Significant reduction in the diagnosis of possible and probable invasive fungal infection by applying the revised EORTC definitions of invasive fungal disease

Objectives: To assess the impact of the revised (2008) European Organization for Research and Treatment of Cancer (EORTC) criteria for the diagnosis of Invasive Fungal Infection (IFI) compared to the previous ones (2002) in defining "possible", "probable" or "proven" IFIin patients with Acute Myeloid Leukaemia (AML) or those undergoing Allogeneic Haemopoietic Stem Cell Transplantation (HSCT). Methods: Data from 330 patient-episodes in our institution between 09/2005 and 12/2007 were analyzed retrospectively. A patient-episode was defined as i) admission for AML chemotherapy, ii) admission for transplant conditioning and transplantation or iii) admission with an infective complication while receiving intensive chemotherapy or in the post transplant period. The scoring of the CT scans was performed by designated radiologists with a special interest in IFIs and the overall scoring and designation of infections as possible, probable or proven, using both the old and the new criteria, was performed independently by two members of the transplant team. Results: By applying the 2002 definitions, there was some evidence of IFIin 85 (26%) of all episodes. Out of these 85 cases, 73 (86%) were "possible", 10 (12%) were "probable" and 2 cases (2%) were "proven". These figures drop significantly by applying the revised criteria, to an overall evidence of IFIin 30 cases (9%), with 25 (83%) "possible" diagnoses, 3 (10%) "probable" and 2 (7%) "proven" (Table). There was overall 83% (n=275) conformity in results when applying the old and the revised criteria, including all the cases where there was no evidence of IFIat all. In 17% (n=57) of all cases, the estimated likelihood of infection was reduced by applying the new criteria, while there was no case where the probability was higher with the new definitions. In 49 of 71 cases a "possible" IFIby the old criteria was reduced to "none" (69% reduction), while in 7 of 10 cases a "probable" diagnosis by the old criteria was also reduced to "none" when the new criteria were applied (70% reduction). Overall, "positive" episodes (possible, probable and proven) were reduced by 65% when the revised criteria were applied (figure). Conclusions: Application of the revised EORTC definitions for Invasive Fungal Disease leads to a significant reduction of the cases that would be previously characterized as possible, but also as probable IFIs, with implications for the interpretation of previous therapeutic and diagnostic studies.