In-vivo fluorescence lifetime tomography for detection and quantification of programmed death ligand-1

Programmed death ligand-1 (PD-L1) is the only clinically approved predictive biomarker for cancer immunotherapy; however, its expression is highly heterogeneous and routinely assessed by ex vivo immunohistochemistry. To enable spatially resolved, noninvasive PD-L1 quantification, we employed fluorescence lifetime (FLT) tomography with a normalization strategy designed to account for nonspecific probe uptake. C57BL/6 mice bearing orthotopic RIL-175 liver tumors (n=10) were injected with alphaPDL1-800 (anti-PD-L1 antibody conjugated to IRDye 800CW). 48 hours post-injection, we performed in vivo time-domain FLT imaging with co-registered CT, followed by in situ FLT imaging and western blotting to validate PD-L1 expression. Bi-exponential fitting of the time-resolved fluorescence data yielded spatial maps of FLT (τ) and amplitude ( a ). The top 10% of FLTs within the tumor were averaged to define the bound lifetime component (τ_b), while the mean FLT in the normal liver was considered as the FLT of the unbound component (τ_u). Amplitudes of bound (a_b) and unbound (a_u) probes were extracted and segmented into tumor (a_b,T; a_u,T) and normal (a_b,N; a_u,N) ROIs. Among four normalization strategies tested, one correcting for background cross-talk (a_b,T - a_u,T) and normalizing by total uptake (a_b,N - a_u,N) showed the strongest correlation with PD-L1 expression (R² = 0.77). 3D reconstructions using asymptotic time-domain analysis confirmed accurate tumor localization and size. FLT tomography distinguished bound from unbound alphaPDL1-800 and enabled quantitative assessment of PD-L1 expression in deep-seated tumors. The normalization method enabled accurate quantification of PD-L1 expression across tumors, supporting the utility of FLT imaging for preclinical immunotherapy evaluation and biomarker-guided therapy monitoring.