| Objective:To establish test model consistency of the body position on computed tomography (CT) and magnetic resonance (MR) scanners, to investigate the accuracy of image registration and provide theoretical basis and the feasible approach of image fusion in clinical.Methods:The two groups each thirty-six consecutive patients were pathologic diagnosis patients with NPC and enrolled in this study. Each patient underwent CT and MRI scan in the planned treatment. There were differenced between the first group and the second group. The first group had not make tag on body surface, the second group made the body position conformed with CT-simulation and magnetic resonance imaging scanning according to request.Transfered simultaneously the two species images to the Toncom workstations. The two groups used respectively Manual and LandMark method to CT/MRI image registration, manual adjustment was made to optimize the accuracy of image fusion. The GTV and lymphonodi retropharyngic of each patient was independently estimated and contoured by two observers and one associate chief physician of radiology department in my hospital. Compared the two groups, case carried out the precision of image fusion respectively. The volume of GTV contoured on CT(V-GTV-CT),MRI(V-GTV-MRI) and their overlap (V-GTV-F) and composite volumes(V-GTV-CT+MRI) were measured. The percentage of V-GTV-CT and V-GTV-MRI on V-GTV-CT+MRI, V-GTV-F on V-GTV-CT was also calculated ,respectively. Computed average errors on X,Y,Z axis by LandMark.Results: The median and quartering percentiles for P-GTV between the first group(un-conformity) and the second group(body-conformity), 60.249±18.261cm3, 90.995±3.978cm3, respectively. There was significant difference in statistical significance(P=0.000). Average errors on X,Y,Z axis by LandMark were (0. 65±0. 17) mm,( 1.2±0. 11) mm,(2.2±0. 20 ) mm, the gross average errors of three-dimension was 2.2439±1.15mm.Conclusions: The CT and MR scanners of body position were critical during image registration. The method of the CT and MR scanners applied by marker on body surface In clinical were feasible. Automatic registration using LandMark method, CT and MRI images should be guided and integrated precisely for NPC GTV delineation in clinical. Objective: To evaluate the influence of CT, MRI, and CT +MRI image fusion technology for delineation of volumes of nasopharyngeal carcinoma (NPC).Methods: Thirty-six consecutive patients of newly diagnosed and recurrenced patients with NPC. Each patient underwent CT and MR scan in the planned treatment position within the immobilization and the same body position. Firstly, make scanning with CT-sim and make tag in different parts of the surface; secondly, make tag when scanned the same position with MR. Transfer the two images to the medical doctors workstations at the same time. Make image fusion on the use of the image fusion software on Tomcon workstation, and use LandMark as the registration methods. Two senior radiotherapy doctors and a Deputy Director physician of radiology in our hospital will make evaluation and sketching for registration. According to ICRU62 document, sketched focus lies in the CT, MRI, CT + MRI (image fusion) target of GTV, which are named GTVnx- CT, GTVnx- MRI, GTVnx- CT + MRI. Compare the three groups, we divide the test targets into two groups according to skull-base erosion and compare these two groups respectively. According to T stage,there are divided to two groups. One group is the earlier period(T1+T2 stage), the other is local advanced(T3+T4 stage).Results: GTV in the target area on CT, MRI, CT + MRI are significantly different. The difference average of these three target volume(cm3) areas is 27.603±3.247,30.992±3.941,31.71±4.038. Compared every two groups of these three Groups, there is significant difference between Group 1 and Group 2(P = 0.016). Group 1 and Group 3 are also significantly different (P = 0.004). Group 2 and Group 3 have no significant difference(sP = 0.199). Divide the test subjects into two groups according to skull base invasion and compare them. There is significant difference of three target areas in these two groups(P=0.000,P=0.832, respectively).There was significant difference between the earlier period (T1+T2 Stage)and the local advanced(T3+T4 stage)P=0.514, P=0.036, respectively.Conclusion: CT +MRI image fusion is helpful to improve the accuracy of sketching, particularly when the skull base and Retroparyngeal node matastasis and T3+T4 stage of patients were broken. Image fusion is ascendant in the clinical. It offres a theoretical basis and guidance for clinicians. |
PDF Full Text Request