Tools of the TradePublished: 11 November 2025Anthony A. Fung ORCID: orcid.org/0000-0003-1631-74511 &Lingyan Shi ORCID: orcid.org/0000-0003-1373-32061 Nature Reviews Nephrology (2025)Cite this articleSubjectsDiagnostic markersEnd-stage renal diseaseMetabolomicsSometimes it’s not what you look at, it’s what you see. In clinical pathology, the ability to identify pathological changes in a biopsy sample can have an immense impact on treatment decisions and patient outcomes. Different histological stains are frequently applied to serial sections of biopsy samples to provide a multiplexed perspective and thereby increase diagnostic power while avoiding issues relating to chromatic overlap due to simultaneous stains. However, the extra sample preparation associated with the use of serial sections requires laboratory and diagnostic expertise, irreversibly consumes precious biopsy volume, and risks physical deformation to the sample. Perhaps most importantly, serial sectioning entails post-processing co-registration — that is, the computational alignment of images from the different tissue sections — but can still miss critical microstructures and rare cells between slices. To address these limitations, we have developed a label-free optical imaging platform that combines stimulated Raman scattering (SRS), second harmonic generation (SHG) and multi-photon autofluorescence (MPAF) to capture the spatial morphology and molecular landscape of same-slide kidney tissue pathology.In the past few years, multimodal integration has led to the development of generative, cross-modality deep learning models that blend histological, spatial transcriptomic and immunofluorescence data. Training the models on spatially aligned datasets simplifies the process of ascribing spatial patterns to phenotypes and enables researchers to switch between modalities, but even adjacent serial tissue sections have minute spatial incongruities. By contrast, nondestructive imaging modalities can obtain rich, multimodal views of the exact same tissue specimen to enhance the diagnostic and research power currently achieved with co-registration approaches.This is a preview of subscription content, access via your institutionAccess optionsAccess Nature and 54 other Nature Portfolio journalsGet Nature+, our best-value online-access subscription27,99 € / 30 dayscancel any timeLearn moreSubscribe to this journalReceive 12 print issues and online access189,00 € per yearonly 15,75 € per issueLearn moreBuy this articlePurchase on SpringerLinkInstant access to full article PDFBuy nowPrices may be subject to local taxes which are calculated during checkoutFig. 1: Label-free indicators of diabetic nephropathies from single needle core biopsy samples.ReferencesOriginal articleFung, A. A. et al. Label-free multimodal optical biopsy reveals biomolecular and morphological features of diabetic kidney tissue in 2D and 3D. Nat. Commun. 16, 4509 (2025)Article CAS PubMed PubMed Central Google Scholar Download referencesAcknowledgementsThe original work was made possible by the funding, support and expertise of S. Jain and the Washington University Kidney Translational Research Center, as well as HuBMAP grants U54HL145608 (L.S., S.J.) and U54DK134301 (S.J.). We thank A. Knoten and K. Conlon for patient enrolment and sample preparation. We also thank the support of NIH R01GM149976 (L.S.) and a Sloan research fellow award (L.S.).Author informationAuthors and AffiliationsShu Chien – Gene Lay Department of Bioengineering, University of California San Diego, La Jolla, CA, USAAnthony A. Fung & Lingyan ShiAuthorsAnthony A. FungView author publicationsSearch author on:PubMed Google ScholarLingyan ShiView author publicationsSearch author on:PubMed Google ScholarCorresponding authorCorrespondence to Anthony A. Fung.Ethics declarationsCompeting interestsA.A.F and L.S. are co-founders and shareholders of Raman Noodle Inc.Rights and permissionsReprints and permissionsAbout this article