Pulse.

a daily field guide to health research that matters

◆ Console

‹ Sun · 12 Jul 2026
Near-term implementable finding

FUO-PETMamba: a pet maximum-intensity-projection-based artificial intelligence framework for aetiological classification of fever of unknown origin with multicentre validation.

An AI tool learns to distinguish fever causes on PET scans, potentially helping doctors diagnose cancers and infections more accurately and consistently.

This multicentre study developed and validated FUO-PETMamba, an AI framework that analyzes PET maximum-intensity-projection images to classify fever of unknown origin into malignant, infectious, autoimmune, and miscellaneous causes, showing good discrimination across development (n=355) and two external validation cohorts (n=195, n=131). Reader study evidence that AI assistance reduces experience-dependent diagnostic variability makes this a candidate clinical decision support tool, with malignancy discrimination being particularly relevant to hematologic cancer detection.

What the study was

Study design
retrospective_multicentre_external_validation
Population
Fever of unknown origin patients
Sample size
681
Category
Diagnostics
Maturity
Validated
Journal
European Journal of Nuclear Medicine and Molecular Imaging

Why it surfaced

FUO is a common presentation of lymphoma and other hematologic malignancies; AI-assisted triage of FUO PET/CT in malignancy vs infection vs autoimmune directly reduces diagnostic delay; multicentre external validation strengthens near-term clinical translation signal.

A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.