Dear Editor,

Fungal liver abscesses (FLA), mainly caused by Candida, are among the most serious complications in patients with acute leukemia [14]. Diagnosing FLA by imaging such as ultrasound (US) [5] or computed tomography (CT) [6] is easily performed when multiple nodular lesions in the patient's liver can be seen. However, these nodular lesions can still be found by imaging for a long time after clinical symptoms begin to improve with anti-fungal treatments. Therefore, it is very difficult to evaluate the early effectiveness of treatments. We first report here a case in which the use of 18F-Fluorodeoxyglucose (FDG) positron emission tomography (PET) imaging was more beneficial than US or CT in evaluating FLA in the early stage of treatment.

A 32-year-old woman with relapsed acute myeloid leukemia was given re-induction chemotherapy and achieved second complete remission. Allogeneic bone marrow transplantation from an HLA-identical unrelated donor was planned, but low-grade fever (37–38°C) appeared before the transplantation. At that time, her C-reactive protein (CRP) level was about 3 mg/dL, but her blood cultures and beta-d-glucan were negative. Abdominal US and CT revealed multiple low-density areas in her liver, which suggested the FLA or leukemic infiltration. FDG-PET imaging examination was performed, and FDG uptake was shown to be increased at these lesions (Fig. 1 (1)). She was diagnosed with FLA when the periodic acid-Schiff staining of the liver biopsy specimens showed granuloma and yeast-like fungus, although the strain of the fungus could not be identified. Liposomal amphotericin B and Micafungin were administered daily, but the low-grade fever continued, and her CRP gradually increased. Consequently, the scheduled transplantation was postponed. After a month, the antifungal treatment was changed to oral Voriconazole, then her fever lowered promptly and her CRP level decreased gradually. At 3 months after the anti-fungal treatment was started, she had no clinical symptoms, and her CRP level was negative. While the US and CT tests at that time showed that the number and sizes of the nodular lesions were the same or had decreased only slightly before the Voriconazole treatment, the FDG-PET results showed that FDG uptake at the nodular lesions was almost absent (Fig. 1 (2)a, b). Based on the clinical symptoms and findings of FDG-PET, we concluded that the FLA had improved. Five months later, the patient received allogeneic bone marrow transplantation from an HLA-identical unrelated donor while still being treated with Voriconazole. We found no recurrence of FLA just prior to and since the transplantation.

Fig. 1
figure 1

Comparison of FDG-PET (upper row), US (lower left), and CT (lower right) imaging for FLA. (1) Before the anti-fungal treatments, FDG-PET, and US imaging tests showed multiple nodular lesions in the patient's liver. (2) Three months after starting treatment, a the FLA was shown to be almost the same or had decreased only slightly by US; however, FDG uptake decreased dramatically and was nearly absent. b The abscess highlighted by the arrow was a residual lesion by plain CT; however, FDG uptake of the same part was nearly absent

This case suggests that FDG-PET can be used to evaluate the treatment for FLA at early stage more accurately than other imaging techniques because FDG uptake increases at lesions with active inflammation and disappears once inflammation is gone. Therefore, we highly encourage the use of FDG-PET exams when the effectiveness of anti-fungal agents for FLA is uncertain.