Résumé : OBJECTIVES: In nuclear medicine, reproducibility studies allow the assessment of the robustness of a technique. They could also be used in continuing education. This study evaluates this educational impact. METHODS: Two series of clinical data and I-123 fluoropropyl-carbomethoxy-iodophenyl- nortropane single photon emission tomography images from 12 patients were prepared before initiating the study. Each series covered similarly a wide spectrum of clinical situations and images. Nuclear medicine physicians having experience with I-123 fluoropropyl-carbomethoxy-iodophenyl-nortropane single photon emission tomography were recruited and assigned two similar groups regarding their level of experience and type of institutions from which they were issued. Clinical data and images were transmitted by e-mail. For each case, observers had to choose among three answers: normal, equivocal, and abnormal. Answers were returned anonymously. Before the second series was sent, observers from group A had a collegiate discussion on the results of the first series of cases. Observers from group B received the second series without any information regarding analysis of series 1. RESULTS: In group A, median agreement increased from 75% (series 1) to 100% (series 2); in group B it was 75% (series 1) and 87% (series 2). In group A, a 100% agreement was observed for four cases (series 1) and for eight cases (series 2). In group B, a 100% agreement was observed for four cases for both series. CONCLUSION: These preliminary results, obtained in a small sample of observers, suggest that participating and discussing results of interobserver reproducibility studies seems to have a positive educational impact and therefore improves interobserver reproducibility. © 2011 Wolters Kluwer Health