MedForge: Interpretable Medical Deepfake Detection via Forgery-aware Reasoning

2026-03-19

AI Summary

This work addresses the critical threat posed by high-fidelity medical image editing—capable of implanting or removing lesions—which undermines clinical trust and patient safety, while existing detection methods often operate as black boxes or lack medical grounding. To bridge this gap, we propose MedForge, the first deepfake detection framework for medical imaging that integrates physician examination guidelines with ground-truth forgery localization supervision. We construct MedForge-90K, a large-scale expert-guided dataset of forged medical images, and introduce a “localize-then-analyze” inference paradigm coupled with a forgery-aware GSPO alignment strategy to enable proactive, evidence-driven detection with interpretability. Leveraging a multimodal large language model, MedForge achieves state-of-the-art accuracy across 19 pathological editing types, generating medical reasoning chains that closely align with expert judgments and significantly outperform existing approaches.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Security and Privacy: Data transparency and provenance
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Text-guided image editors can now manipulate authentic medical scans with high fidelity, enabling lesion implantation/removal that threatens clinical trust and safety. Existing defenses are inadequate for healthcare. Medical detectors are largely black-box, while MLLM-based explainers are typically post-hoc, lack medical expertise, and may hallucinate evidence on ambiguous cases. We present MedForge, a data-and-method solution for pre-hoc, evidence-grounded medical forgery detection. We introduce MedForge-90K, a large-scale benchmark of realistic lesion edits across 19 pathologies with expert-guided reasoning supervision via doctor inspection guidelines and gold edit locations. Building on it, MedForge-Reasoner performs localize-then-analyze reasoning, predicting suspicious regions before producing a verdict, and is further aligned with Forgery-aware GSPO to strengthen grounding and reduce hallucinations. Experiments demonstrate state-of-the-art detection accuracy and trustworthy, expert-aligned explanations.

Problem

• medical deepfake
• forgery detection
• interpretable AI
• medical image manipulation
• trustworthy explanation

Innovation

• Medical Deepfake Detection
• Interpretable AI
• Forgery-aware Reasoning
• Expert-aligned Explanation
• Pre-hoc Interpretability