, azygos vein, diaphragm or peritoneum; and T4b-tumor invades other adjacent structures, such as the aorta, vertebral body, and trachea.

DECT image acquisition

All patients were scanned using a 64-detector CT scanner (SOMATOM Drive, Siemens Healthineers) in dual-energy mode through two X-ray tubes with different kV tube voltages (tube A, 100 kV; tube B, Sn 140 kV), using a tin filter for the high-voltage tube. Automatic exposure control (CARE Dose 4D, Siemens Healthineers) was used in all scans. The parameters of scanners were as follows: collimation, 64 × 0.6 mm; rotation time, 0.28 s; pitch, 0.55; reference tube current time product, 71 mAs for the 100 kV tube and 60 mAs for the Sn140 kV tube; reformatted section thickness, 1.5 mm; reformatted section increment, 1.5 mm.

First, all patients were scanned non-contrast DECT images of chest. Then, acquired contrast-enhanced images and iodinated nonionic contrast media (ioversol, Hengrui Medicine) were administered through the ulnar vein at a dose of 1.5 mL/kg with a flow rate of 2.5 mL/s, followed by a bolus injection of 30 mL of saline at the same flow rate. The arterial phase was acquired after the injections 10 s (average, 35 ± 5 s). The scan delay time for the venous phase scanning was 25 s after the end of the arterial phase scanning (average, 60 ± 7 s).

DECT image reconstruction

Reconstructed DECT image data were post-processed on syngo.via workstation (VB20A, Dual Energy, Siemens Healthineers). The MEI (+) images were reconstructed at 40, 50, 60, 70, and 80 keV levels, and the PEI was reconstructed by applying the blending factor of 0.4 (M_0.4; 40% of the low kV and 60% of the high kV spectrum).

Subjective image analysis

The images were analyzed on MEI (+) images (40–80 keV) and PEI independently by two radiologists with 3 and 5 years of chest CT experience, respectively. Two readers individually rated each image series regarding the following categories using 5-point Likert scale: (i) image sharpness (ranging from 1 = distinct blurring to 5 = no apparent blurring); (ii) image noise (defined as image graininess: ranging from 1 = extensive image noise to 5 = no apparent noise); (iii) lesion margin (demarcation of lesion margins: ranging from 1 = no visual demarcation to 5 = perfect demarcation of contours); and (iv) lesion inside (the definition of cystic necrosis inside the lesion: 1 = nondiagnostic, 2 = poor, 3 = sufficient, 4 = good, 5 = excellent).

Objective image analysis

Objective image analysis was also performed on five sets of MEI (+) and PEI. The first region of interest (ROI) was located in the primary EC at the maximum diameter without areas of apparent cystic necrosis, blood vessel and air. The second ROI was located in the normal esophageal wall; and the last ROI was located background of the air. The average size of the three ROIs was 5–10 mm2. The position and size of the ROIs were kept constant in all sets of MEI (+) and PEI in both the arterial and venous phases. The mean attenuation (Hu) of EC lesion and normal esophageal wall and the standard deviation (SD) of the air were recorded; then, the SNR and CNR of EC lesions were calculated according to the following formulas:

$$} = }_})}} - }_})}} /}_})}}$$

$$} = }_}} \right)}} /}_}} \right)}}$$

Subjective evaluation of T staging

Subjective evaluation of T staging was performed in EC patients who received radical surgery. Two radiologists with 5 and 3 years of experience, who were blinded to the histopathological data, reviewed the optimal MEI (+) and the PEI to evaluate the T staging according to the UICC/AJCC TNM classification for EC (7th edition, 2011) independently. When the two reader's assessment of T staging appears inconsistent, they would discuss to achieve a consensus result.

Multiparameter differential T staging

Multiple quantitative parameters were measured to differential T1–2 from T3–4 staging, including: (i) the effective atomic number (Zeff) of non-contrast image; (ii) the attenuation (Hu) in arterial and venous phases of the optimal MEI (+); (iii) the normalized iodine concentration (NIC) obtained by iodine concentration of lesions that divided the iodine concentration of the aorta in arterial and venous phases; and (iv) electron density (Rho) in arterial and venous phases.

Statistical analysis

The statistical analyses were performed using software (IBM SPSS software, version 23). The data distribution was assessed using the Kolmogorov–Smirnov test. Subjective Likert scores and CNR and SNR were compared using the Wilcoxon test with adjustment for multiple comparisons, where applicable. The Kappa concordance test was used to evaluate the interobserver agreement of subjective Likert scores; and a kappa value ≤ 0.20 indicates poor agreement, 0.21–0.40 is fair, 0.41–0.60 is moderate, 0.61–0.80 is good, and 0.81–1.00 is excellent. The agreement of T staging between MEI (+) and PEI with those assigned after postoperative histopathologic examination was calculated. For all multiple quantitative parameters analysis to identify T1-2 and the T3-4 in EC patients, the area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated at maximal Youden’s index. The level of significance was set at p ≤ 0.05.

留言 (0)

沒有登入
gif