In fractionated stereotactic radiotherapy (SRT), the accuracy of patient relocation is very important. The Gill-Thomas-Cosman (GTC) re-locatable stereotactic frame is used for patient immobilization. A depth helmet and measuring probe are used to check the stability of the patient's head position relative to the GTC frame. However, displacement error caused by rotation of the patient's head is not considered in the depth-helmet measurement. Consequently, displacement of the isocenter position cannot be confirmed by the measurement obtained from the depth helmet. In this study, we evaluated the precision of reproducibility by comparing measurement values of the depth helmet with the displacement of anatomical position on a CT image. We analyzed 21 setups of 8 patients immobilized for SRT using the GTC frame, between June 2001 and June 2003. The reproducibility of the GTC frame was checked at each treatment by comparing it with the treatment planning position. The average discrepancy of the GTC frame set-up measured by the depth helmet was 0.6 mm, with a standard deviation of 0.3 mm. The result measured by CT was 0.7 mm, with a standard deviation of 0.4 mm. When the error of each measurement point was within 1.0 mm, the accuracy of relocation of the patient could be considered clinically acceptable. Displacement error not considered in the measurement of the depth helmet could be evaluated by using CT.