Build And Comparison Of Target Volume For Peripheral Lung Cancer Based On Four Dimensional CT And Three Dimensional CT Assisted With Active Breathing Control

Objective: To analyze artifacts of gross tumor volume (GTV) and correlated factors ineach phase images of four dimensional CT (4DCT) for peripheral lung cancer based onthree dimensional CT (3DCT) assisted with active breathing control (ABC). To comparethe position, displacement, degree of inclusion (DI) and matching index (MI) of GTV forperipheral lung cancer based on4DCT and3DCT assisted with ABC. To compare thedifferences of the position, volume and MI between internal gross tumor volumes (IGTVs)determined by four methods based on4DCT and3DCT scans during free breathing andIGTV determined by one method based on3DCT assisted with ABC.Methods: Patients with peripheral lung cancer underwent3DCT (CT3D) and4DCTsimulation scans during free breathing and then underwent3DCT simulation scans in endinspiration hold (CTEIH) and end expiration hold (CTEEH) assisted with ABC. The4DCTimages from each respiratory cycle were sorted into10phases,0%phase was defined asend-inspiratory phase (CT0),50%phase was defined as end-expiratory phase (CT50).TheGTVs were delineated separately on CT0, CT10…CT90, CTEIH, CTEEHand CT3Dimages,and then GTV0, GTV10…GTV90, GTVEIH, GTVEEHand GTV3Dwere constructedrespectively. GTV0, GTV10…GTV90were combined into IGTV10; GTV0and GTV50werecombined into IGTV2; IGTVMIPwas the GTV contour delineated from the maximumintensity projection (MIP); IGTV3Dwas acquired from the enlargement of3DCT-basedGTV by each spatial direction on the motion amplitude measured in4DCT; IGTVABCwasacquired by combining GTVEIHand GTVEEH. The relative deviations (Devref) between thereference GTV (GTVref) and the GTVs delineated based on CT3D(GTV3D) and all phases of 4DCT were calculated respectively. The correlations between GTVrefand Devmaxandbetween the tumor motion in the cranio-caudal (CC) direction and Devrefwere analyzed.The differences between4D displacements and ABC displacements in three dimensionaldirections were compared. Mutual DI and MI between GTV0and GTVEIH, GTV50andGTVEEHwere compared. The differences of the position, volume, DI and MI betweenIGTVABCand IGTV10, IGTVMIP, IGTV2, IGTV3Dwere analyzed respectively.Results: The maximum median Devrefof GTV was GTV3Dwith17.83%, and themaximum median Devrefof the GTVs in all phases of4DCT was GTV30with17.20%. Asignificant negative correlation was found between GTVEIHand Devmax(r=-0.691,P=0.001). The Devrefcorrelated with the tumor motion amplitude in the CC directionpositively(r=0.323-0.617, P=0.005-0.150). The partial regression coefficient of influenceof GTVrefsize and motion amplitude in the CC direction to the tumor Devmaxwere-0.500and0.583, P=0.002and0.001, respectively.The mean distances between the centroid positions of GTV0and GTVEIH, GTV50andGTVEEHwere4.78mm and4.03mm in all patients,4.28mm and3.50mm in upper lobegroup,6.10mm and5.41mm in lower lobe group respectively. The mean mutual DIs ofGTV0and GTVEIH, GTV50and GTVEEHwere62.97%-70.57%in upper lobe group and24.66%-32.28%in lower lobe group. The mean MIs of GTV0and GTVEIH, GTV50andGTVEEHwere0.53and0.56in upper lobe group,0.19and0.25in lower lobe grouprespectively. Whether in upper or lower lobe group, the differences between thedisplacements of centroid positions of GTVEIHand GTVEEH(ABC displacement) andGTV0and GTV50(4D displacement) were all smaller than1mm in three dimensionaldirections (z=-1.963-0.000, p=0.052-1.000).In upper lobe group, there were significant differences in centroid position betweenIGTVABCand IGTV10, IGTVMIPin lateral(LR) and anterio-posterior (AP) directionsrespectively (t=-2.377,-2.199, P＜0.05), and there were significant differences in centroidposition between IGTVABCand IGTV2, IGTV3Din both LR and AP directions (t=-2.185--2.767, p＜0.05). In lower lobe group, there were significant differences in centroidposition between IGTVABCand IGTV10, IGTVMIP, IGTV3Din LR direction (t=3.950-4.543,P＜0.05). Either in the upper lobe group or in the lower lobe group, IGTVABCwas smaller than IGTV10or IGTV3Dsignificantly (Z=-3.180--2.023, P＜0.05) and the MIs betweenIGTVABCand other IGTVs were larger in the upper lobe group than in the lower lobe groupsignificantly (Z=-2.419--2.119, P＜0.05).Conclusions: The GTV artifacts in each phase of4DCT for the peripheral lung cancerwere different to each other, and the influence of target displacement to artifacts was largerthan target volume, so artifacts could be reduced by controlling breathing to reduce targetdisplacement.The displacements of tumor based on3DCT scanning in end-inspiratory hold andend-expiratory hold can be used to construct internal target volume instead of that based on4DCT scanning in extreme phases for peripheral lung cancer, but spatital mismatches ofGTVs were obvious between extreme phases in4DCT and corresponding phases in3DCTassisted with ABC especially for tumors with smaller volume and larger motion amplitude.IGTVABCis not completely coincide with IGTVs acquired based on4DCT or3DCTduring free breathing, and the spatial mismatch corrrelated with the tumor position, thespatial mismatch is more significant for middle-lower lobe tumor than upper lobe tumor.IGTVABCare larger than IGTV2but smaller than other IGTVs acquired based on4DCT andIGTV acquired based on3DCT during free breathing.