Heart failure & cardiomyopathiesDiagnostic or prognostic study
Natural Language Processing Framework in Early Detection of Amyloidosis
- development and validation of natural language processing (NLP)-predictive models
Authors' conclusion
NLP-based prediction models have an excellent diagnostic performance for SA in patients with HF and/or neuropathy. NLP-enabled workflow can potentially lead to earlier diagnosis of amyloidosis, starting therapy and possibly improving outcomes.
Study context
- Population
- Patients with a diagnosis of heart failure (HF) and/or neuropathy and sufficient chart datadevelopment cohort (Oregon Health & Science University) 109,275 · validation cohort (Baylor Scott & White Health System, HF) 116,658
- Participants
- 109,275 vs 116,658
- Comparison
- NLP-predictive models for the diagnosis of systemic amyloidosis (SA), developed and validated in an independent population
- development cohortn = 109,275
- validation cohortn = 116,658
Performance
As reported in the abstract
Diagnostic accuracy in predicting SA, development cohort (gradient boosted regression applied to combined notes, best of 420 models)
| Test or model | AUC | sensitivity | specificity | positive predictive value | negative predictive value |
|---|---|---|---|---|---|
| development cohort | 0.88 | 73.5% | 86.0% | 3.4% | 99.8% |
Further findings
Cohort characteristics
age development cohort 57.6 ± 18.6 years · validation cohort 71.6 ± 14.3 years
females
White
- Diagnostic accuracy in predicting SA, validation cohort
- 54.1% · sensitivity · specificity 94.0% · positive predictive value 8.8% · negative predictive value 99.1% · AUC 0.85
Original publication
Natural Language Processing Framework in Early Detection of Amyloidosis. JACC Adv. doi:10.1016/j.jacadv.2026.102818
Summarised from the published abstract; every number is checked against it