Comparison of Measurement and Prognostic Power of SUV Between High-Definition and Standard PET Imaging in Non–Small Cell Lung Cancer Patients

Patient Recruitment

The institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived. We performed a retrospective review of the medical records of patients with NSCLC. We merged the list of PET/CT cases performed at the University of Chicago Medical Center between April 2018 and January 2021, using our Siemens Biograph mCT (Siemens Healthcare) scanner, with the list of NSCLC patients diagnosed and treated at our institution. This information was obtained from our cancer registry database. From these 2 sources, we identified 242 consecutive NSCLC patients for this study, based on the following inclusion criteria: all patients had undergone a pretherapy baseline PET/CT scan with our mCT, none of the patients had known brain metastasis (as our whole-body PET/CT scans did not cover the entire brain, usually just eyes to thighs), and the patients had no history or concurrent diagnosis of another type of cancer. The purpose of the PET/CT scan for this group of patients was to stage the disease or diagnose lung lesions.

PET/CT Imaging

The baseline¹⁸F-FDG PET/CT scans were performed on the study patients in accordance with National Cancer Institute guidelines (9). The patients fasted for at least 4 h before intravenous administration of 370–555 MBq of ¹⁸F-FDG. In addition, the serum glucose levels were tested via finger stick sampling before injection.

The ¹⁸F-FDG PET images were acquired on a Siemens Biograph mCT, equipped with lutetium oxyorthosilicate detectors and a 64-detector-row equivalent CT. The PET camera had a bed position length of 22.1 cm and a bore 78 cm in diameter.

A whole-body unenhanced CT scan with no intravenous contrast administration was performed first for PET attenuation correction and fusion of PET with CT images. We used a standard protocol for the CT with 120 kVp, 70 mAs (quality reference mAs for dose modulation), a transaxial field of view 50 cm in diameter, a tube rotation time of 0.5 s per rotation, and a pitch of 1.5.

Ninety ± 30 min after injection of the ¹⁸F-FDG, a whole-body static PET scan was acquired for about 15–20 min, starting at the thighs and proceeding to the head. PET scans were obtained with an acquisition time of 3 min per cradle position, with 45% bed overlap to avoid artifacts.

The reconstructions were performed using HD and SD PET for all study patients. For HD PET, reconstructions were performed with PSF plus ordered-subsets expectation maximization plus TOF, 3 iterations, 21 subsets, a 2-mm gaussian filter, and 1.85 × 1.85 × 2.03 mm voxels. The image matrix was 400 × 400 pixels. For SD PET, reconstructions were performed with the 3-dimensional (ordered-subsets expectation maximization) iterative algorithm plus TOF, 2 iterations, 21 subsets, a 4.5-mm gaussian filter, and 2.89 × 2.89 × 3.00 mm voxels. The image matrix was 256 × 256 pixels. In contrast to SD PET, the HD PET reconstruction algorithm incorporates a specific correction for the PSF. We used a 3-dimensional imaging mode with model-based scatter correction. Daily calibrations were conducted using a ⁶⁸Ge tub phantom.

Analysis of ¹⁸F-FDG PET Data

The SUV was calculated using Equation 1. (Eq. 1)

The SUV_max, metabolic tumor volume (MTV), and total-lesion glycolysis (TLG) for all tumor lesions throughout the body were measured by segmenting each tumor in the attenuation-corrected SD PET/CT images. This was done with the gradient method outlined in previous studies (1,10), using the PET Edge tool from MIMvista software. The measurements were conducted by one of the authors, a radiologist who has over 20 y of experience in clinical nuclear medicine and PET. He was aware of the clinical information but remained uninformed about the patients’ survival status and survival time. He used the MIMvista software along with our PACS to detect, characterize, and segment tumor lesions. This involved analyzing PET/CT images (including attenuation-corrected, maximum-intensity projection, and fusion images) and relevant diagnostic CT images and reports. The tumor segmentation was only performed on the SD PET/CT images, and the tumor contours were copied directly from SD PET to HD PET/CT images. The SUV_max MTV and TLG for each body tumor were recorded in a Microsoft Excel spreadsheet. The highest SUV_max across all lesions, termed SUV_maxwb, was identified. This comprehensive approach ensured no omission of the highest SUV_maxwb. Additionally, the percentage difference in SUV_maxwb was calculated, defined as Equation 2. (Eq. 2)

The whole-body MTV, which is the sum of all individual tumor MTVs, and the whole-body TLG, which is the sum of all individual tumor TLGs, were also calculated.

For calculating the tumor-to-background ratio (TBR) of all lesions, background SUV_mean was measured in specified areas in normal lung, right atrial blood pool, right liver lobe, and a lumbar vertebral body (diameters of the background regions were at least twice the full width at half maximum). These measurements facilitated the calculation of TBR for primary lung tumors and mediastinal nodal, soft-tissue, and bony metastases, respectively. TBR is defined as the ratio of tumor SUV_max to the SUV_mean of the corresponding background (10). The background SUV_mean of the lung, blood pool, liver, and bone was used for calculating the TBR of lung tumoral lesions, chest nodal metastasis, soft-tissue metastases (including liver, spleen, and muscle metastases), and bony metastases, respectively. The percentage difference in TBR was calculated as Equation 3. (Eq. 3)

Statistical Analysis

Statistical analyses were performed using Stata software version 14. The SUV_maxwb measurements in all study patients and the TBRs of all tumoral lesions of the patients were compared using a nonparametric Wilcoxon signed-rank test, since the data were not normally distributed. The comparison of background SUV_mean of normal lung, right atrial blood pool, normal liver, and bone on SD and HD PET was performed with paired samples t tests, as the data were normally distributed. The overall survival (OS) was considered the primary endpoint and defined as the time from baseline PET/CT to the time of any cause of mortality. Univariate and multivariate analyses were performed with Cox proportional-hazards regression models to test for an association with the primary endpoint, OS. A P value of less than 0.05 from a 2-tailed test was considered significant. Kaplan–Meier curves were plotted to visualize the time dependance of survival in the 2 patient groups stratified by the median of SUV_maxwb, and a log-rank test was performed to compare the curves. This was done for both HD and SD.