PF-Atlas › Treatments › PMFF-Net AI Diagnostic
PMFF-Net AI Diagnostic
Parallel Multi-scale Feature Fusion Network classifies ILD patterns from CT with 97.6% accuracy
PF-Atlas model — Computational / preclinical hypothesis — not clinically validated. How PF-Atlas grades evidence.
Development stage
Available Now
Evidence level
Computational
Mechanism class
Parallel Multi-scale Feature Fusion Network classifies ILD patterns from CT with 97.6% accuracy
Timeline
< 1 year
Risk profile
Low
Developer
Multiple
Key researchers
Multiple
Overview
Transformer+CNN hybrid architecture. Diagnoses UIP pattern, NSIP, and other ILD subtypes. Enables earlier treatment initiation. Reduces diagnostic delay from years to minutes.
How PMFF-Net AI Diagnostic fits into IPF treatment
PMFF-Net AI Diagnostic is classified as Computational at the Available Now stage. Explore how it compares to other options and where it sits in the development pipeline:
Other IPF treatments
Medical disclaimer. This page summarizes published research and PF-Atlas model output about PMFF-Net AI Diagnostic for information only. It is not medical advice or a treatment recommendation. Drug availability, approval status and evidence change — verify with a pulmonologist and official sources (FDA, ClinicalTrials.gov) before any decision.