Splenic 18F-FDG uptake on baseline PET/CT is associated with oncological outcomes and tumor immune state in uterine cervical cancer

2020 
Abstract Objective The spleen represents an important contributor to tumor immune escape, but the relevance of increased splenic metabolic activity remains to be fully elucidated. Methods We retrospectively measured the spleen-to-liver standard uptake value (SLR) on 18F-FDG PET/CT examinations of 92 consecutive patients with FIGO stage IB1 to IVA cervical cancer and integrated the results with survival, response to treatment, tumor immune infiltrate, and baseline characteristics. Results SLRmax > 0.92 (p = .026) and SLRmean > 0.94 (p = .005) were significantly associated with decreased DFS in univariable analysis. Multivariable models were built using best subset selection; ΔSLRmax and either SLRmax or SLRmean were consistently selected, strongly reinforcing the association between SLR variables and DFS in relation to potential confounders (all models p ≤ .002). Independent associations were found for SLRmax using multivariable Cox regression models for DFS (all p ≤ .003). Further, uni- and multivariable analyses demonstrated the negative impact of higher SLR values on pathological complete response. A statistically significant higher proportion of patients with high SLRmax had a dense infiltrate of CD20+ (p = .036) and CD68+ (p = .015) immune cells, as well as PD-L1+ tumor cells (p = .019) as compared to those with low SLRmax. Finally, high SLRmax status was neither associated with systemic inflammatory markers (except for an increased white blood cell count; p = .038), nor with clinically overt infection. Conclusion This hypothesis-generating study provides the first evidence that increased splenic metabolic activity is a negative prognostic and predictive biomarker in locally advanced cervical cancer. In addition, it might help to discriminate immunologically ‘hot’ from ‘cold’ cervical tumors.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    40
    References
    4
    Citations
    NaN
    KQI
    []