December 8, 2025BloodOpen Access

AI-assisted diagnosis of paroxysmal nocturnal hemoglobinuria

View Full Paper

Ask AI

Bookmark

GPT-4.5 achieved 89.7% diagnostic accuracy for paroxysmal nocturnal hemoglobinuria, compared to 80.9% for both Claude and Gemini.
Statistical analyses revealed significant performance differences among AI models, particularly when key diagnostic features were omitted.
Analysis of clinical vignettes shows robust performance of AI even with reduced clinical information, but highlights critical thresholds for effective diagnosis.
Cluster analysis indicated that removing both hemoglobinuria and laboratory indicators significantly impairs diagnostic accuracy.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

AI models demonstrate improved diagnostic accuracy in paroxysmal nocturnal hemoglobinuria, indicating potential for timely recognition of rare conditions.

Cite This Study

Bowen et al. (2025) studied this question.

synapsesocial.com/papers/69362f574fa91c937236dacd https://doi.org/10.1182/blood-2025-4670

View Full Paper

Ask AI

Bookmark

Implication

AI models demonstrate improved diagnostic accuracy in paroxysmal nocturnal hemoglobinuria, indicating potential for timely recognition of rare conditions.

AI-assisted diagnosis of paroxysmal nocturnal hemoglobinuria

Key Points

Discussion

Implication

Cite This Study

AI-assisted diagnosis of paroxysmal nocturnal hemoglobinuria

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider