GLOW-FDG: Generalized cancer LesiOn Whole-body segmentation model for $^{18}$F-FDG-PET/CT

2026-07-04

AI Summary

This study addresses the challenges of manual lesion delineation in whole-body $^{18}$F-FDG-PET/CT imaging—namely its time-consuming nature, subjectivity, and limited scalability—by proposing the first deep learning model capable of cross-cancer and multi-center generalization. The model integrates multimodal PET and CT information, trained on 1,563 scans and validated on 185 independent external cases. It demonstrates significantly superior performance over existing public methods across breast cancer, lung cancer, head and neck cancer, and metastatic melanoma, achieving low false-positive rates and lesion segmentation accuracy approaching inter-expert variability. Furthermore, the model enables precise quantification of key imaging biomarkers, including total tumor burden and total lesion glycolysis.

Abstract

Whole-body fluorodeoxyglucose positron emission tomography combined with computed tomography is widely used in cancer care, but manual lesion delineation is slow, subjective, and difficult to scale. We present GLOW-FDG, an open-source artificial intelligence model for whole-body cancer lesion segmentation in fluorodeoxyglucose positron emission tomography and computed tomography. The model was trained on 1,563 scans spanning multiple cancer types and evaluated on 185 external scans from independent institutions. Across breast cancer, nonmetastatic and oligometastatic lung cancer, head and neck cancer, and metastatic melanoma, GLOW-FDG consistently outperformed publicly available benchmark models in lesion detection, while reducing false positives and maintaining strong segmentation accuracy. Quantification of total tumor burden and total lesion glycolysis was robust across cohorts, and performance approached the variability observed between expert radiation oncologists. These results support GLOW-FDG as a generalizable tool for automated cancer segmentation and quantitative imaging biomarker extraction in whole-body imaging.

Problem

• lesion segmentation
• FDG-PET/CT
• cancer imaging
• manual delineation
• whole-body imaging

Innovation

• whole-body segmentation
• FDG-PET/CT
• generalizable AI model
• quantitative imaging biomarkers
• cancer lesion detection