RÉSUMÉ
Objective To explore the influence of delineator and contouring criteria training on the delineation of the tumor bed and whole breast target after breast-conserving surgery.Methods Twelve brcast cancer patients after breast conserving surgery were selected.Tumor bed marked by clips was defined as gross target volume 1 (GTV1),tumor bed formed by seroma was defined as GTV2 and the whole breast was defined as clinical target volume (CTV).Five junior radiation oncologists first delineated GTV1,GTV2 and CTV for each patient following their own criteria.After contouring criteria training,they then delineated GTV1,GTV2 and CTV for the same group of patients again.The differences of the volumes of GTV1,GTV2 and CTV before and after training among different delineators were compared.One-way ANOVA or matching t-test was performed.Results The inter-delineator variability on GTV1,GTV2 and CTV delineation before training was statistically significant (F =11.16,7.54 and 3.78,P =0.000,0.000 and 0.009).After training,the inter-delineator variability on GTV1 and GTV2 delineation had statistical significance (t =4.78 and 4.24,P =0.002 and 0.005),but the inter-delineator variability on CTV delineation had no statistical significance (t =1.52,P =0.209).The coefficient of variance of the GTV1,GTV2 and CTV before and after training was significantly different (t =3.14,2.81,2.70,P =0.009,0.017 and 0.021).The matching index of GTV1,GTV2 and CTV before and after training was significantly different (F =16.08,8.61,8.48,P =0.000,0.000 and 0.000).Conclusions In delineating the target of breast cancer,application of the criteria of target delineation can reduce the difference among the delineators,especially for CTV.
RÉSUMÉ
Objective To investigate the impact of delineator and delineating criteria on the target delineation of the peripheral lung cancer based on four-dimensional computed tomography (4D-CT).Methods The 4D simulation CT images of twelve patients with peripheral lung cancer were selected.Before and after the establishment of the target delineation criteria,six radiation oncologists were asked to delineate the targets based on 4D-CT images at the end-inhalation phase (0％),end-exhalation phase (50％),and three-dimensional computed tomography (3D-CT) images respectively.The delineated targets were denominated as GTV0,GTV50,GTV3D.IGTVIN+Ex was created by combining GTV0 and GTV50.IGTVMIP was delineated based on the maximum intensity projection (MIP) of 4D-CT.The interand intra-observer variability before and after the establishment of the delineation criteria was compared.Results The mean coefficients of variation of GTV0,GTV50,GTV3D,IGTVMIP and IGTVIN+Ex delineated by the six delineators before and after the establishment of the delineation criteria were 0.50 ± 0.25 vs 0.24 ±0.10,0.52 ±0.38 vs 0.26 ±0.12,0.45 ±0.19 vs 0.20 ±0.07,0.54 ±0.27 vs 0.23 ±0.09 and 0.44±0.23 vs 0.26 ±0.09,respectively.The differences were statistically significant(t =3.38,2.44,3.60,4.20,3.11,P ＜ 0.05).No statistically significant difference was found in the volume of the same target for GTV0,GTV50,GTV3D,IGTVMIP and IGTVIN+Ex delineated by the six oncologists before and after the establishment of the delineation criteria.For delineator 3 and 6,the differences of GTV0,GTV50,IGTVIN+EX before and after establishment of the delineation criteria were statistically significant(t =2.46,2.91,3.28,P ＜0.05 ;t =2.40,2.79,3.22,P ＜0.05).For delineator 4,the differences of GTV0,GTV50,IGTVIN+EX,IGTVMIP,GTV3D before and after establishment of the delineation criteria were statistically significant (t =2.70,3.21,3.04,3.99,3.00,P ＜ 0.05).Conclusions The unified delineation criteria can significantly reduce the inter-observer variability in delineating the GTVs based on 3D-CT or 4D-CT images,and in delineating the IGTVMIP based on 4D-CT images of peripheral lung cancer patients.However,regarding the same delineator,various influences of the unified delineation criteria on the target delineation are observed.