Digital severity scoring and viral metagenomics: A feasibility study on integrated diagnosis of pediatric influenza-like illness.

INTRODUCTION

Metagenomic next-generation sequencing (mNGS) holds promise for identifying diverse pathogens in complex cases of influenza-like illness (ILI). Interpreting results requires comprehensive clinical context. We aimed to explore the feasibility of an integrated diagnostic approach by linking shotgun viral mNGS with standardized clinical data for unbiased ascertainment and hypothesis generation in pediatric ILI patients.

PATIENTS AND

METHODS

We studied a cohort of 6,073 pediatric ILI patients (mean age 3.1 years, range 0-18.8 years), assessed using the VIVI ScoreApp for immediate computation of Disease Severity and Risk Factor Scores. Nasopharyngeal samples were tested for nine respiratory viruses by PCR. In a nested pilot feasibility study, we linked the clinical dataset of 100 ILI patients with neurological complications (mean age 3.9 years, range 0-17.8 years) to additional viral mNGS. PCR and mNGS were compared by agreement rates and Cohen's κ for inter-method reliability.

RESULTS

In the pilot feasibility study, the mean VIVI Disease Severity Score was above the cohort average (>67th percentile, p < 0.0001), with 'age < 2 years' as the most prevalent risk factor (n = 44/100). mNGS identified 15 viruses, expanding the range of viral identifications by six viruses compared to PCR. Linking VIVI Scores with mNGS-discovered viruses suggested high disease severity. Sensitivity of mNGS was relatively low; overall agreement with PCR was 77-98 % and overall reliability was 'moderate' (κ scores of 0.1-0.85).

CONCLUSIONS

Digital surveillance tools can successfully integrate with mNGS to capture complex clinical patterns and generate data-driven hypotheses. Large-scale investigation and technical refinement are warranted.