Every year, thousands of patients die on liver transplant waiting lists while potentially usable donor organs are discarded because clinicians lack reliable, whole-organ tools to assess viability. A noninvasive optical technology that could scan the entire liver surface — rather than sampling a tiny biopsy core — would fundamentally change how transplant teams make accept-or-discard decisions, potentially rescuing organs that are currently deemed too risky.
Polarization-sensitive optical coherence tomography (PS-OCT) uses light-scattering patterns sensitive to tissue microstructure and collagen architecture to quantify four critical hepatic parameters simultaneously: steatosis (fat accumulation), fibrosis, inflammation, and necrosis. In a study published in Science Translational Medicine, PS-OCT was applied across multiple surface regions of human donor livers, with machine learning and texture-analysis algorithms translating raw optical signals into pathological grades. The resulting quantifications correlated with conventional histopathology at greater than 80% accuracy. Critically, PS-OCT measurements also predicted functional performance during normothermic machine perfusion — the gold-standard ex-vivo stress test — and aligned with actual post-transplant clinical outcomes.
This work sits at an important intersection: optical coherence tomography has been refined for retinal and coronary imaging for over two decades, but its application to solid-organ transplant assessment is relatively nascent. The polarization-sensitive variant adds a layer of specificity for fibrillar collagen, which is precisely the architecture driving fibrosis scoring. The 80% correlation threshold, while promising, should be interpreted carefully — histopathology itself carries significant inter-observer variability, so this figure may partially reflect the noise floor of the reference standard rather than a ceiling for PS-OCT. The study appears to be single-center and does not yet report the size of the donor liver cohort in full granularity publicly available, which limits generalizability. Nonetheless, the combination of whole-surface mapping, machine learning interpretation, and correlation with functional and clinical endpoints makes this an unusually comprehensive preliminary validation. If confirmed in multicenter trials, PS-OCT could become a meaningful adjunct that reduces unnecessary organ discard from marginal donors.