Objective: To compare the performance of observers with different knowledge when detecting periapical lesions. Methods: We obtained nine dry human hemimandibles. Periapical lesions were created with a bur and radiographic sets were taken for each hemimandible in four stages (initial, no cancellous bone, removal of 0.5 and 1 mm of cortical bone). Each radiographic set comprised five digital radiographs taken at different angulations (ortho, mesio, disto, lingual and vestibuloradial). Thirty six radiographs sets were obtained with a total of 180 images. The images were evaluated by experienced examiners and students. Mann Whitney test and Spearman's correlation were applied for data comparison. Results: In all study stages, no statistically significant differences were found (P ≤ 0.05) between observers, with a positive association established by Spearman's correlation in two stages (removal of 0.05 and 1 mm of cortical bone; 0.866 and 0.500, respectively). Conclusions: There were no statistically significant differences in diagnostic accuracy between experienced examiners and students when detecting simulated periapical lesions in dry human mandibles.