Purpose The body contour of patients with cervical cancer is prone to change between radiotherapy sessions. This study aimed to investigate the effect of body contour changes on the setup and dosimetric accuracy of radiotherapy. Methods 15 patients with cervical cancer after surgery were randomly selected for retrospective analysis. The body contours on the once-per-week cone-beam computed tomography (CBCT) were registered to the planning CT (pCT) for subsequent evaluation. A body contour conformity index (CI body ) was defined to quantify the variation of body changes. The body volume measured by CBCT was collected, and its relative difference in reference with the first CBCT was calculated and denoted by ΔV n . The relative setup errors, denoted by ΔSE LR , ΔSE AP , ΔSE SI , and ΔSE vec for left–right, anterior–posterior, superior–inferior, and vectorial shifts, respectively, were defined as the difference in measured setup errors between the reference and following CBCTs. The planned dose was calculated on the basis of virtual CT generated from CBCT and pCT by altering the CT body contour to fit the body on CBCT without deformable registration. The correlations between body contour changes and relative setup errors as well as dosimetric parameters were evaluated using Spearman’s correlation coefficient r s . Results CI body was found to be negatively correlated with the superior–inferior and vectorial relative setup errors ΔSE SI ( r s = −0.448, p = 0.001) and ΔSE vec ( r s = −0.387, p = 0.002), and no significant correlation was found between relative setup errors and ΔV n . Moreover, ΔV n was negatively correlated with ΔD 2 ( r s = −0.829, p < 0.001), ΔD 98 ( r s = −0.797, p < 0.001), and ΔTV PIV ( r s = −0.819, p < 0.001). ΔD 2 , ΔD 98 , and ΔTV PIV were negatively correlated with ΔV n ( p < 0.005). No correlation was found for other examined dosimetric parameters. Conclusion The body contour change of patients could be associated with the setup variability. The effect of body contour changes on dose distribution is minimal. The extent of body change could be used as a metric for radiation therapists to estimate the setup errors.