Unproven hypothesis that leflunomide is better than methotrexate as measured by magnetic resonance imaging: comment on the article by Reece et al.