Autoinflammatory & rare systemic diseases
Amyloidosis secondary to familial Mediterranean fever: machine learning–based prediction models
Authors' conclusion
Amyloidosis, the most severe FMF complication, is driven by key clinical and genetic factors, with ensemble machine-learning models outperforming conventional regression for early risk prediction.
Study context
- Population
- FMF patients diagnosed 1990–2022 at Cerrahpaşa Faculty of Medicine58 (9.4%) with amyloidosis
- Participants
- 615 patients
- Exposure
- Logistic regression vs random forest vs gradient boosting for predicting amyloidosis
- Outcome
- Prediction of amyloidosis (AUC on held-out test set)
Methodological detailssingle-centre, machine learning prediction models, held-out test set
- Single-centre
- Machine learning prediction models
- Held-out test set
Main finding
As reported in the abstract
AUC for amyloidosis prediction, test set
Random forest
0.784
AUC
Logistic regression
0.766
AUC
Gradient boosting
0.749
AUC
Observational data: associations, not effects.
Further findings
Independent associations with amyloidosis (multivariate analysis)
- Disease duration, M694V homozygosity, comorbidity, parental consanguinity, frequent infections and erysipelas-like erythema independently associated with amyloidosis
SHAP analysis
- Strongest contributors: M694V homozygosity, myalgia, erysipelas-like erythema, disease duration, arthritis
Original publication
Amyloidosis secondary to familial Mediterranean fever: machine learning–based prediction models. Rheumatology (Oxford). Published 22 April 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →