Whole-brain metabolic activity on pretreatment [18 F]FDG PET/CT provides prognostic information beyond metabolic tumour burden in patients with breast cancer.

Tatoğlu, Mehmet Tarık; Canıtez, Burak; Uslu, Hatice; İbişoğlu, Ebru; Tatoğlu, Ferda Yerdelen · Eur J Nucl Med Mol Imaging · 2026

retrospective_cohort · Level III

Where this comes from

Abstract

To investigate whether whole-brain (WB) metabolic activity measured on pretreatment [18 F]FDG PET/CT provides prognostic information beyond global metabolic tumour burden in patients with breast cancer. This single-centre retrospective cohort included 358 patients with biopsy-confirmed breast cancer who underwent pretreatment [18 F]FDG PET/CT for staging. WB SUVmean, WB SULmean, and total metabolic tumour volume (TMTV) were measured using a semi-automated PET/CT analysis workflow. The primary endpoint was overall survival (OS). Sequential Cox models evaluated the incremental prognostic value of WB SUVmean beyond clinical variables and TMTV; WB SULmean was assessed in sensitivity analysis. Complete-case data were available for 354 patients. In univariable analysis, higher TMTV was associated with worse OS (HR 1.167 per 100 mL; P < 0.001), whereas higher WB SUVmean (HR 0.785; P < 0.001) and WB SULmean (HR 0.733; P < 0.001) were associated with better OS. After adjustment for age, body mass index, disease extent, fasting glucose, and TMTV, WB SUVmean remained independently associated with OS (HR 0.810, 95% CI 0.679-0.967; P = 0.019), while TMTV retained prognostic value (HR 1.104, 95% CI 1.053-1.157; P < 0.001). WB SULmean showed similar results in sensitivity analysis (HR 0.764, 95% CI 0.586-0.997; P = 0.047). Combined TMTV/WB SUVmean stratification identified three prognostic groups (global log-rank P = 2.14 × 10⁻¹⁰). Pretreatment WB metabolic activity provides prognostic information beyond global metabolic tumour burden in breast cancer and may represent a complementary host-related prognostic signal, although the underlying biological mechanisms require further validation.