Abstract
Abstract
Fluorescence lifetime imaging microscopy (FLIM) provides quantitative contrast based on fluorescence decay dynamics, but its high-resolution implementation is generally restricted to a small field of view, limiting the tissue-scale analysis of spatially heterogeneous pathological sections. Here, we present FLIStitcher, an automated, high-fidelity method for reconstructing large-field FLIM mosaics from sequentially acquired small field tiles. The method combines shading correction, brightness correction, feature-based tile alignment, phase-correlation-based strip alignment, and seam-guided fusion to suppress illumination nonuniformity, tile-to-tile brightness mismatch, and boundary artifacts while maintaining consistent lifetime-based pseudocolor contrast. Reference-based validation using FLIM images from different tissue types demonstrates accurate reconstruction with a high structural similarity and low pixel-level deviation. In large-field tissue data sets, FLIStitcher preserves local structural continuity and shows improved robustness compared with existing stitching approaches, particularly for irregular tissue boundaries, repetitive structures, and weak-texture regions. Its application in autofluorescence FLIM measurements of ovarian cancer enables regional lifetime analysis of red blood cell-enriched, nontumor stromal, and tumor-containing tissue regions. FLIStitcher provides a practical computational strategy for extending high-resolution FLIM from local measurements to large-area tissue analyses.