Background: Accurate and timely malaria diagnosis with species-level identification of Plasmodium parasites is critical for guiding effective treatment and disease management. Although light microscopy remains the diagnostic gold standard, its dependence on highly trained personnel limits accessibility, particularly in resource-constrained settings. Recent advances in deep learning have enabled automated image-based diagnosis with promising performance; however, reliable differentiation among Plasmodium species continues to pose a major challenge. This study aims to evaluate and compare the effectiveness of different deep learning architectures for automated species identification from microscopy images. Methods: Three deep learning architectures were systematically assessed: a convolutional backbone (ResNet50), a Vision Transformer (ViT), and a hybrid ResNetViT model. All models were trained from scratch, without using pretrained weights, on a dataset comprising real-world thick blood smear images augmented with publicly available microscopy data from Kaggle. In the hybrid architecture, the ResNet module was used to extract robust local morphological features, while the ViT component captured long-range dependencies and contextual relationships within images. Results: In cross-validation experiments, all three architectures consistently achieved high diagnostic performance. ResNet50 attained the highest accuracy (96.9%), an F1-score of 96.3%, and an ROC-AUC of 0.997. The Vision Transformer achieved 93.1% accuracy, 91.7% F1-score, and an ROC-AUC of 0.989. The hybrid ResNetViT reached 95.2% accuracy, 94.2% F1-score, and an ROC-AUC of 0.995. These results confirm that all architectures can reliably distinguish among Plasmodium species. Although ResNet50 achieved the highest raw accuracy, the hybrid model showed the most stable calibration and the closest qualitative agreement between its attention maps and expert-annotated parasite locations. Conclusions: The findings demonstrate that convolutional, transformer-based, and hybrid deep learning architectures can be successfully trained on real microscopy data for species-level malaria diagnosis. These results support the feasibility of deploying scalable, automated diagnostic systems to improve both accuracy and accessibility of malaria detection, particularly in resource-limited healthcare settings.
F. Branda, Lilia Andriani, Riccardo Lucis et al.· Infectious Disease Reports· 0 citations
Influenza A(H1N1)pdm09 remains one of the predominant seasonal influenza viruses worldwide and continues to evolve under the combined effects of host immunity, vaccination, and changing epidemiological conditions. However, the long-term impact of the COVID-19 pandemic on its evolutionary dynamics remains poorly understood. We investigated the genetic variability and phylodynamic evolution of the hemagglutinin (HA) and neuraminidase (NA) genes of Italian A(H1N1)pdm09 viruses collected between 2009 and 2026. Time-calibrated phylodynamic analyses, Bayesian Skyline Plots (BSPs), Lineages Through Time (LTT) graphs, principal component analysis (PCA), and codon-based selection analyses were used to characterize long-term evolutionary patterns. Both HA and NA followed continuous evolutionary trajectories, with a marked post-2020 genetic shift associated with the emergence of 6B.1A.5a.2-related clades. HA showed greater evolutionary variability than NA, whereas selection analyses identified only one positively selected site in HA and none in NA, consistent with reduced adaptive diversification in recent strains. Phylodynamic analyses revealed a marked decline in effective population size and lineage accumulation during the COVID-19 pandemic, followed by renewed expansion after 2022. Overall, Italian A(H1N1)pdm09 viruses exhibited reduced genetic diversity, limited evidence of positive selection, and coordinated genomic restructuring following the pandemic. These findings provide new insights into the long-term evolutionary dynamics of A(H1N1)pdm09 in Italy and reinforce the importance of sustained genomic surveillance for anticipating evolutionary changes and informing evidence-based public health strategies.
Maria Perra, Ilaria Deplano, I. Azzena et al.· Pathogens· 0 citations
INTRODUCTION
Marburg virus disease (MVD) is a highly lethal filovirus infection with case fatality rates of up to 88%. Although no virus-specific interventions have yet been approved, recent laboratory breakthroughs have markedly accelerated translational progress in vaccine and therapeutic development.
AREAS COVERED
A hybrid methodological design was employed, combining macro-level bibliometric analysis with a targeted qualitative narrative review. As per the findings, the human monoclonal antibody MR191N provides complete post-exposure protection in non-human primates. Advanced RNA-targeted siRNAs, antisense oligonucleotides, and small-molecule nucleoside analogues (e.g. galidesivir) demonstrate robust preclinical efficacy in suppressing viral replication. Prophylactic platforms - including adenovirus-vectored (ChAd3-MARV), rVSV-MARV, and mRNA-based vaccines - are progressing through early-phase clinical and preclinical evaluation. Bibliometric analysis, however, reveals pronounced geographic disparities: high-income nations dominate scientific output (United States: 31.7%), while endemic African countries remain severely underrepresented in global collaboration networks.
EXPERT OPINION
Addressing MVD requires a coordinated transition from experimental research to licensed, field-ready interventions. This necessitates adaptive trial designs, expanded diagnostic models, and synergistic combination therapies. Ultimately, correcting global research inequities and embedding robust, cross-border One Health surveillance infrastructure are essential to transforming scientific advances into equitable, rapid-response solutions for endemic regions.
Venkataramana Kandi, F. Branda, Mohammed Alissa et al.· Expert Review of Anti-Infect...· 0 citations
The study supports the potential usefulness of combining contact-based models with explainable graph neural networks for scenario-based epidemic analysis and suggests that the k-GCN model captures relevant temporal and structural dependencies in the simulated graph-organized data.
Francesco Branda, G. Ceccarelli, Massimo Ciccozzi et al.· Network Modeling Analysis in...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
