Febrile neutropenia, a serious complication that can occur during the treatment of hematological malignancies, can sometimes be fatal owing to fungal infection. Prospective randomized trials indicated the utility of liposomal amphotericin B or caspofungin as an empirical antifungal therapy. Itraconazole, a broad-spectrum tri azole antifungal agent, is poorly absorbed in the intestines after oral absorption and makes it difficult to achieve a stable serum drug concentration. Therefore, an intravenous formulation might offer a potentially safer and more effective alternative. To compare the efficacy and safety of empirical antifungal therapy, patients will be randomly assigned to either the liposomal amphotericin B 3.0 mg/kg once daily group or the intravenous itraconazole 200 mg dose group with five stratification factors (disease risk, previous antifungal prophylaxis, age, sex, and institute). The primary endpoint will be overall favorable response, comprising five secondary endpoints: successful treatment of baseline infection by the end of the treatment; absence of breakthrough infection; no discontinuation of the antifungal treatment due to drug-related toxicity; fever resolution during neutropenia; and 7-day survival after termination of the antifungal treatment. The target sample size of 850 subjects is sufficient to prove the non inferiority of itraconazole compared with liposomal amphotericin B, with a non-inferiority margin of 10%, one sided significance level of 5%, and power of 90%.