Abstract: Objective To compare the consistency of different MR sequences in measuring the target area of radiotherapy for cervical cancer. Methods A retrospective analysis was made on the data of 38 patients with cervical cancer initially diagnosed by pathology（ⅡB or above）. Tumor contours were delineated manually on axial T2WI，T1WI enhanced scan（T1+C） and apparent diffusion coefficient （ADC），respectively. Gross tumor volume（GTV） was calculated by area summation method. The GTV obtained from different sequences were divided into T2WI group，T1+C group and ADC group. One-way ANOVA was used to calculate the average values of three groups of GTV and compare their differences. Bland-Atlman method was used to calculate 95% consistency limits（95% LoA） of three groups of GTVs，and the consistency of three groups of GTVs was compared. Results The mean value of GTV in T2WIgroup，T1+C group and ADC group were （46.35±28.04） cm³、（44.27±26.66） cm³ and （47.50±28.55） cm³，respectively. There was no significant difference in mean GTV among the three groups（F=0.132，P=0.877）. Bland-Atlman results showed that 95% LoA in T2WI group and T1+C group，T2WI group and ADC group，T1+C group and ADC group were（-16.43-20.59） cm³，（-19.01-25.47） cm³ and  （-20.13-22.44） cm³，respectively. Bland-Atlman scatter plots in different groups were 5.26%（2/38） out of the 95% LoA range，which was beyond the clinical alternative range. Conclusions The consistency of different MR sequence measurements of cervical cancer radiotherapy target areas is not good，and can not be replaced by each other. Comprehensive analysis of T2WI，T1+C and ADC is more conducive to the accurate delineation of cervical cancer radiotherapy target areas.

Key words: Cervical , canler, Radiotherapy target area, Magnetic resonance imaging, Consistency comparison