Safire

Image datasets were acquired in spiral (helical, pitch factor 0.9) technique using a regularly calibrated 64-detector row 128 slice CT system (Somatom Definition Flash, Siemens Healthineers, Forchheim, Germany). Each dataset acquired was reconstructed with both FBP (B40f kernel, medium-soft) and corresponding IR (I40f kernel, medium-soft) at 0.75 mm slice thickness, 128x0.6 mm collimation, 0.6 mm increment, 512x512 pixel matrix, and 300×300 mm² field-of-view. The medium-soft kernel was chosen as it is recommended for use in densitometry [19 (link), 20 (link), 22 (link), 23 (link)]. For IR, a raw-data based algorithm (Safire, Siemens Healthineers, Forchheim, Germany) was used and “strength” 3 of 5 was selected as previously described [13 (link), 19 (link), 20 (link), 24 (link)]. With ten lung explants, eight dose levels, and two reconstructions per acquisition, 160 datasets were obtained for this comparative study. Exposure settings were chosen to represent “moderate” to “low-dose” chest CT protocols as currently used in clinical routine, and to challenge IR by acquisitions at very low dose (Table 1). Of note, automatic exposure control (AEC) was disabled to ensure homogenous exposure of the whole scanning volume. Image data can be made available by the authors upon request.

At baseline (prior to treatment initiation), all patients underwent a clinically indicated 18F-FDG-PET/CT scan (Biograph mCT; Siemens Healthcare GmbH, Erlangen, Germany). Patients fasted for at least six hours prior to the injection of 18F-FDG (mean dose: 311.71 +/- 12.72 MBq). The uptake time for the 18F-FDG-PET/CT was 60 min. During PET/CT, a diagnostic contrast-enhanced CT (Ultravist 370; Bayer Healthcare) of the whole body (head to thighs, portal venous phase, expiratory breath-hold) was acquired. An additional thorax scan (inspiratory breath-hold) for the detection of lung metastases was performed. CT acquisition and reconstruction parameters were as follows: 120–140 kV; reference dose: 200 mAs; pitch: 0.7; slice collimation: 128 × 0.6; gantry rotation time: 0.5 s; iterative reconstruction (Safire^®, Siemens Healthcare GmbH, Erlangen, Germany) kernel B70 or I31f; slice thickness: 3 mm; increment: 2.5 mm; image resolution: 0.9 × 0.9 × 3 mm³; matrix size: 512 × 512× 340–1000 (depending on the patient’s size). Whole-body PET scans were performed with a usual scan time per bed position of 2 min. The PET reconstruction parameters were: 3D OP-OSEM time-of-flight reconstruction; 2 mm Gaussian filter; matrix size: 400 × 400. For the attenuation correction, the whole-body CT scan was used.

The patient was in the supine position. After breath holding at the end of inhalation, the dual-energy plain scan and dual-energy enhanced scan in AP and VP were performed from the thoracic inlet to the bottom of the lung.
CT scans were performed in DE mode on a second-generation dual-source CT scanner (SOMATOM Definition FLASH, Siemens Healthcare, Germany). After unenhanced CT was performed, 350 mg I/mL of nonionic iodinated contrast agent (Ioversol) at a dose of 1.2 mL/kg weight was injected into an antecubital vein together with 20 mL of saline at rates of 3 mL/s and 4 mL/s, respectively. AP and VP dual-energy contrast-enhanced CT images were obtained after post-injection delays of 30 and 60 s, respectively. The scan parameters for the DECT mode were summarized as follows. The tube voltages of A and B were set at 100 kVp and 140 kVp, respectively, with a real-time adjustable variable tube current. Collimation was 128 × 0.6 mm; rotation speed was 0.28 s/r; gantry rotation was 330 ms; slice thickness was 5 mm. Finally, 100 kVp and 140 kVp images were acquired in the arterial and venous phases, respectively, and 120 kVp equivalent mixed images were generated (linear fusion coefficient, 0.4). These images were reconstructed with a slice thickness of 1 mm and an interval of 1 mm using iterative reconstruction software (SAFIRE, Siemens Healthcare, Germany).

Dedicated iterative reconstruction technique was used for second-generation (Safire) and third-generation (Admire) DSCT (Siemens Healthineers, Forchheim, Germany) at a strength level of 3 of 5 with a medium smooth reconstruction kernel (B30f for second-generation; Br40 for third-generation). All data were calculated as transverse and coronal images with a slice thickness of 2.0 mm and increment of 1.0 mm. Table 1 summarizes all image acquisition data.

In both the low- and high-iodine groups, CTA images were reconstructed by a conventional filtered back projection (FBP) algorithm with a medium smooth kernel designed for cardiac imaging (B26f). In the low-iodine group, images were also reconstructed by a sinogram-affirmed IR algorithm (SAFIRE, Siemens Healthcare) with the corresponding vascular kernel (I26f). With the IR algorithm, five adjustable strength settings (strength 1–5) were available for adaptation of the noise model (SAFIRE). As recommended by manufacturer, a medium strength of 3 was used.
In both groups, transverse images were reconstructed with a slice thickness of 1 mm in 1-mm increments. Patient information was removed from all images, which were transferred to an external workstation (Syngo Multi-Modality Work Place, CT 2011A, Siemens Healthcare) for further image analysis. Using the axial data, two cardiac radiologists with more than 8 years of experience in cardiac imaging reconstructed the images by volume rendering technique, maximum intensity projection, and multiplanar reconstruction (Fig. 1).

From the ECG-synchronized CT raw datasets, morphological images were generally reconstructed through the entire chest and abdomen at 70% of the R-R interval with 1-mm slice thickness, 0.7-mm increments, I26f reconstruction kernel and 32 cm FOV. Functional imaging was then reconstructed in 5% steps between 0 and 95% of the R-R interval from the ECG with 1-mm slice thickness, 0.7-mm increments, I26f reconstruction kernel and 24 cm FOV. All datasets were reconstructed with sinogram affirmed iterative reconstruction (SAFIRE, Siemens Healthcare, Forchheim, Germany) and a medium strength level of 3 was used in all patients. All datasets were transferred to a Siemens Workstation (Syngo Workplace; Siemens Medical Solutions) for image analysis.