| (Purpose]1.To compare the application values between the early dynamic 18F-FDG PET/CT (ED 18F-FDG PET/CT), conventional 18F-FDG PET/CT (whole-body static 18F-FDG PET/CT, WB 18F-FDG PET/CT) and one-stop 18F-FDG PET/CT (combined ED 18F-FDG PET/CT with WB 18F-FDG PET/CT) in detecting hepatocellular carcinoma (HCC).2.To compare the application values between the early dynamic 18F-FDG PET/CT (ED 18F-FDG PET/CT), conventional 18F-FDG PET/CT (whole-body static 18F-FDG PET/CT, WB 18F-FDG PET/CT) and one-stop 18F-FDG PET/CT in differentiating malignant from benign focal liver lesions.[Materials and Methods]1. Study objectsFrom April 2013 to September 2014, a total of 56 patients with focal liver lesions who underwent PET/CT examination in our PET center and eventually confirmed by pathological diagnosis or by clinical diagnosis as well as over 6 months imaging follow-up were prospectively enrolled. Among these patients, there were 35 patients with HCC,5 patients with intrahepatic cholangiocarcinoma (ICC), 2 patients with hepatic metastasis,7 patients with hepatic hemangioma,4 patients with liver abscesses,1 patient with liver cirrhosis regenerative nodules,1 patient with focal nodular hyperplasia and 1 patient with hepatic angiomyolipoma.1.1 The group of hepatocellular carcinomaA total of 52 HCC lesions from 35 patients who diagnosed by histopathology or by primary liver cancer diagnosis standard of the People’s Republic of China (PRC) (2011) were included in this group. Among these patients,16 cases were diagnosed by surgical pathology,3 were diagnosed by biopsy pathology and 16 were diagnosed in accordance with primary liver cancer diagnosis standard of the PRC. Of the 35 HCC patients,23 patients each had 1 HCC lesion,7 patients each had 2 HCC lesions and 5 patients each had 3 HCC lesions.1.2 The group of other malignancyA total of 10 lesions from 7 patients with other liver malignant tumor who diagnosed by histopathology were included in this group. Among these patients,4 cases were diagnosed by surgical pathology,3 were diagnosed by biopsy pathology. Of the 7 patients,4 patients with ICC each had 1 liver tumor lesion,1 patient with ICC had 2 liver tumor lesions,1 patient with liver metastasis of tongue carcinoma had 3 liver tumor lesions,1 patient with liver metastasis of carcinoma the floor of mouth had 1 liver tumor lesion.1.3 The group of benign lesionsA total of 18 lesions from 14 patients with liver benign lesions who confirmed by histopathological diagnosis or by clinical diagnosis as well as over 6 months imaging follow-up were included in this group. Among these patients,4 cases were diagnosed by surgical pathology (including one each case of hepatic abscess, liver cirrhosis regenerative nodules, focal nodular hyperplasia of the liver and hepatic angiomyolipoma),10 cases confirmed by clinical diagnosis and over 6 months imaging follow-up (including 7 cases of hepatic hemangioma and 3 cases of hepatic abscess). Of the 14 patient with liver benign lesions,1 patient with hepatic hemangioma had 3 lesions,2 patients with hepatic abscess each had two lesions, and the remaining 11 patients each had 1 lesion.2. Main instrument and imaging agentThe imaging instrument was Biography mCTx PET/CT Scanner (Siemens company of Germany), as PET scanner was 52-ring LSO crystal, and CT was of 128 layers. The imaging agent was 18F-FDG, as 18F was derived from GE PET trace accelerator (GE company, USA), while 18F-FDG was automatically compound by chemical compound mould (PET Biotechnology Limited Company of Beijing, China), with a radiochemical purity> 95%.3. Method and conditions for imagingFasting patient over 6 h, after checking blood-sugar and weighing, a bed position of ED 18F-FDG PET/CT of liver-center was conducted, then after supine rest, a WB 18F-FDG PET/CT was conducted.3.1 ED18F-FDG PET/CTA bed position of low-dose CT (120 kV,50 mAs) was conducted before injection of imaging agent so as to scan as much as possible the entire liver, and at the moment of bolus injecting 18F-FDG, a dynamic PET scan was immediately made lasting 1 min.3.2 WB 18F-FDG PET/CTA conventional whole body CT scan (120 kV, 100mAs) was made 60 min after bolus injection of 18F-FDG, scanning from skull to upper part of leg, around 6-8 bed positions and each of them acquired for 1.5 min.4. Imaging reconstruction and fusionPET and CT imaging were both reconstructed by standard iteration method (maximum likelihood ordered subsets), the CT reconstruction matrix was with a pixel of 512 × 512, a thickness of 3 mm and an interval of 2 mm; the reconstruction of PET imaging had undergone attenuation correction by using CT imaging, with a reconstruction matrix of a pixel of 200 × 200. Early dynamic PET data reconstructed a frame every 15s and gained 4 frames of image in total. Conventional whole-body PET imaging data reconstruction was 1 frame of image. And then PET imaging and CT imaging were sent to the Syngo MMWP work station of Siemens and underwent imaging fusion frame by frame by using TrueD software.5. PET/CT imaging analysis5.1 Semi-quantitative analysisThe lesion was visual-observed confirmed by two senior doctors with PET/CT diagnosis experience. By adopting region of interest (ROI) technology, delineated the borderline of lesion and calculated the maximal Standardized Uptake Value (SUVmax) automatically by the computer. In respect of lesions with indefinite radioactive concentration, ROI was set in accordance with CT imaging. In order to compare the difference on the radioactivity between HCC and surrounding tumor-free liver tissue, we divided a corresponding ROI in the tumor-free liver tissue of each patient. During the process of dividing HCC and ROI in the tumor-free liver tissue, vascular section of liver could not be included. Also we calculate the SUVmax ratio between tumor and tumor-free liver tissue (T/L ratio). Out of the purpose of distinguishing, SUVmax and T/L ratio of ED 18F-FDG PET/CT were respectively shorted as ED SUVmax and ED T/L ratio, while SUVmax and T/L ratio of WB 18F-FDG PET/CT were respectively shorted as WB SUVmax and WB T/L ratio. Since the acquisition time of the first frame imaging (0-15 s) of ED 18F-FDG PET/CT was relatively early, and there was no 18F-FDG concentration (ie, SUVmax= 0) in HCC and tumor-free liver tissue of most cases, therefore, in the event that the denominator was 0 when calculating the ED SUVmax ratio between tumor and tumor-free liver tissue, the calculation could not continue. Thus, the T/L ratio of first frame imaging of ED 18F-FDG PET/CT was not applied in the calculation in this study. In addition, in this study, the maximum SUVmax and maximum T/L ratio among all the frames of ED 18F-FDG PET/CT were respectively recorded as peak ED SUVmax and peak ED T/L ratio. Time to peak (s) refers to the time needed for reaching the peak ED SUVmax, and the median time of the scanning frame where peak ED SUVmax situated was taken as time to peak. Rate-to-peak ED SUVmax (SUV/s) refers to the ration of peak ED SUVmax and time to peak.5.2 Visual analysisUnder the circumstance of not informing the clinical information of patients and other imaging result, the two senior doctors conducted analysis over all the PET, CT and PET/CT fusion imaging, and used CT imaging to ascertain the location of lesion. In ED 18F-FDG PET/CT, the focal 18F-FDG concentration shade shown in any frame of imaging, after eliminating vascular imaging, were considered as hyperperfusion lesion (positive finding). In WB 18F-FDG PET/CT, focal 18F-FDG concentration shade was considered as hypermetabolism lesion (positive finding). When different views came up, it would be settled after discussion.6. Statistic MethodThe data were expressed as the means ± standard deviations (X±S). All statistical analyses were performed with SPSS, Version 13.0.In detecting HCC lesion, the compare of 18F-FDG perfusion parameters (peak ED SUVmax, time to peak and rate-to-peak ED SUVmax) and metabolic parameters (WB SUVmax) between HCC lesion and tomor-free liver tissue was examined with independent samples t-test, and the compare between peak ED T/L ratio and WB T/L ratio was examined with paired samples t-test. In size-based analysis, accessing the relationship between the size of HCC and the 18F-FDG perfusion parameters (peak ED SUVmax, time to peak, rate-to-peak ED SUVmax and peak ED T/L ratio) as well as metabolic parameters (WB SUVmax and WB T/L ratio) underwent Spearman correlation analysis. In differentiation-based analysis, accessing the relationship between the differentiation level of HCC and the 18F-FDG perfusion parameters (peak ED SUVmax, time to peak, Ratio-to-peak ED SUVmax and peak ED T/L ratio) as well as metabolic parameters (WB SUVmax and WB T/L ratio) underwent Spearman correlation analysis. The compare over the HCC detection rate of ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT was conducted by Cochran’s overall Test. If there is no significant difference shown in the overall Test, the multiple comparisons would not be conducted; otherwise, the multiple comparisons were conducted by McNemar Test and the significant level was used 0.05/3. The compare over the HCC detection rate of ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT between small and large size HCC lesions as well as the compare over the HCC detection rate of ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT among different differentiation HCC lesions were conducted by independent samples Χ2 Test.In the differential diagnosis of begin and malignant liver lesions, the assessment on the differential diagnosis value over 18F-FDG perfusion parameters (peak ED SUVmax, time to peak, rate-to-peak ED SUVmax and peak ED T/L ratio) and metabolic parameters (WB SUVmax and WB T/L ratio) was by applying Receiver Operating Characteristic (ROC) curve estimation. The compare on the differential diagnosis value of ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT was conducted by Cochran’s overall Test. If there is no significant difference shown in the overall Test, the multiple comparisons would not be conducted; otherwise, the multiple comparisons were conducted by McNemar Test and the significant level was used 0.05/3.[Results]1. Detection of hepatocellular carcinoma1.1 Lesion-based analysis1.1.1 Semi-quantitative analysisFor ED 18F-FDG PET/CT, the peak ED SUVmax in HCC was significantly higher than that of tumor-free liver tissue, being 9.7 ± 4.8 and 7.8 ± 4.5(t= 1.999, P = 0.048), respectively; time to peak was significantly less than that of tumor-free liver tissue, being 45.3 ± 8.1 and 51.6 ± 3.5 (t=-5.163, P= 0.000), respectively; rate-to-peak ED SUVmax was significantly higher than that of tumor-free liver tissue, being 0.22 ±0.12 and 0.15 ± 0.09 (t= 3.412, P= 0.001), respectively.For WB 18F-FDG PET/CT, WB SUVmax in HCC was significantly higher than that of tumor-free liver tissue, being 5.5 ± 3.4 and 2.6 ± 0.5 (t= 6.146, P= 0.000) respectively.The maximum T/L ratio of the ED 18F-FDG PET/CT (i.e., peak ED T/L ratio) was significantly higher than the T/L ratio of WB 18F-FDG PET/CT (i.e., WB T/L ratio), being 4.4 ± 3.4 and 2.2 ± 1.5 (t=4.385, P= 0.000), respectively.1.1.2 Visual analysisThere was significant disparity over the detection rate of HCC lesion among ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 78.8%(41/52),59.6%(31/52) and 92.3% (48/52), Cochran Q= 18.250, P= 0.000, respectively. Multiple comparison showed the HCC detection rate of ED 18F-FDG PET/CT was slightly higher than that of WB 18F-FDG PET/CT, being 78.8%(41/52) and 59.6%(31/52), P= 0.064 (P> 0.05/3), respectively; one-stop 18F-FDG PET/CT significantly improved the detection rate of WB 18F-FDG PET/CT and ED 18F-FDG PET/CT from 59.6% and 78.8% to 92.3% (48/52), P= 0.000 (P< 0.05/3) and P= 0.016 (P< 0.05/3), respectively.1.2 Size-based analysis1.2.1 Semi-quantitative analysisOf the 52 HCC lesions, the average size of lesions was 5.0 ± 3.4 cm (ranged from 1.1 to 15.0 cm). The 18F-FDG perfusion parameters in HCC lesions (peak ED SUVmax, time to peak, rate-to-peak ED SUVmax and peak ED T/L ratio) were 9.7 ± 4.8,45.3 ±8.1,0.22 ± 0.12 and 4.4 ± 3.4, respectively. The 18F-FDG metabolic parameters (WB SUVmax and WB T/L ratio) in HCC lesions were 5.5 ± 3.4 and 2.2 ± 1.5, respectively.Spearman’s correlation analysis showed that, there was no significant correlation between the size of HCC lesion and 18F-FDG perfusion parameters (peak ED SUVmax (r= 0.104, P= 0.463), time to peak (r= 0.013, P= 0.926), rate-to-peak ED SUVmax (r= 0.106, P= 0.456) and peak ED T/L ratio (r= 0.033, P= 0.815)); there was of a positive correlation with 18F-FDG metabolic parameters (WB SUVmax (r= 0.292, P= 0.035) and WB T/L ratio (r= 0.287, P= 0.039)).1.2.2 Visual analysisFifty-two HCC lesions were divided into two groups according to their maximum diameter; small size group included 28 HCC lesions (1.0 cm< maximum diameter of tumor≤ 5.0 cm), large size group included 24 HCC lesions (maximum diameter of tumor> 5.0 cm).There was no distinct disparity over the detection rate of small size group and large size group in ED 18F-FDG PET/CT, being 75.0% (21/28) and 83.3% (20/24), Χ2-0.538, P= 0.463, respectively. While the detection rate in small size group under WB 18F-FDG PET/CT was significantly lower than that of large size group, being 42.9%(12/28) and 79.2%(19/24),Χ2= 7.077, P= 0.008, respectively. There was no distinct disparity over the detection rate of small size group and large size group in one-stop 18F-FDG PET/CT, being 89.3% (25/28) and 95.8%(23/24), Χ2= 0.780, P= 0.377, respectively.There was significant disparity over the detection rate of small size group among ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 75.0%,42.9% and 79.2%, Cochran Q= 15.647, P= 0.000, respectively. Multiple comparison showed the detection rate under ED 18F-FDG PET/CT in small size group was slightly higher than that under WB 18F-FDG PET/CT, being 75.0% and 42.9%, P= 0.049 (P> 0.05/3); one-stop 18F-FDG PET/CT imaging significantly improved the detection rate of WB 18F-FDG PET/CT and ED 18F-FDG PET/CT of small size group from 42.9% and 75.0% 9.3%, P= 0.000 (P< 0.05/3) and P= 0.046 (P> 0.05/3), respectively.There was no distinct disparity over the detection rate of large size group in ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 83.3%,79.2% and 95.8, (Cochran Q= 3.714, P= 0.156).1.3 Differentiation-based analysis1.3.1 Semi-quantitative analysisAmong the HCC patients,18 patients with 23 HCC lesions got pathological diagnosis. Grouping according to the differentiation degree of tumor, there were highly, moderately, poorly differentiated lesions in 5,11 and 7, respectively. Peak ED SUVmax were 7.6 ± 3.8,9.8 ± 5.9 and 9.0 ± 2.3, respectively; time to peak were 52.5 ± 0.0,44.3 ± 7.8 and 41.8 ± 11.3, respectively; rate-to-peak ED SUVmax were 0.14 ± 0.07,0.22 ±0.12 and 0.23 ±0.10, respectively; peak ED T/L ratio were 2.4 ± 1.2,4.2 ± 3.5 and 3.8 ± 3.1, respectively; WB SUVmax were 2.7 ± 0.9,6.7 ± 5.3 and 5.5 ± 3.2, respectively; WB T/L ratio were 1.1 ± 0.1,3.0 ± 2.6 and 2.0 ± 1.2, respectively.Spearman’s correlation analysis showed that, there was no significant relevant between differentiation degree of HCC lesions and 18F-FDG perfusion parameters as well as metabolic parameters:peak ED SUVmax (r=-0.166, P= 0.448), time to peak (r=0.400, P= 0.059), rate-to-peak ED SUVmax (r=-0.324, P= 0.132), peak ED T/L ratio (r=-0.142,P= 0.519), WB SUVmax (r=-0.339, P= 0.113), WB T/L ratio (r=-0.172, P= 0.434).1.3.2 Visual analysisThe detection rates of ED 18F-FDG PET/CT towards highly, moderately and poorly differentiated of HCC lesions were 4/5,8/11 and 6/7, and there was no significantly difference (Χ2= 0.435, P= 0.804). The detection rates of WB 18F-FDG PET/CT towards highly, moderately and poorly differentiated of HCC lesions were 0/5,7/11 and 6/7, and there was no significantly difference (Χ2= 0.374, P= 0.052). The detection rates of one-stop 18F-FDG PET/CT towards highly, moderately and poorly differentiated of HCC lesions were 4/5,10/11 and 4/7, and there was no significantly difference (Χ2= 5.929, P= 0.829).There was significantly difference on detection rates towards highly differentiated HCC among ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 4/5,0/5 and 4/5 (Cochran Q= 8.000, P= 0.018), respectively. Multiple comparison showed the detection rates of ED 18F-FDG PET/CT towards highly differentiated HCC was slightly higher than that of WB 18F-FDG PET/CT, Χ2= 2.250,P= 0.134 (P> 0.05/3); one-stop 18F-FDG PET/CT slightly improved the detection rate of WB 18F-FDG PET/CT and ED 18F-FDG PET/CT from 0/5 and 4/5 to 4/5, P= 0.134 (P> 0.05/3) and P= 1.000 (P> 0.05/3), respectively.There was no significantly difference on detection rates towards moderately differentiated HCC among ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 8/11,7/11, and 11/10 (Cochran Q= 2.800, P= 0.247), respectively.There was no significantly difference on detection rates towards poorly differentiated HCC among ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT, being 6/7,4/7, and 6/7 (Cochran Q= 4.000, P= 0.135), respectively.2. Differential diagnosis between benign and malignant liver lesions2.1 Semi-quantitative analysisFor ED 18F-FDG PET/CT, the peak ED SUVmax of begin and malignant liver lesion was 8.2 ± 6.6 and 9.2 ± 4.7, time to peak was 50.0 ± 5.8 and 45.7 ± 8.4, rate-to-peak ED SUVmax was 0.17 ± 0.16 and 0.18 ± 0.10, and peak ED T/L ratio was 2.8 ± 2.2 and 4.1 ± 3.3, respectively. For WB 18F-FDG PET/CT, the WB SUVmax of begin and malignant liver lesion was 2.5 ± 0.9 and 5.9 ± 3.5, and WB T/L ratio was 1.1 ± 0.4 and 2.4 ± 1.6, respectively.ROC curve estimation showed that Area Under the Curve (AUC) of these parameters in differential diagnosis between benign and malignant liver lesions, from high to bottom, was WB SUVmax (AUC= 0.864, P= 0.000), WB T/L ratio (AUC= 0.856, P= 0.000), peak ED T/L ratio (AUC= 0.612, P= 0.150), peak ED SUVmax (AUC= 0.592, P= 0.235), rate-to-peak ED SUVmax (AUC= 0.583, P= 0.284) and time to peak (AUC= 0.371, P= 0.097).2.2 Visual analysisIf considering 18F-FDG hyperperfusion and (or) hypermetabolic lesion as malignant liver tumor, the diagnostic sensitivity of ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT was 75.8% (47/62),64.5% (40/ 62) and 93.5% (58/62), Cochran Q= 17.034, P= 0.000, respectively; specificity was 44.4%(8/18),77.8%(14/18) and 38.9%(7/18), Cochran Q= 10.750, P= 0.005, respectively; accuracy was 68.8%(55/80),67.5%(54/80) and 81.3%(65/ 80), Cochran Q= 6.000, P= 0.050, respectively.Multiple comparison showed the diagnostic sensitivity of ED 18F-FDG PET/CT and WB 18F-FDG PET/CT was 75.8%(47/62) and 64.5%(40/62), P= 0.265 (P> 0.05/3), respectively; specificity was 44.4%(8/18) and 77.8%(14/18), P= 0.070 (P> 0.05/3), respectively; accuracy was 68.8%(55/80) and 67.5%(54/80), P= 1.000 (P> 0.05/3), respectively. One-stop 18F-FDG PET/CT improved the sensitivity of WB 18F-FDG PET/CT and ED 18F-FDG PET/CT from 64.5% and 75.8% to 93.5% (58/62), P= 0.000 (P< 0.05/3) and P= 0.001 (P< 0.05/3), respectively; reduced the specificity from 77.8% and 44.4% to 38.9% (7/18), P= 0.016 (P< 0.05/3) and P= 1.000 (P> 0.05/3); improved the accuracy from 67.5% and 68.8% to 81.3% (65/80), P= 0.046 (P> 0.05/3) and P= 0.006 (P< 0.05/3), respectively.Conclusion1. ED 18F-FDG PET/CT could be applied in distinguishing the difference on blood perfusion between HCC lesion and tumor-free liver tissue; WB 18F-FDG PET/CT could be applied in distinguishing the difference on glucose metabolism between HCC lesion and tumor-free liver tissue.2. ED 18F-FDG PET/CT performs slightly better than conventional imaging in detecting HCC lesion; one-stop 18F-FDG PET/CT has significantly improved the detection of HCC lesion.3. The 18F-FDG perfusion shows less relevance with the size of HCC lesion; the 18F-FDG metabolism shows positive correlation with the size of lesion.4. ED 18F-FDG PET/CT performs no significantly distinct in positive rate of different size of lesions; WB 18F-FDG PET/CT has a significant lower detection rate of small size group of lesion than that of the large group; ED 18F-FDG PET/CT has a significant better detection rate towards small group lesion than that of WB 18F-FDG PET/CT, and shows no significant difference in detecting large size lesion as compared with WB 18F-FDG PET/CT; one-stop 18F-FDG PET/CT has significantly improved the detection rates of small lesion, but has not significantly improved the detection of large lesion.5. Both the 18F-FDG blood perfusion and metabolism show less relevance with the differentiation degree of HCC lesion.6. ED 18F-FDG PET/CT, WB 18F-FDG PET/CT and one-stop 18F-FDG PET/CT appears to be of no significant difference in positive rate of HCC lesion with different differentiation degrees.7. The 18F-FDG metabolic parameters seem to perform higher value than the 18F-FDG perfusion parameters in differential diagnosis of benign and malignant focal liver lesions.8. ED 18F-FDG PET/CT imaging in differential diagnosis of benign and malignant liver lesions, shows slightly higher sensitivity as compared with WB 18F-FDG PET/CT, shows lower specificity than that of WB 18F-FDG PET/CT and shows less difference on accuracy as compared with WB 18F-FDG PET/CT. One-stop 18F-FDG PET/CT has improved the sensitivity of WB 18F-FDG PET/CT, reduced the specificity and raised the accuracy. |
PDF Full Text Request