PURPOSE: To evaluate the neutron and photon dose equivalent rates at the treatment vault entrance (Hn,D and HG), and to study the secondary radiation to the patient in GRID therapy. The radiation activation on the grid was studied.METHODS: A Varian Clinac 23EX accelerator was working at 18 MV mode with a grid manufactured by .decimal, Inc. The Hn,D and HG were measured using an Andersson-Braun neutron REM meter, and a Geiger Müller counter. The radiation activation on the grid was measured after the irradiation with an ion chamber γ-ray survey meter. The secondary radiation dose equivalent to patient was evaluated by etched track detectors and OSL detectors on a RANDO(®) phantom.RESULTS: Within the measurement uncertainty, there is no significant difference between the Hn,D and HG with and without a grid. However, the neutron dose equivalent to the patient with the grid is, on average, 35.3% lower than that without the grid when using the same field size and the same amount of monitor unit. The photon dose equivalent to the patient with the grid is, on average, 44.9% lower. The measured average half-life of the radiation activation in the grid is 12.0 (± 0.9) min. The activation can be categorized into a fast decay component and a slow decay component with half-lives of 3.4 (± 1.6) min and 15.3 (± 4.0) min, respectively. There was no detectable radioactive contamination found on the surface of the grid through a wipe test.CONCLUSIONS: This work indicates that there is no significant change of the Hn,D and HG in GRID therapy, compared with a conventional external beam therapy. However, the neutron and scattered photon dose equivalent to the patient decrease dramatically with the grid and can be clinical irrelevant. Meanwhile, the users of a grid should be aware of the possible high dose to the radiation worker from the radiation activation on the surface of the grid. A delay in handling the grid after the beam delivery is suggested.