Label-free melanoma boundary identification: a multiphoton-imaging-based approach

Precise intraoperative delineation of tumor boundaries is critical for achieving negative margins in melanoma surgery. However, conventional methods are limited by subjectivity and time constraints. In this study, we employed label-free multiphoton microscopy (MPM) to image collagen fibers in normal subcutaneous, tumor boundary, and melanoma tissues in female nude mice. Eight quantitative morphological features of collagen architecture were extracted. Statistical analysis revealed significant differences among the three tissue types (P < 0.0001). Tumor tissues exhibited distinct collagen remodeling, characterized by increased fiber alignment (elevated Feature 5, reduced Feature 8) and structural degradation (reduced Feature 6, elevated Feature 7). Pairwise comparisons showed that Features 5 and 6 did not differ significantly between normal and boundary tissues (P > 0.05), while all other features distinguished the groups with high significance (P < 0.05), except Feature 4 between boundary and tumor tissues (P = 0.39). These findings demonstrate that MPM-derived collagen metrics can serve as robust, real-time biomarkers for intraoperative tumor boundary identification. This approach offers a rapid, objective, and high-resolution method to guide surgical resection, potentially reducing recurrence and improving patient outcomes.