PET imaging of Aspergillus infection using Zirconium-89 labeled anti-ß-glucan antibody fragments.

PURPOSE: Invasive fungal diseases, such as pulmonary aspergillosis, are common life-threatening infections in immunocompromised patients and effective treatment is often hampered by delays in timely and specific diagnosis. Fungal-specific molecular imaging ligands can provide non-invasive readouts of deep-seated fungal pathologies. In this study, the utility of antibodies and antibody fragments (Fab) targeting ß-glucans in the fungal cell wall to detect Aspergillus infections was evaluated both in vitro and in preclinical mouse models. METHODS: The binding characteristics of two commercially available ß-glucan antibody clones and their respective antigen-binding Fabs were tested using biolayer interferometry (BLI) assays and immunofluorescence staining. In vivo binding of the Zirconium-89 labeled antibodies/Fabs to fungal pathogens was then evaluated using PET/CT imaging in mouse models of fungal infection, bacterial infection and sterile inflammation. RESULTS: One of the evaluated antibodies (HA-ßG-Ab) and its Fab (HA-ßG-Fab) bound to ß-glucans with high affinity (KD = 0.056 & 21.5 nM respectively). Binding to the fungal cell wall was validated by immunofluorescence staining and in vitro binding assays. ImmunoPET imaging with intact antibodies however showed slow clearance and high background signal as well as nonspecific accumulation in sites of infection/inflammation. Conversely, specific binding of [89Zr]Zr-DFO-HA-ßG-Fab to sites of fungal infection was observed when compared to the isotype control Fab and was significantly higher in fungal infection than in bacterial infection or sterile inflammation. CONCLUSIONS: [89Zr]Zr-DFO-HA-ßG-Fab can be used to detect fungal infections in vivo. Targeting distinct components of the fungal cell wall is a viable approach to developing fungal-specific PET tracers.

Kiñit classification in Ethiopian chants, Azmaris and modern music: A new dataset and CNN benchmark.

In this paper, we create EMIR, the first-ever Music Information Retrieval dataset for Ethiopian music. EMIR is freely available for research purposes and contains 600 sample recordings of Orthodox Tewahedo chants, traditional Azmari songs and contemporary Ethiopian secular music. Each sample is classified by five expert judges into one of four well-known Ethiopian Kiñits, Tizita, Bati, Ambassel and Anchihoye. Each Kiñit uses its own pentatonic scale and also has its own stylistic characteristics. Thus, Kiñit classification needs to combine scale identification with genre recognition. After describing the dataset, we present the Ethio Kiñits Model (EKM), based on VGG, for classifying the EMIR clips. In Experiment 1, we investigated whether Filterbank, Mel-spectrogram, Chroma, or Mel-frequency Cepstral coefficient (MFCC) features work best for Kiñit classification using EKM. MFCC was found to be superior and was therefore adopted for Experiment 2, where the performance of EKM models using MFCC was compared using three different audio sample lengths. 3s length gave the best results. In Experiment 3, EKM and four existing models were compared on the EMIR dataset: AlexNet, ResNet50, VGG16 and LSTM. EKM was found to have the best accuracy (95.00%) as well as the fastest training time. However, the performance of VGG16 (93.00%) was found not to be significantly worse (P < 0.01). We hope this work will encourage others to explore Ethiopian music and to experiment with other models for Kiñit classification.