What question did this study set out to answer?

To optimize machine learning models for diagnosing neurosyphilis based on international diagnostic criteria.

April 3, 2026SHILAP Revista de lepidopterología0 citationsOpen Access

Optimizing Machine Learning–Based Diagnosis of Neurosyphilis in HIV-Negative Patients: A Multicenter, Real-World Comparison of International Diagnostic Criteria

YJYinbo JiangKey Laboratory of Guangdong Province XDXinlei DengUniversity at Albany, State University of New York XZXiaohui ZhangSouthern Medical University

Key Points

To optimize machine learning models for diagnosing neurosyphilis based on international diagnostic criteria.
Developed machine learning models using various diagnostic guidelines.
Collected data from 1,648 suspected neurosyphilis cases across four centers.
Utilized five machine learning algorithms with randomized search and cross-validation for model training.
Performance assessed using metrics like AUC, PRAUC, and calibration.
All models achieved excellent discrimination with AUC and PRAUC values above 0.90.
Neurological symptoms, CSF protein, and CSF white blood cell count were identified as key predictors.
Model performance showed some variation based on the diagnostic guideline used.

Abstract

AbstractNeurosyphilis continues to rise globally, yet diagnosis remains challenging, often requiring multidisciplinary expertise and multiple CSF assays that are difficult to access in resource-limited settings. We developed and compared machine-learning (ML) models tailored to six international diagnostic guidelines and built a free, web-based tool for guideline-adapted decision support. We assembled 1,648 suspected neurosyphilis cases from four centers, using Guangzhou as the training cohort and Beijing, Xiamen (China) and Seattle (USA) as external validation cohorts. Five algorithms (Random Forest, Adaboost, SVC, NuSVC, XGboost) were trained with randomized search and three-fold cross-validation; performance was assessed by AUC, PRAUC, calibration, decision-curve net benefit, Brier score, and SHAP for model explainability. Across models and guidelines, neurological symptoms, CSF protein, and CSF white blood cell count consistently ranked as the strongest predictors. All models achieved excellent discrimination (AUC and PRAUC >0.90) with good calibration, reliability, and positive clinical utility, though performance varied modestly by guideline. These findings indicate that the optimal ML approach depends on the diagnostic definition applied. Our freely available online tool operationalizes these models to provide clinicians worldwide with context-adapted support aligned to local criteria.

KI fragen

Bookmark

View Full Paper