Invasive fungal disease (IFD) is a major infectious complication in patients with hematological malignancies.In this study,we examined 4889 courses of chemotherapy in patients with hematological diseases to establish a training dataset (n--3500) by simple random sampling to develop a weighted risk score for proven or probable IFD through multivariate regression,which included the following variables:male patients,induction chemotherapy for newly diagnosed or relapsed disease,neutropenia,neutropenia longer than 10 days,hypoalbuminemia,central-venous catheter,and history of IFD.The patients were classified into three groups,which had low (0-10,～1.2％),intermediate (11-15,6.4％),and high risk (＞ 15,17.5％) of IFD.In the validation set (n =1389),the IFD incidences of the groups were ～1.4％,5.0％,and 21.4％.In addition,we demonstrated that antifungal prophylaxis offered no benefits in low-risk patients,whereas benefits were documented in intermediate (2.1％ vs.6.6％,P =0.007) and high-risk patients (8.4％ vs.23.3％,P =0.007).To make the risk score applicable for clinical settings,a pre-chemo risk score that deleted all unpredictable factors before chemotherapy was established,and it confirmed that anti-fungal prophylaxis was beneficial in patients with intermediate and high risk of IFD.In conclusion,an objective,weighted risk score for IFD was developed,and it may be useful in guiding antifungal prophylaxis.