Temporal Fusion of RT-qPCR and IgM/IgG Serology for Nipah Virus Diagnosis

What is it about?

Nipah virus infection is a high-consequence zoonosis of South and South-East Asia for which laboratory confirmation rests on two assay families whose informativeness moves in opposite directions across the course of illness. Reverse transcription quantitative polymerase chain reaction performs best while viral nucleic acid remains detectable in blood and the oropharynx whereas enzyme-linked immunosorbent assays for immunoglobulin M and immunoglobulin G become informative as nucleic acid disappears. Prevailing practice interprets these assays as independent dichotomous results obtained at a single moment, discarding quantitative magnitude, specimen provenance and the day of illness on which each specimen was taken. This review critically evaluates whether the longitudinal structure of the two assay families can be exploited jointly, and what explainable artificial intelligence would need to deliver before such a model could be used in outbreak conditions. Human clinical studies describing the temporal behaviour of molecular and serological markers were appraised alongside the methodological literature on diagnostic accuracy without an adequate reference standard, and the computational literature on temporal and multimodal fusion and on explanation methods in clinical decision support. The evidence consistently indicates an ordered transition in which clearance of viral nucleic acid from blood precedes the appearance of virus-specific immunoglobulin G, and that the combination of detectable viraemia with absent immunoglobulin G at presentation carries adverse prognostic weight. Confidence in these observations is nevertheless constrained by very small samples, over-representation of the minority who survive, the absence of an international nucleic-acid standard, and geographical concentration of evidence in Bangladesh and Kerala. No clinical prediction model that fuses molecular and serological results over time was identified for this infection. A conceptual dynamic framework is developed that separates established temporal relationships from hypothesised gains, specifies the evaluation and governance requirements that would apply, and identifies prospective serial-sampling cohorts as the rate-limiting evidential step rather than model architecture.

Featured Image

Why is it important?

This review is bounded to the diagnosis of human Nipah virus infection using molecular and serological assays interpreted as a function of time since symptom onset, and to the computational and methodological requirements of combining them. Animal and environmental surveillance, vaccine and therapeutic development, ecological modelling of spillover risk, genomic epidemiology, and in silico drug discovery fall outside the scope except where they bear directly on the interpretation of human diagnostic results. Imaging-based diagnosis is not covered, since imaging findings in this infection are non-specific and no imaging-based classifier has been reported. The gap addressed is specific. Clinical descriptions of Nipah virus infection contain repeated, mutually consistent observations that molecular and serological positivity are ordered in time and that their joint pattern is associated with outcome, yet no published work converts these observations into a formal inferential structure, and the diagnostic reviews that summarise the field do not identify this as a deficiency. Conversely, the computational methods that would be required are mature in other clinical domains but have not been applied to any henipavirus. The result is that the temporal information generated routinely during outbreak investigations is recorded and then discarded. The intended contribution is analytical rather than empirical. No new data are presented, no model is fitted, and no claim is made that the proposed framework would improve diagnostic performance. The framework is offered as a structured research agenda whose principal value is to make explicit what would have to be measured, and how it would have to be evaluated, before temporal fusion could be considered for use. Four objectives follow. The first is to synthesise and critically appraise the human clinical evidence on the temporal behaviour of molecular and serological markers in Nipah virus infection, distinguishing well-supported patterns from observations resting on single small series. The second is to evaluate the methodological quality of that evidence, with particular attention to the absence of an adequate reference standard, the over-representation of survivors, and assay standardisation. The third is to assess which temporal and multimodal modelling strategies, and which explanation approaches, are defensible given the quantity and quality of data that realistically exist. The fourth is to develop a conceptual dynamic diagnostic framework that separates relationships already established by evidence from those that remain hypothesised, and to specify the validation, calibration and governance conditions under which such a framework could be tested.

Perspectives