Truace
Subscribe

An interpretable machine learning model integrating mucosal and cellular immunity to predict SARS-CoV-2 reinfection: a multicenter cohort study

SARS-CoV-2 reinfection remains a significant public health challenge that requires accurate risk assessment across populations. Using prospectively collected data from two independent cohorts, we developed and externally validated an interpretable machine learning model for predicting symptomatic SARS-CoV-2 reinfection. The cohort comprised 1,156 adults from Wuxi and Nantong, China, with a median follow-up duration of 10 months. Reinfection occurred in 128 participants (11.1%). The model integrated eighteen clin…

By the Truvace desk1 min read1 sourcePeer-reviewedTRV-2026-2218· cite & verify
An interpretable machine learning model integrating mucosal and cellular immunity to predict SARS-CoV-2 reinfection: a multicenter cohort study
SARS-CoV-2 scanning electron microscope image by National Institute of Allergy and Infectious Diseases (NIAID). CC BY 2.0 · https://creativecommons.org/licenses/by/2.0

In brief

Researchers built a machine learning model to predict symptomatic SARS-CoV-2 reinfection using prospectively collected data from 1,156 adults in Wuxi and Nantong, China. Participants were followed for a median of 10 months, and 128 of them (11.1%) were reinfected.

The work suggests that mucosal and cellular immune measurements could guide reinfection risk assessment where population immunity varies. The abstract reports external validation but gives no figure for it, and the findings come from two Chinese cohorts, so wider applicability is not shown.

Main points

  1. The cohort included 1,156 adults from Wuxi and Nantong, China, followed for a median of 10 months.
  2. Reinfection occurred in 128 participants, or 11.1%.
  3. The model combined eighteen clinical and immune features using extreme gradient boosting.
  4. Higher nasal spike-specific sIgA was associated with lower reinfection risk in SHAP analysis.
  5. Spike-specific CD4+ central memory T cells showed a non-linear association with reinfection.

The gain

An XGBoost model using nasal spike-specific sIgA and CD4+ central memory T cell measures, among other features, can stratify adults by symptomatic SARS-CoV-2 reinfection risk, with external validation.

The rundown

The model was developed from two independent prospectively followed cohorts and tested in internal and external validation sets. Its internal validation AUC was 0.835, and the authors describe strong discrimination in the external cohort.

Interpretability analysis with SHAP identified which immune markers drove predictions. Nasal sIgA pointed toward protection at higher levels, while the CD4+ TCM relationship was not linear, so more of the marker does not simply mean lower risk.

The authors present the tool as a framework for immune-informed risk stratification and for managing reinfection risk in settings with heterogeneous population immunity.

Sources

  1. Peer-reviewedEmerging Microbes & Infections2026-10-21

The debate