Talk Summary:
The molecular identity of a cell and its position within the native three-dimensional architecture of a tissue together shape how disease develops. Yet, molecular assays such as immunofluorescence and spatial omics are read out on thin two-dimensional sections, and no existing approach recovers their molecular breadth within an unperturbed, intact 3D tissue environment at cellular-resolution. Here we present μSTiCH, a framework that decouples structural 3D imaging from molecular readout, bringing the multi-omic breadth of 2D assays into a micron-resolution native 3D tissue architecture. Using depth-encoded optical barcodes, μSTiCH aligns 2D hematoxylin and eosin (H&E), immunofluorescence, spatial proteomic and spatial transcriptomic sections at cellular resolution (~2 µm) to a label-free innate tissue scaffold imaged by high-definition optical coherence tomography (HD-OCT), a non-destructive microscopy. The scaffold then supports three uses. First, it explains the molecular basis of the HD-OCT signal by linking each signal feature to its underlying markers. Second, it directs targeted molecular analysis, locating regions of interest in the intact tissue and guiding assays precisely to them in place of blind sampling. Third, it maps molecular signatures back onto the native 3D volume for spatial visualization. In a murine glioblastoma model, μSTiCH pinpoints invasive, stem-like tumor cells and associated neural structures within the unperturbed 3D architecture of the intact tissue. These features extend beyond the visible tumor boundaries that 2D histology misses and that may seed local recurrence. In human glioblastoma, μSTiCH resolved distinct vascular niches and enabled the microregistration of 6,000-plex spatial transcriptomic profiles onto these niches in their corresponding 3D intact tissue volume. This spatial integration uncovered divergent molecular programs associated with these niches revealing intratumoral vascular heterogeneity. Beyond these three uses, and because μSTiCH links non-destructive HD-OCT signal to molecular state, these signatures could be read in vivo to follow molecular changes within a single tumor over time, across tissues and diseases.

Presenter Biography:
Yonatan Winetraub is an Instructor in the Department of Structural Biology at Stanford University School of Medicine and a member of the Wu Tsai Neurosciences Institute. His group builds optical imaging that reads tissue the way a pathologist reads a biopsy, without cutting it, combining optical coherence tomography, machine learning and spatial omics to resolve cancer at single-cell resolution. He has a PhD in biophysics from Stanford and received an NIH Director’s Early Independence Award.

Author
Yonatan Winetraub
Stanford University