Machine learning prediction of
immune TTP diagnosis with Missing Data

TTP-MD

Risk Interpretation

  • High probability of immune TTP.
    Urgent hematology evaluation and empiric treatment should be strongly considered.

  • High Risk
    (≥50%)

  • Immune TTP is a significant possibility.
    Clinical context and competing diagnoses should guide further evaluation and management.

  • High-Intermediate Risk
    (25–49.9%)

  • Low-Intermediate Risk
    (5–24.9%)

  • Immune TTP is less likely but cannot be excluded.
    Continued clinical assessment and diagnostic evaluation remain important.

  • Low Risk
    (<5%)

  • Immune TTP is unlikely.

About This Calculator

Developed at the University of Utah (Abou-Ismail et al.) using an XGBoost machine learning model trained on a multicenter development cohort (University of Utah, Case Western Reserve University, Rochester General Hospital, and University of Illinois Peoria). The model was externally validated using an independent cohort at Mayo Clinic Rochester. Data was presented at the American Society of Hematology Annual Meeting (2025) and is currently under peer-review.

  1. Mouhamed Yazan Abou-Ismail, Meera Sridharan, Chong Zhang, Peter Kouides, Ming Lim; External validation of TTP-14: A novel machine learning model that improves rapid diagnosis of immune TTP. Blood 2025; 146 (Supplement 1): 850. doi: https://doi.org/10.1182/blood-2025-850