EXPLORING THE MULTI-OMIC LANDSCAPE OF THYROID NODULES: A RETROSPECTIVE STUDY

Indeterminate lesion diagnosis in thyroid nodules remain an open problem for the risk of patient overtreatment [1]. In this context multi-omic analyses able to maintain the native spatial information could represent an unpreceded source of molecular information to assist clinicians in patients’ stratification. Here we analysed two archival Tissue Microarrays (TMA) including thyroid nodules from around 80 patients with different diagnosis. Sequential MALDI-MSI of N-glycans and tryptic peptides were acquired with a timsTOF flex operating in positive ion mode and at a spatial resolution of 20 µm/px. Following MSI acquisition, slides were H&E stained and whole-slide images were used to automatically define regions of interest (ROIs) through a pixel-based classifier, as described in Coelho et al [2]. N-glycans digestion significantly enhanced spectral quality by minimizing colloidderived interference, allowing more robust spatial molecular profiling. The tryptic peptides observed in the second molecular layer were more abundant in the stromal component of each core. Additionally, the automated ROI selection based on histological features ensured consistent and reproducible data extraction. The integration of N-glycans and tryptic peptides revealed distinct molecular signatures associated with different thyroid lesion types, highlighting alterations in glycosylation patterns alongside proteomic changes. Despite the need for further investigations, the proposed workflow opens new avenues to explore molecular heterogeneity in tissue neoplasms, enhancing the diagnostic potential of challenging thyroid lesions. [1] Nikiforov, Y. E. et al. JAMA oncology, 2.8 (2016), pp. 1023-1029; [2] Coelho, V. et al., Journal of Proteome Research, (2026).