| Objects: To evaluate the effect of respiratory motion on the gross tumorvolume contoured by PET image and PET-CT fusion image using a respiratorymotion phantom system. To evaluate the effect of respiratory motion onthe gross tumor volume (GTV) contoured by CT image using a respiratorymotion phantom system. To evaluate the effect of respiratory motion onaccuracy of PET-CT image coregistration using a respiratory motionphantom system.Methods: This study is divided into three parts. In the first part,Twenty-eight kinds of respiratory motions were simulated using arespiratory motion phantom system. The parameters of the respiratorymotion phantom system included the respiratory frequency, the amplitudeon z axis and x axis. The respiratory frequencies were 16 times per minute,18 times per minute and 20 times per minute. The amplitudes on z axis were1.0cm, 1.5cm, 2.0cm and on x axis 0.5cm, 1.0cm and 1.5cm, respectively.The PET-CT scanning was performed under different conditions. Then thegross tumor volume (GTV) was contoured by PET image and PET-CT fusion image.According to phantom volume, the data were divided into five groups(GTV1group, GTV2 group, GTV3 group, GTV4 group and GTV5 group). In the secondpart, the respiratory motions were simulated using a motion phantom system and the parameter of the respiratory motion phantom system was the sameas part one. The CT images were obtained and the GTVs were contoured byCT images. The groups in details were as the same as part one too. In thethird part, twenty-eight kinds of respiratory motions were simulatedusing a respiratory motion phantom system. The registration accuracy ofPET-CT image under different respiratory conditions and different phantomvolumes was measured by calculating the R value with formula: R=VPET-CT/VPET+CT.Results: 1. (1) The results of PET image: The GTVs of the 5 groups onquiescent condition were 2.78±0.09 cm3, 7.90±0.21 cm3, 14.15±1.23cm3, 22.05±0.48 cm3 and 40.65±1.02 cm3, and on motional condition 3.96±0.24 cm3 5.89±1.09 cm3, 9.68±2.32 cm3 14.41±3.19 cm3, 16.39±2.22 cm3 22.58±6.07 cm3, 27.69±3.85 cm3 37.12±8.11 cm3 and44.44±4.72 cm3 56.77±10.83cm3, respectively. The differences ofthe GTV in the 5 groups under different respiratory frequency (F=7.43,4.95, 6.28, 9.20, 7.61, respectively; all P＜0.01), on z axis (F=6.51,5.91, 4. 13, 5.77, 3.57, respectively; all P＜0.01), and on x axis (F=13.17,29.23,23.23,29.01,23.99, respectively; all P＜0.01) weresignificant.(2) The results of PET-CT fusion image: The GTVs of the 5 groupson quiescent condition were 2.80±0.53 cm3, 8.30±0.11 cm3, 14.85±0.69 cm3, 24.20±0.11 cm3 and 41.65±0.59 cm3, and on motional conditionwere 5.30±1.28 cm3 6.79±1.91 cm3, 13.38±4.42 cm3 17.96±5.16cm3, 19.54±4.59 cm3 26.11±8.06 cm3, 30.91±7.21 cm3 40.93±9.56 cm3 and 49.14±7.06 cm3 61.44±11.62 cm3, respectively. Thedifferences of the. GTVs in the 5 groups under different respiratoryfrequency (F=8.26, 5.10, 8.86, 9.40, 8.66, respectively; all P＜0.01),on z axis (F=6.50, 5.20, 4.87, 6.39, 6. 00, respectively; all P＜0.01), and on x axis (F=8.28, 20.73, 15.04, 19.89, 21.74, respectively; allP＜0.01) were significant.(3) The result of univariate indicated that, both in PET andin PET-CT image, the interaction between the motion on z axis and x axis,on z axis and frequency, on x axis and frequency, were significantrespectively (P＜0.01).2. The GTVs of the 5 groups on quiescent condition were 2.65±0.05 cm3, 6.80±0.43 cm3, 13.15±1.12 cm3, 23.65±0.27 cm3 and 36.55±0.27 cm3, and on motional conditio 3.02±0.46 cm3 3.42±1.27cm3,8.50±1.73 cm3 9.44±3.31 cm3, 15.02±3.64 cm3 16.24±3.14 cm3,25.49±6.51 cm3 30.82±6.63 cm3 and 40.19±4.98 cm3 45.48±6.78cm3, respectively. The ANOVA showed that on different respiratoryfrequency condition, the difference of GTV in GTV2, 4, 5 group wassignificant (F=3.18, 6.68, 7.20, respectively, P=0.027, 0.000, 0.000,respectively), but in GTV1, 3 group were not (F=2.20, 2.42, respectively,P=0.093, 0.070, respectively). On different amplitude on z axis, thedifference of GTV in GTV4, 5 group was significant (F=7.86, 9.10,respectively, all P=0.000) but in GTV1, 2, 3 group was not (F=1.19,2.55, 1.88, respectively, P=0.317, 0.061, 0.138, respectively). Ondifferent amplitude on x axis, the difference of GTVin GTV4, 5 group wassignificant (F=5.40, 5.07, respectively, P=0.002, 0.003, respectively)but in GTV1, 2, 3 group were not (F=1.391, 2.14, 2.10, respectively,P=0.249, 0.099, 0.104, respectively). The result of univariate indicatedthat, both in PET and in PET-CT image, the interaction between the motionon z axis and x axis, on z axis and frequency, on x axis and frequencywas significant (P＜0.01).3. The R value under static condition (R0) was 0.88±0.21, andwith motion frequencies of 16, 18, 20 times per minute were 0.61±0.21 (R16), 0.58±0.25 (R18), 0.54±0.22 (R20), respectively. The differencesamong R0, R16, R18 and R20 were significant (F=6.24, P＜0.01). The R valueswith motion amplitudes of 1.0cm, 1.5cm and 2.0cm on z axis were 0.55±0.25 (Rz-1.0), 0.60±0.20 (Rz-1.5) and 0.58±0.23 (Rz-2.0) respectively.The R values with motion amplitudes of 0.5 cm, 1.0 cm and 1.5cm on x axiswere 0.63±0.24 (Rx-0.5), 0.60±0.18 (Rx-1.0) and 0.50±0.24 (Rx-1.5)respectively. The differences of R values on z axis and x axis weresignificant. The R values with phantom volumes of 1.6 cm3, 6.1 cm3, 13.4cm3, 23.4 cm3 and 36.7cm3 were 0.36±0.18 (R1.6), 0.51±0.22 (R6.1), 0.580.18 (R13.4) 0.71±0.20 (R23.4) and 0.72±0.15 (R936.7)) respectively. Thedifferences of R values among R0, R1.6, R6.1, R13.4, R23.4 and R36.7 were significant(F=18.23, P＜0.01). The interaction between the motion on z axis andx axis, on z axis and frequency, on x axis and frequency were significant(P＜0.01).Conclusions: 1. When the respiratory frequency were 16, 18 and 20 timesper minute, when amplitudes on z axis were 1.0 cm, 1.5 cm and 2.0 cm, andon x axis 0.5 cm, 1.0 cm and 1.5 cm, the GTVs contoured by PET and PET-CTimages on motional condition were 3.8％91.7％and 18.0％142.1％respectively lager than those on quiescent condition. The difference ofthe GTVs contoured by PET and PET-CT images on motional condition was-59.87％32.53％and -72.9％54.7％, respectively. The interaction of allparameters was significant and the respiratory motion could affect GTVcontouring on both PET and PET-CT fusion image;2. When the respiratory frequency were 16, 18, 20 times perminute, when amplitudes on z axis were 1.0 cm, 1.5 cm and 2.0 cm, and onx axis 0.5 cm, 1.0 cm and 1.5 cm, the GTVs contoured by CT images on motionalcondition were 7.8％38.8％lager than those on quiescent condition. Thedifference of the GTV contoured by PET images on motional condition was -20.4%～14.4%. The interaction of all parameters was significant and therespiratory motion could affect GTV contouring on CT image.3. Misregistration will be occurred when the respiratory motionexists and may impact on the target volume delineation in radiationtreatment planning with PET-CT fused images for NSCLC, so theprecautionary measures should be adoped to lower it. |
PDF Full Text Request