Application of spectral CT combined with perfusion scan in diagnosis of pancreatic neuroendocrine tumors – Insights into Imaging

Patients

This study was approved by the e Ethics Committee of The First Affiliated Hospital of Zhengzhou University. All patients in this research were informed and provided written consent. 58 patients with clinically suspected pNETs were prospectively enrolled in the study for one-stop spectral CT combined with perfusion scanning from December 2018 to December 2020. Suspected pNETs were referred as patients with symptoms such as epigastric pain, pancreatitis, obstructive jaundice, intractable hypoglycemia or splenomegaly and wasting and was initially diagnosed as pNETs by experienced clinicians. The inclusion criteria were as follows: (a) the CT images of the patients did not have severe motion artifacts or any metal artifacts that affect the assessment of the lesion; (b) patients did not undergo relevant treatment affecting perfusion and energy spectrum parameters before surgery; and (c) Patients with pathologically confirmed pNETs with complete clinical information. Exclusion criteria were: (a) the patient was allergic to iodine contrast agents during pregnancy; (b) presence of severe hepatic and renal insufficiency; (c) poor vascular status of the patient and inability to tolerate high speed intravenous injection. With these criteria 36 pNETs patients were recruited. An independent cohort of 30 patients with pathologically confirmed pNETs who underwent routine pancreatic perfusion scanning in our hospital during the same period were retrospectively collected, the same inclusion and exclusion criteria applied to this cohort.

CT imaging technique

Scans were performed on a GE Revolution CT machine. Patients received iv injection of non-ionic iodine contrast agent (350 mgI/ml, Iopamidol injection, HengRui Healthcare, Jiangsu, China). Energy spectrum combined with perfusion scan method: upper abdominal plain scan plus conventional pancreas perfusion and dual-phase enhanced energy spectrum scan (arterial phase and venous phase). Scanning range: (1) Upper abdominal CT plain scan: from the top of the diaphragm to the level of the umbilicus; (2) Perfusion scan range: the upper and lower boundaries of the pancreas (determined according to the plain scan of the upper abdomen), that is, from the level of the first hepatic hilum to the level of the duodenum; (3) Two-phase energy spectrum scanning range: from the top of the diaphragm to the level of the lower border of the liver. Scanning parameters: (1) Conventional upper abdomen plain scan: helical scan, tube voltage 120 kVp, tube current 100–450 mAs, pitch 0.992:1, scanning slice thickness 5 mm, and slice spacing 5 mm. (2) Perfusion parameters: axial scan, automatic tube voltage 100 kVp, automatic tube current 100 mAs, 50ASIR-V, and the other settings were same as the plain scan. 3) Energy spectrum enhancement parameters: helical scan, tube voltage 80 kVp and 140 kVp, high and low tube voltage was switched instantaneously (25 ms), tube current automatic adjustment, and the other settings were same as the plain scan (Table 1). The scanning phase was divided into 4 parts, the 1st and 3rd parts were perfusion inflow phase and outflow phase, of which 8 images were collected in the inflow phase and 14 images were scanned in the outflow phase; in the spectral enhancement phase, the 2nd phase was the arterial phase and the 4th phase was the venous phase, and each image was collected once; a total of 24 images were collected. After the localization image was obtained, a routine upper abdominal scan was performed first, and the parameters of the perfusion combined with energy spectrum were set. The non-ionic iodine contrast medium was injected through the anterior cubital vein indwelling needle (20G Y-II type) using a double-barrel high-pressure syringe, and then 20 ml of saline was added with bolus tracking to ensure that all of the contrast medium reached the central veins, serving as a bolus chaser to increase peak arterial enhancement. Contrast agent injection: with the use of a power injector the flow rate was 5 mL/s (depending on the patient’s blood vessel tolerable flow rate) with bolus tracking, and the dose was 1 ml/Kg. The perfusion scan started 6 s after contrast injection, with 8 consecutive perfusion scans of 3 s each, followed by an energy spectrum enhancement CT scan of 3 s, then 14 consecutive perfusion scans of 1.5 s each, and when 54 s after contrast injection the last energy spectrum enhancement CT scan of 3 s. After scanning, images were reconstructed: the energy spectrum scanning range was adjusted to the perfusion range, and the original 22 perfusion images and 2 energy spectrum enhanced images were fused to reconstruct the perfusion image.

Conventional pancreas perfusion scan method: continuous pancreas perfusion scan after upper abdominal plain scan. The perfusion scanning adopted axial scanning, tube voltage 120 kV, current 150 mA, and pitch 1.375:1. Contrast agent was injected intravenously at 3–4.5 ml/s at a dose of 1 ml/kg, and 20 ml of saline was added at the same time; perfusion images were continuously collected for 25 times.

Imaging analysis and measurement

Analysis and measurement of the images were performed by two independent radiologists on professional workstation GE ADW4.7, both with intermediate title or above; final decision was made by consultation with a third senior radiologist in case of disagreement; the histopathological nature of each tumor was unknown to both radiologists. Perfusion images were fused and reconstructed (thickness 1.25 mm), and then sent to the post-processing workstation for analysis and measurement. The reconstructed perfusion image was subjected to automatic motion artifact correction (CT Dynamic Registration), and then loaded into CT perfusion 4D software for pancreatic perfusion analysis. The abdominal aorta was used as the inflow artery to calculate and generate a pseudo-color map of pancreatic perfusion. A region of interest (ROI) was delineated at the normal pancreatic parenchyma to obtain the perfusion parameters of the corresponding tissue, including blood flow (BF), blood volume (BV), capillary surface permeability (capillary surface permeability), permeability of surface (PS), mean transit time (MTT), time to peak (TTP), transit time to pulse peak (transit max, Tmax), mean slope of increase (MSI), contrast agent arrival delay (idoine remaining function time 0, IRF T0). Perfusion parameters between tumor lesions and normal parenchyma as well as between different grades of pNETs were compared.

Energy spectrum images in the arterial and venous phases were loaded into the GSI viewer General software, images at the axial Mono interface 70 keV were used to create ROIs in the abdominal aorta, pancreatic lesions, and adjacent normal parenchyma, respectively. Iodine-based map, water-based map and energy spectrum curve were automatically generated, iodine concentration (IC) and water concentration (WC) of the corresponding parts, normalized iodine concentration (NIC, ratio of the IC of the lesion to that of the abdominal aorta at the same layer), and the CT value of a single energy level at 60 keV to 140 keV in the lesion and adjacent pancreatic parenchyma (with 10 keV intervals) were recorded. The longest diameter of the lesions was measured and recorded on the spectrally enhanced images. ROIs were placed at the solid part of the lesions, avoiding necrotic area, calcification, and blood vessels; the scope of ROIs were kept consistently as possible; all data were measured 3 times and averaged to reduce errors. pNETs detection rate is defined as: total number of lesions detected in imaging/total number of lesions detected in histopathology.

Scanning radiation dose

After each scan, the system automatically generated the radiation metering parameters, recorded dose-length product (DLP, mGy × cm), and calculated effective dose (ED, mSv), ED = k × DLP, where k = 0.015 (mSv × mGy⁻¹ × cm⁻¹) (k is the conversion factor for adult abdominal CT examination).

Statistical analysis

Statistical analysis of the recorded data was performed using SPSS 21.0 package (SPSS Inc, Chicago, IL). Data that conformed to the normal distribution were expressed as mean ± standard deviation, those that did not conform to the normal distribution were expressed as median (95% confidence interval). Paired sample t test was used to compare the difference between tumor and normal tissues. Independent sample t test or nonparametric test was used to compare the differences of perfusion parameters at different grade. p < 0.05 was considered statistically significant.