PURPOSE: Radiation oncologists are responsible for deciding which day-to-day variations are acceptable or not in the treatment setup. However, properly qualified and trained radiation therapists might be capable to perform image registration. We evaluated in our centre the capability and accuracy of radiation therapists to validate positioning images in a prospective study. METHODS AND PATIENTS: A total of 84 patients treated for prostate, head and neck, lung or breast cancer was prospectively and randomly included from July 2011 to July 2013 in radiotherapy unit of our institution. For each patient, three positioning images were randomly analysed. Two radiation oncologists analysed all positioning images and shifts decided by the radiation therapists in an independent and blinded way. The radiation oncologists had to decide whether to validate or not this shift and give a corresponding additional shift, if any. A theoretical disagreement rate less than 5% between radiation therapists and radiation oncologists was planned. RESULTS: A total of 240 images were analysed (head and neck: 15.0%; prostate: 14.2%; breast: 55.0%; lung: 15.8%). The global disagreement between radiation oncologists and radiation therapists for all the images analysed was 2.5% 95% confidence interval (95% CI) [1.0-5.0], corresponding to six images out of 240. A 100% agreement was reached for prostate and lung images, a 97.2% agreement for head and neck images and a 96.2% agreement for breast images. CONCLUSIONS: The radiation therapist validation for repositioning images seemed accurate for image-guided radiotherapy in our institution. Periodic evaluation and in-house training are warranted when routine delegation of image registration to radiation therapists is considered.