In the aftermath of a major radiological accident, the medical management of overexposed individuals will rely on the determination of the dose of ionizing radiations absorbed by the victims. Because people in the general population do not possess conventional dosimeters, after the fact dose reconstruction methods are needed. Free radicals are induced by radiations in the tooth enamel of victims, in direct proportion to dose, and can be quantified using Electron Paramagnetic Resonance (EPR) spectrometry, a technique that was demonstrated to be very appropriate for mass triage. The presence of dimethacrylate based restorations on teeth can interfere with the dosimetric signal from the enamel, as free radicals could also be induced in the various composites used. The aim of the present study was to screen irradiated composites for a possible radiation-induced EPR signal, to characterize it, and evaluate a possible interference with the dosimetric signal of the enamel. We investigated the most common commercial composites, and experimental compositions, for a possible class effect. The effect of the dose was studied between 10 Gy and 100 Gy using high sensitivity X-band spectrometer. The influence of this radiation-induced signal from the composite on the dosimetric signal of the enamel was also investigated using a clinical L-Band EPR spectrometer, specifically developed in the EPR center at Dartmouth College. In X-band, a radiation-induced signal was observed for high doses (25-100 Gy); it was rapidly decaying, and not detected after only 24h post irradiation. At 10 Gy, the signal was in most cases not measurable in the commercial composites tested, with the exception of 3 composites showing a significant intensity. In L-band study, only one irradiated commercial composite influenced significantly the dosimetric signal of the tooth, with an overestimation about 30%. In conclusion, the presence of the radiation-induced signal from dental composites should not significantly influence the dosimetry for early dose assessment.