Patients
Between September 2007 and June 2011, 202 patients were referred to our institution for the execution of multimodal PET/CT, including both non-enhanced and contrast-enhanced diagnostic CT. Clinical indication comprised diagnosis of suspected oncologic disease, diagnosis of a suspected recurrence, and staging, restaging, or post-therapy evaluation of a known oncologic disease. All patients accepted the use of imaging data for research purposes by providing written informed consent that was approved by the local regulatory bodies. Patient characteristics are summarized in Table 1. In these patients, the indications for adding a multiphase ceCT to the standard PET/CT exam protocol were: (1) new patients with no state-of-the-art, whole-body staging examination available; (2) equivocal or insufficient results of previous examinations; (3) precise assessment of tumor extent before local radiation therapy; (4) post-therapy assessment of an inpatient who required restaging with both ceCT and PET/CT.
Table 1
Patient characteristics
Positron emission tomography/computed tomography scanning
Patients were prepared and standard PET/CT exams were performed according to the European guidelines [11]. Briefly, patients fasted overnight prior to the intravenous administration of FDG; this was performed in a quiet room, with the patient lying in a recumbent position and asked not to move. Blood glucose was measured before tracer injection was administered, to ensure blood glucose levels < 160 mg/dl. The dose of FDG varied between 350 and 450 MBq, depending on the patient’s weight, and was injected through a peripheral vein catheter. To minimize artifacts caused by the presence of radioactive urine in the excretory system, patients were asked to drink 500–1000 ml of water 1 h prior to image acquisition and to void just before the scan. No urinary bladder catheterization was used. Whole-body imaging was performed using a combined PET/CT scanner (BioGraph 16 Hi-Rez PET/CT scanner; Siemens AG, Erlangen, Germany). The technical parameters of the 16-detector row, helical CT scanner included a gantry rotation speed of 0.5 s and a table speed of 24 mm per gantry rotation. The PET component of the combined imaging system had an axial view of 16.2 cm (per bed position), with an interslice spacing of 3.75 mm in one bed position. The transaxial field of view and pixel size of the reconstructed PET images were 58.5 cm and 4.57 mm, respectively, with a matrix size of 128 × 128. Data acquisition started 60 ± 10 minutes after intravenous tracer administration. First, unenhanced, low-dose CT was performed at 140 kV and 40 mA for emission-based attenuation correction, immediately followed by a PET scan, which was executed in three-dimensional (3D) mode, with a 3-min acquisition per bed position. The scan was performed starting from the orbital plane on to the mid-thigh, except for those cases where the clinical history demanded a whole-body, head-to-toes scan (e.g. multiple myeloma or melanoma). Attenuation-corrected PET images were reconstructed by means of an ordered-subset expectation maximization, iterative reconstruction algorithm (three iterations, eight subsets). Finally, diagnostic ceCT was performed for the same axial coverage. The entire ceCT data set was automatically fused with the 3D PET images using the integrated software interface provided by the manufacturer (syngo Image Fusion; Siemens AG, Erlangen, Germany) to create contrast-enhanced anatomical images superimposed with FDG uptake. We did not experience significant fusion mismatch as the ceCT was performed immediately after the PET emission scan, while the patient was maintaining the same position.
Contrast-enhanced computed tomography technique
The CT scan was performed immediately after completion of PET acquisition, planned with the same scout view. In most cases, a pre-contrast diagnostic scan was not acquired and the standard acquisition protocol consisted of an arterial phase scan of the upper abdomen, starting 35 s after the start of contrast injection, followed by a portal phase scan, extended from the skull base to the symphysis pubis, starting 70 s after the administration of the intravenous contrast. The scan parameters were as follows. Arterial phase: slice thickness of 5 mm, pitch 0.8, tube rotation speed 0.5 s, 120 kV, reference 175 mA. A dose modulation system was applied to optimize total exposure according to the patient’s body size; an additional set of 1-mm thick slices was reconstructed to obtain high-resolution, multiplanar reformations. Portal phase: slice thickness 5 mm, pitch 0.8, tube rotation speed 0.5 s, 120 kV, reference 175 mA with the same modulation system; 2-mm thick slices at 1.5 mm intervals were reconstructed for multiplanar reformations.
In selected cases, a delayed scan was performed at equilibrium or in the urographic phase, according to the clinical question. In the follow-up of known lesions or in patients with lymphoproliferative disorders and with previous negative imaging examination, only the portal phase scan was obtained, so as to reduce the dose administered to the patient.
Iodinated contrast medium, with a concentration of 350 mg/ml, was injected using a power injector at a flow rate of 3 ml/s and a dose of 80–130 ml, depending on body weight, followed by 40 ml of saline at the same flow rate. Contrast medium administration protocols and measures aiming to prevent contrast-related adverse effects were performed according to the guidelines of the European Society of Urogenital Radiology [12].
Standard, 5-mm thick images were used for rapid evaluation by the radiologist and for reviewing by the referring physician, while thinner slices were used for multiplanar imaging of vessels, bone (ribs and spine), and for high-resolution scanning of lung and liver lesions.
Image interpretation
Positron emission tomography, ceCT, and fused images were reconstructed for review on a dedicated computer workstation (syngo Image Fusion; Siemens AG, Erlangen, Germany).
Initially, PET/low-dose CT and multiphase diagnostic CT images were evaluated independently by two experienced nuclear medicine specialists and by two experienced CT radiologists, respectively. All readers had access to the patient’s clinical history, but were blinded to the other modality results. Subsequently, fused, cePET/CT images were evaluated in consensus, by the combined team of radiologists and nuclear physicians. This latter consensus report was compared with the previous evaluations to determine the additional value of diagnostic CT on PET/CT image interpretation.
In the evaluation of FDG-PET/CT, a lesion was considered positive whenever it showed a non-physiological increase of FDG uptake. In particular, the diagnosis of a PET-positive lesion was also supported by a maximum standardized uptake (SUVmax) value of at least 2.5 or by an FDG uptake exceeding the surrounding background tissue, the blood pool radioactivity, or the average liver uptake. However, the differentiation between malignant and benign lesion was not based solely on SUVmax, as the qualitative assessment of increased FDG uptake areas played a major role in the clinical reporting. For instance, if a lesion showed clearly abnormal focal FDG avidity but displayed a SUVmax lower than the 2.5 threshold (e.g. due to partial volume effect in small-sized lesions), that lesion was deemed malignant.
On diagnostic ceCT-only images, detection of a pathologic lesion was performed according to published criteria [13–16]. For example, lymph node (LN) assessment in the neck, thorax, and abdomen was based on size criteria (1-cm short-axis diameter threshold). However, the presence of peripheral low attenuation, suggesting a fatty hilum within a LN, was considered a benign sign, regardless of node size. Finally, criteria such as abnormal enhancement, central necrosis, irregular borders, and the presence of infiltrative behavior were used to characterize soft tissue lesions. Consensus interpretation of the PET/ceCT exam was performed according to the following criteria: (1) positive lesions were diagnosed when an abnormal area of focal FDG uptake, as observed in PET images, corresponded to an abnormal finding on CT; (2) LNs with increased glucose uptake were deemed positive for metastatic spread even if they were smaller than 1 cm in short-axis diameter; (3) conversely, LNs with no detectable tracer uptake were deemed negative for metastatic spread, even if they were larger than 1 cm in short-axis diameter; (4) in all the remaining cases, the two readers decided to emphasize either functional information from the PET or morphological information from the ceCT on a case-by-case basis, according to type and site of disease and in relation to the patient’s clinical history. For example, in cePET/CT images, small pathological lesions, such as millimetric lung metastases, which often lack FDG accumulation, were evaluated on the basis of the ceCT results.
Standard of reference
The final diagnosis was obtained from the results of the histopathologic examination, obtained following surgery or biopsy (73 patients, 147 lesions), or clinical/radiological 12-month follow-up or, again, on the basis of tumor marker levels or on the evolutionary pattern of known findings at subsequent imaging (129 patients, 450 lesions). Follow-up information included physical examination, laboratory tests, tumor markers, other independent imaging studies, such as multislice CT, magnetic resonance imaging, FDG-PET/CT, 18 F-NaF PET/CT, X-ray studies, and bone scans.
The criteria used as the standard of reference were: (1) laboratory findings, such as increasing tumor markers; (2) combination of either negative follow-up imaging findings and negative clinical findings or positive clinical findings with decreasing lesion size after therapy, as determined by subsequent imaging studies; (3) increasing lesion size or metabolic activity in the course of follow-up; (4) subsiding of pathological findings on follow-up PET/CT studies combined with negative clinical follow-up.
Statistical analysis
Statistical analysis was carried out using the ‘R’ software program [17] and the DiagnosisMed software package [18].
We performed patient-based and lesion-, site-, tumor-, and question-based analyses of the cePET/CT results compared with PET/CT. Tables 1 and 2 list the tumor type, clinical question, and site of disease that were used in the tumor-, question-, and site-based analyses.
Table 2
Distribution of lesions
A lesion was included in the analysis if at least one of the three modalities (cePET/CT, ceCT, PET/CT) deemed it positive or if it was considered as non-pathologic at these three modalities but resulted positive at a subsequent histopathologic analysis.
Sensitivity, specificity, positive predictive value, negative predictive value (NPV), accuracy, likelihood ratios, diagnostic odds ratio, error rate, and Youden’s index were calculated using standard statistical formulae, and the 95% confidence interval was determined for each parameter. Differences among imaging modalities were assessed with the Cochran Q test, followed by multiple comparisons using the McNemar test with continuity correction and Bonferroni adjustment. Probability values inferior to 0.05 were considered as statistically significant. The value of adding a diagnostic CT to the standard PET/CT protocol was also evaluated in terms of the impact on patient management, following the information derived from the cePET/CT exam only. This evaluation was performed on a patient basis for each tumor type and clinical question.
In order to better characterize the additional value of ceCT with respect to the low-dose PET/CT study, the management-changing findings were classified into the following categories: (1) metabolism-related; (2) site-related; (3) dimension-related; (4) related to local infiltration or additional findings on ceCT.
Category 1 referred to the improved characterization of a lesion that presented no increment in FDG uptake or whose uptake did not reach the significance threshold. Category 2 encompassed the scenario of a more accurate interpretation of lesions due to improved localization of abnormal FDG uptake, which also allowed a better distinction between abnormal and physiological FDG uptake. Category 3 was applied to the cases of identification of pathologic lesions with size falling below PET image resolution. Finally, category 4 referred to the evaluation of infiltrative behavior or to meaningful findings that had gone undetected at the standard PET/CT scan.