Anikó Ekárt

Objective: To review multimodal artificial intelligence (AI) models for predicting adverse outcomes in pneumonia, with emphasis on prognosis, validation, benchmarking, explainability, and clinical implementation readiness. Materials and Methods: We conducted a PROSPERO-registered systematic review. PubMed, Scopus, IEEE Xplore, and supplementary register/citation-search sources were searched between 01 January 2000 and 27 February 2026. Eligible studies integrated at least two heterogeneous clinical data modalities within one predictive framework for pneumonia-related deterioration, mortality, escalation of care, ventilation, or related adverse outcomes. We extracted modality composition, fusion architecture, outcome alignment, validation, performance metrics, clinical benchmarking, explainability, predictive uncertainty, workflow integration, governance reporting, and risk of bias. Results: Of 10,460 identified records, 49 studies were included. The evidence base was dominated by COVID-era studies. Structured electronic health record data, laboratories, and vital signs were more common than imaging or clinical notes. Early fusion was the predominant strategy, while attention-based and graph-based fusion were rare. External validation was reported in 19/49 studies, calibration in 9/49, temporal validation in 5/49, decision curve analysis in 3/49, and prospective evaluation in 3/49. Incremental AUROC over the best unimodal baseline was computable in 19 studies and was usually modest. Discussion: The literature emphasised retrospective discrimination more than clinical utility, with limited reporting of added value, workflow, predictive uncertainty, and governance. Conclusion: Multimodal pneumonia prognosis research is active but not yet implementation-ready. Future studies should prioritise externally validated, calibrated, explainable, clinician-aligned, and workflow-aware models.

The advent of large language models (LLMs), their sudden popularity, and their extensive use by an unprepared and, therefore, unskilled public raises profound questions about the societal consequences that this might have on both the individual and collective levels. In particular, the benefits of a marginal increase in productivity are offset by the potential for widespread cognitive deskilling or nonskilling. While there has been much discussion about the trust relationship between humans and generative AI technologies, the long-term consequences that the use of generative AI can have on the human capability to make trust decisions in other contexts, including interpersonal relations, have not been considered. We analyze this development using the functionalist lens of a general trust model and deconstruct the potential loss of the human ability to make informed and reasoned trust decisions. From our observations and conclusions, we derive a first set of recommendations to increase the awareness of the underlying threats, laying the foundation for a more substantive analysis of the opportunities and threats of delegating educative, cognitive, and knowledge-centric tasks to unrestricted automation.

We introduce a phylogeny-aware framework for predicting linear B-cell epitope (LBCE)-containing regions within proteins. Our approach leverages evolutionary information by using a taxonomic scaffold to build models trained on hierarchically structured data. The resulting models present performance equivalent or superior to generalist methods, despite using simpler features and a fraction of the data volume required by current state-of-the-art predictors. This allows the utilization of available data for major pathogen lineages to facilitate the prediction of LBCEs for emerging infectious agents. We demonstrate the efficacy of our approach by predicting new LBCEs in the monkeypox (MPXV) and vaccinia viruses. Experimental validation of selected targets using sera from infected patients confirms the presence of LBCEs, including candidates for the differential serodiagnosis of recent MPXV infections. These results point to the use of phylogeny-aware predictors as a useful strategy to facilitate the targeted development of immunodiagnostic tools.

The identification of linear B-cell epitopes is an important task in the development of vaccines, therapeutic antibodies and several diagnostic tests. Recently, organism-specific training has been shown to improve prediction performance for data-rich organisms. This article investigates the limits of organism-specific training for epitope prediction, by systematically quantifying the effect of the amount of training data on the performance of the models developed. The results obtained indicate that even models trained on small organism-specific data sets can outperform similar models trained on much larger heterogeneous and mixed data sets, as well as widely-used predictors from the literature, which are trained on heterogeneous data. These results suggest the potential for a much broader applicability of pathogen-specific models, which can be used to accelerate the development of diagnostic tests and vaccines in the context of emerging pathogens and to support faster responses in future disease outbreaks.