Digital subtraction is a common technique used in radiological studies of periapical lesions and other dental disorders, where treatment has to be evaluated over time. This paper shows a fast and reliable registration method for subtracting two digitized radiographs showing an unpredicted mismatch. An optimal affi ne transformation is found using an adaptive Genetic Algorithm (GA) as the optimizing strategy and a correlation ratio as the similarity measure. When processing pairs of standard intra-oral radiographs, the parallel GA implemented takes advantage of the CPU idle cycles of a computational grid, resulting in an application that exploits an existent infrastructure with a computational time of twenty seconds. By using an existing hardware infrastructure and software of free distribution, the proposed approach can * Correspondencia: edromero@unal.edu.co Direccion postal: Universidad Nacional de Colombia, Ciudad Universitaria. Centro de Telemedicina. Carrera 30 No. 45-03, Bogota DC. Colombia Recibido: Marzo 2 de 2007. Aceptado: Julio 11 de 2007. MANANA G., ROMERO E. Y FORERO M. Volumen 15 • No. 2 Julio de 2007 181 be used in public hospitals and other institutions carrying low budget. The validation process shows that there are not signifi cant differences between the automatic system and the manual registration. The results show that in endodontic surgery the technique of digital radiographic subtraction is a determining tool in the evaluation of post-surgical bone mineralization. Furthermore, it will be useful in clinical environments where there is not viable to follow a standardization protocol.