In vivo dosimetry in bladder and rectum as well as determining doses on suggested reference points following the ICRU report 38 contribute to quality assurance in HDR-brachytherapy of cervical carcinoma, especially to minimize side effects. In order to gain information regarding the radiation exposure at ICRU reference points in rectum, bladder, ureter and regional lymph nodes those were calculated (digitalisation) by means of orthogonal radiographs of 11 applications in patients with cervical carcinoma, who received primary radiotherapy. In addition, the doses at the ICRU rectum reference point was compared to the results of in vivo measurements in the rectum. The in vivo measurements were by factor 1.5 below the doses determined for the ICRU rectum reference point (4.05 +/- 0.68 Gy versus 6.11 +/- 1.63 Gy). Reasons for this were: calibration errors, non-orthogonal radiographs, movement of applicator and probe in the time span between X-ray and application, missing connection of probe and anterior rectal wall. The standard deviation of calculations at ICRU reference points was on average +/- 30%. Possible reasons for the relatively large standard deviation were difficulties in defining the points, identifying them on radiographs and the different locations of the applicators. Although 3 D CT, US or MR based treatment planning using dose volume histogram analysis is more and more established, this simple procedure of marking and digitising the ICRU reference points lengthened treatment planning only by 5 to 10 minutes. The advantages of in vivo dosimetry are easy practicability and the possibility to determine rectum doses during radiation. The advantages of computer-aided planning at ICRU reference points are that calculations are available before radiation and that they can still be taken into account for treatment planning. Both methods should be applied in HDR-brachytherapy of cervical carcinoma.