Explainable AI boosts radiologist trust in PET/CT lung cancer staging

In a study of 10 UK nuclear medicine radiologists, all explainable AI (XAI) methods significantly increased willingness to adopt an AI staging system for PET/CT lung cancer over a black-box model (p<0.05). Explanations were rated useful (p<0.001), but a depth-usability trade-off…

• User-centred observation study: 10 radiologists from 8 institutions evaluated a simulated AI-CDSS for TNM staging of lung cancer with whole-body PET/CT.
• Three explanation types tested: input feature attribution, high-level concepts, and global algorithmic transparency.
• Qualitative analysis highlighted a trade-off between explanation depth and clinical usability.

Small sample (10 radiologists) and simulated AI setting; no external validation or real-world deployment data.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.