Understanding disease requires more than identifying which molecules are present—it requires knowing where they are located within tissues. Imaging mass spectrometry (MSI) enables us to map thousands of metabolites and lipids directly from tissue sections, generating detailed molecular “maps” without the need for labels or dyes. This powerful approach forms the basis of spatial omics, allowing molecular-level visualization of disease processes in their anatomical context. In our laboratory, we specialize in both DESI-MSI (Desorption Electrospray Ionization Imaging) and MALDI-MSI (Matrix-Assisted Laser Desorption Ionization Imaging) to investigate diverse disease models. DESI-MSI allows rapid, minimally destructive analysis of fresh or frozen tissues, making it highly suitable for translational and intraoperative applications. MALDI-MSI provides high spatial resolution and broad molecular coverage, enabling detailed molecular pathology and diagnostic classification.
A major focus of our interdisciplinary research, conducted in close collaboration with multiple medical centers and hospitals, is human cancer margin assessment—determining whether tumor tissue has been completely removed during surgery. Traditional pathology can be time-consuming and sometimes ambiguous. MSI, in contrast, provides rapid molecular information that distinguishes tumor from healthy tissue based on lipidomic and metabolomic signatures. We integrate advanced machine learning approaches with high-dimensional omics datasets to identify disease-specific lipid and metabolite biomarkers, pinpoint molecular features that define tumor margins, classify tissue types with high accuracy, enable data-driven intraoperative decision-making, by extracting diagnostic molecular patterns directly from MSI data, our approach aims to assist surgeons in real time, improving surgical precision and reducing recurrence rates.
our imaging mass spectrometry (MSI) research extends to non-cancerous diseases such as nephrotic syndrome, epilepsy, and other disorders where biopsy specimens are central to diagnosis. By applying MSI to kidney and brain tissues, we generate spatially resolved lipidomic and metabolomic maps that reveal subtle molecular alterations not visible through conventional histopathology. Integrating machine learning with high-dimensional MSI datasets enables us to identify disease-specific molecular signatures, stratify pathological subtypes, and uncover potential diagnostic biomarkers. This approach enhances the interpretive power of biopsy analysis, offering more objective, data-driven insights for clinical decision-making and advancing biomarker discovery in complex non-malignant diseases.
