A study published in Nature Methods has demonstrated the utility of iSCALE, an artificial intelligence platform developed at Vanderbilt University Medical Center, for the imaging of gene expression in large-sized tissues.
The advancement transforms how scientists and clinicians can visualize and analyze complex tissues on a high-resolution map of cellular landscapes across entire tissue sections, including regions that aren’t directly measured. In the proof-of-concept study, the researchers analyzed postmortem tissue samples from patients with two different diseases: gastric cancer and multiple sclerosis.
Their findings demonstrated “the utility of iSCALE in analyzing large tissues by enabling unbiased annotation, resolving cell type composition, mapping cellular microenvironments and revealing spatial features beyond the reach of standard spatial transcriptomic analysis or routine histopathological assessment,” the study stated.
The study has important clinical implications because existing spatial transcriptomic platforms are constrained by high costs, long turnaround times, low resolution, limited gene coverage and small tissue capture areas, the authors noted.
“This work fundamentally reimagines how we can scale spatial biology to clinically relevant samples. Instead of testing thousands of points across a side, we can now learn from just a few regions and generate a molecular map of the whole tissue. That changes the equation for translational science and diagnostics,” said Tae Hyun Hwang, PhD, professor of Surgical Research, founding director of Molecular AI Initiative and director of AI Research in the Division of Surgical Sciences at VUMC.
Notably, in the tissue sections from patients with multiple sclerosis, iSCALE revealed fine-grained immune features — such as activated microglia and lesion rim boundaries — that were not captured using conventional methods. The team confirmed these spatial patterns using immunostaining and pathology review, demonstrating the platform’s accuracy.
The research team is pursuing additional validation studies across cancer types, organ transplants and neuroinflammatory tissues, as well as exploring collaborations for clinical deployment of the iSCALE technology.
The iSCALE platform was developed under the leadership of Hwang. The platform reconstructs gene activity across full pathology slides with near single-cell resolution. iSCALE overcomes the constraints of traditional spatial transcriptomics technologies through a machine learning model trained on a small number of “daughter” molecular captures and matched hematoxylin and eosin images. The model infers gene expression as subcellular resolution (approximately 8 microns) across the entire tissue landscape, enabling a continuous molecular profile that spans full clinical slides.
Hwang received research support from National Institutes of Health grants (R01CA276690), (R37CA265967) and (U01CA294518) and Department of Defense grant (CA190578). His work was also partly supported by the Eric and Wendy Schmidt Fund for AI Research and Innovation through the Mayo Clinic Foundation and the American Association for Cancer Research Innovation and Discovery Grant. The MS tissue samples and associated clinical and neuropathological data were supplied by the Multiple Sclerosis Society Tissue Bank, funded by the Multiple Sclerosis Society.
Other Vanderbilt authors of the study are Jean Clemenceau, PhD, Inyeop Jang, PhD, Minji Kim, and Isabel Barnfather.
