OBJECTIVETo compare the position and magnitude of internal target gross volume (IGTV) of primary hepatocarcinoma delineated by three methods based on four-dimensional computed tomography (4D-CT) and to investigate the relevant factors affecting the position and magnitude.

METHODSTwenty patients with primary hepatocarcinoma after transcatheter arterial chemoembolization (TACE) underwent big bore 4D-CT simulation scan of the thorax and abdomen using a real-time position management (RPM) system for simultaneous record of the respiratory signals. The CT images with respiratory signal data were reconstructed and sorted into 10 phase groups in a respiratory cycle, with 0% phase corresponding to end-inhale and 50% corresponding to end-exhale. The maximum intensity projection (MIP) image was generated. IGTVs of the tumor were delineated using the following three methods: (1) The gross tumor volume (GTV) on each of the ten respiratory phases of the 4D-CT image set was delineated and fused ten GTV to produce IGTV10; (2) The GTVs delineated separately based on 0% and 50% phase were fused to produce IGTV(IN+EX); (3) The visible tumor on the MIP image was delineated to produce IGTV(MIP). Twenty patients were divided into groups A and B based on the location of the target center,and were divided into groups C and D based on the tumor maximum diameter. The patients were divided into groups E and F based on the three-dimensional (3D) motion vector of the target center. The position of the target center, the volume of target, the degree of inclusion (DI) and the matching index (MI) were compared reciprocally between IGTV10, IGTV(IN+EX) and IGTV(MIP), and the influence of the tumor position and 3D motion vector on the related parameters were compared based on the grouping.

RESULTSThe average differences between the position of the center of IGTVs on direction of X, Y and Z axes were less than 1.5 mm, and the difference was statistically not significant. The volume of IGTV10 was larger than that of IGTV(IN+EX), but the difference was not significant (t = 0.354, P = 0.725). The volume of IGTV10 was larger than that of IGTV(MIP) but the difference was not significant (t = -0.392, P = 0.697). The ratio of IGTV(IN+EX) to IGTV10 was 0.75 +/- 0.15 and the ratio of IGTV(MIP) to IGTV10 was 0.78 +/- 0.14. The DI of IGTV(IN+EX) in IGTV10 was (74.85 +/- 15.09)% and that of IGTV(MIP) in IGTV10 was (68.87 +/- 13.69)%. The MI between IGTV10 and IGTV(IN+EX), IGTV10 and IGTV(MIP) were 0.75 +/- 0.15 and 0.67 +/- 0.13, respectively. The median of ratio of IGTV(IN+EX)/ IGTV10 was 0.57 in group A versus 0.87 in group B, statistically with a significant difference between the groups A and B (Z = -3.300,P = 0.001). The median of ratio of IGTV(MIP)/IGTV10 was 0.51 in the group A and 0.72 in group B, with a significant difference between the groups A and B (Z = -3.413, P = 0.001). The median of ratio of IGTV(IN+EX)/IGTV10 was 0.79 in group C versus 0.74 in group D, with a difference not significant (Z = -0.920, P = 0.358). The median of ratio of IGTV(MIP)/IGTV10 was 0.85 in group C versus 0.80 in group D, with a non-significant difference (Z = -0.568, P = 0.570). The median of ratio of IGTV(IN+EX)/IGTV10 was 0.87 in group E versus 0.68 in group F, with a significant difference between the two groups (Z = -2.897, P = 0.004). The median of ratio of IGTV(MIP)/IGTV10 was 0.85 in the group E versus 0.81 in the group F, with a non-significant difference (Z = -0.568, P = 0.570).

CONCLUSIONSThe center displacement of the IGTVs delineated separately by the three techniques based on 4D-CT images is not obvious. IGTV(IN+EX) and IGTV(MIP) can not replace IGTV10, however, IGTV(IN+EX) is more close to IGTV10 comparing with IGTV(MIP). The ratio of IGTV10 and IGTV(MIP) is correlated to the 3D motion vector of the tumor. When the tumor is situated in the upper part of the liver and with a 3D motion vector less than 9 mm, IGTV10 should be the best IGTV.