Senescent cell detection could become faster and less destructive through a new technique developed by researchers at MIT, Massachusetts General Hospital and collaborating institutions. The team combined Raman microscopy, spatial gene-expression measurements and machine learning to generate biochemical “barcodes” capable of identifying senescent cells in intact tissue. The approach, called RamanOmics, was demonstrated in mouse skin and lung tissue and could eventually provide researchers with new ways to study aging without destroying the cells they want to examine.
The research was published September 21, 2026, in Nature Aging and forms part of the NIH Cellular Senescence Network, or SenNet, initiative.
Senescent Cell Detection Targets Zombie Cells
Senescent cells are sometimes described as “zombie cells” because they stop dividing but remain metabolically active instead of dying.
Cellular senescence can occur after DNA damage or other forms of cellular stress. The process is not inherently harmful. Senescence helps prevent damaged cells from multiplying and also participates in normal biological processes such as embryonic development, tissue repair and wound healing.
Problems can develop when these cells accumulate as people age.
The immune system normally helps remove senescent cells, but that clearance becomes less efficient over time. Accumulating senescent cells can release inflammatory signals and have been associated with tissue dysfunction and multiple age-related conditions.
Senescent Cell Detection Has Been Difficult
Scientists already use biomarkers such as the proteins p16 and p21 to help identify senescence.
However, conventional techniques often require cells to be fixed, stained, sequenced or otherwise processed in ways that destroy them.
That makes it difficult to observe the same living tissue repeatedly or study what happens to identified senescent cells over time.
MIT researchers wanted a nondestructive alternative.
Their approach uses Raman microscopy, which shines visible or near-infrared light onto tissue and measures how that light scatters. The resulting spectrum reveals information about chemical bonds and can provide a biochemical fingerprint of individual cells without requiring labels or destroying the tissue.
RamanOmics Combines Molecular Data With Imaging
Raman imaging alone can provide rich biochemical information, but identifying senescence requires understanding which spectral changes actually correspond with the cellular state.
The researchers therefore combined Raman microscopy with spatial RNA sequencing and single-cell molecular information.
Their RamanOmics platform integrates label-free hyperspectral Raman imaging, single-nucleus RNA sequencing, STARmap in situ sequencing and machine-learning analysis.
This allowed the team to compare biochemical signatures with gene-expression programs occurring in the same tissues.
Senescent Cell Detection Gains AI Barcode
Machine learning was then used to identify combinations of Raman signals that most reliably distinguished senescent cells.
The researchers combined the most informative Raman features with molecular gene signatures to construct a multimodal barcode for cellular senescence.
One particularly notable finding was a conserved lipid-linked Raman signature at approximately 1,131 to 1,135 inverse centimeters that marked p21-positive senescent cells across tissues.
Instead of analyzing an entire Raman spectrum every time, future systems could potentially focus on a smaller collection of highly informative bands.
That could make the technique considerably faster.
Aging Cells Show Major Lipid Changes
The researchers studied skin and lung samples from young 2-month-old mice and aged 26-month-old mice.
One of the clearest changes across both tissues involved lipid metabolism.
Older and senescent cells showed increased lipid synthesis and accumulation, suggesting that cellular aging is accompanied by substantial biochemical remodeling.
The exact biological consequences of those lipid changes are not yet fully understood.
Researchers also observed tissue-specific differences.
In aged skin, senescence was associated with changes in pathways involving muscle contraction, collagen and extracellular-matrix remodeling. In lung tissue, researchers detected stronger activation of immune and inflammatory gene programs.
Wound Healing Reveals Dynamic Senescence
The team also evaluated senescence during wound repair.
Their analysis found that senescent cells involved in healing showed changes involving epidermal differentiation genes as well as increased lipid-associated Raman signatures.
That finding reinforces an important point: not every senescent cell should necessarily be eliminated.
Senescence can contribute to beneficial repair processes as well as harmful age-related inflammation.
NIH’s SenNet program specifically aims to understand these differences so future therapies might suppress harmful senescence while preserving its beneficial biological functions.
Faster Imaging Is Needed for Clinical Use
RamanOmics remains a research technology rather than a clinical diagnostic.
The MIT team currently requires about 30 hours to image a tissue sample measuring roughly one square millimeter. Researchers are developing a faster system capable of concentrating on the Raman barcode bands found to be most informative.
A faster approach could eventually allow larger tissue areas to be screened.
Researchers have even suggested that future Raman-based systems might be incorporated into tools such as endoscopes to identify cellular senescence inside the body.
That remains a future possibility rather than an established medical application.
Human Tissue Is the Critical Next Step
Another major limitation is that the current study was performed in mouse tissues.
Researchers are now working to adapt the technique for human samples.
Human validation will be essential because senescent cells are highly heterogeneous. Their characteristics can differ depending on tissue type, age, disease state and the biological event that triggered senescence.
Recent NIH SenNet research underscores that complexity. The consortium has begun building comprehensive human atlases that classify different senescent-cell states, or “senotypes,” across tissues and stages of life.
Senescent Cell Detection Could Transform Aging Research
The immediate significance of RamanOmics is not a new anti-aging treatment.
Its value lies in giving researchers a potentially nondestructive way to identify and characterize cells that are otherwise difficult to recognize reliably.
Better senescent cell detection could help scientists understand which cells contribute to inflammation and disease, which support normal tissue repair and how experimental senolytic therapies affect specific cellular populations.
If the Raman barcode can eventually be validated in human tissues and imaging speeds improve substantially, the technology could also support future diagnostics for age-related disorders.
For now, the study provides an important research platform linking biochemical imaging, spatial genomics and machine learning—and offers a clearer view of the “zombie cells” that accumulate as tissues age.
For more healthcare industry updates, insights and news, visit DistilINFO. Click here to subscribe to stay informed
