medwireNews: A study published in Heart suggests that detailed analysis of retinal blood vessels may significantly improve stroke risk prediction.
The researchers found that analysis of 29 retinal indicators using routine fundus photography performed better than traditional risk stratification models alone, "indicating its potential application as a non-invasive screening method for individuals with increased risk."
They suggest: “Further investigation of these parameters may provide valuable insights into the intricate pathophysiological processes associated with stroke, thereby contributing to the refinement of preventive and therapeutic strategies.”
Mingguang He, from The Hong Kong Polytechnic University in China, and colleagues analyzed fundus images along with demographic and healthcare data for 45,161 individuals (54.9% women), aged a mean of 55.4 years, from the UK Biobank. Of these, 749 experienced incident strokes over a median follow-up of 12.5 years.
The researchers used the Retina-based Microvascular Health Assessment System (RMHAS), a deep-learning algorithm, to perform automated segmentation and quantification of retinal vascular networks, arteries, and veins within and outside the macular region from the participants' fundus images, identifying a total of 118 retinal vascular parameters.
In all, 29 of these parameters were significantly associated with incident stroke on Cox regression analysis, after adjusting for traditional risk factors (including age, sex, social deprivation, blood pressure, total cholesterol, high-density lipoprotein, glycated hemoglobin, smoking status, and BMI).
Density-related parameters dominated the findings, with 17 of the 29 indicators falling into this category. Six of the parameters associated with arc length (arteries, vessels in the macular region, and arterial terminal and nonterminal points) and chord length (arteries and vessels in the macular region) significantly increased the risk for stroke by 10–14% with each standard deviation (SD) increase. Whereas each SD decrease in the remaining 11 parameters significantly increased the risk by 9.9–19.0%. These included bifurcation density in the arteries; branching density in the arteries, veins, and vessels in the macular region; vessel area density in the vessels and arteries within and outside of the macular region; and skeleton density in the arteries within and outside of the macular region.
The other retinal parameters associated with stroke risk related to three groupings, namely caliber, complexity, and tortuosity. With regards to caliber, there were two measures associated with stroke risk, namely central retinal artery equivalent, which increased the risk by 10.4% with each SD decrease, and the length to diameter ratio of the arteries and vessels in the macular region, which increased the risk by 10.1–14.1% with each SD increase.
Reduced vascular network complexity and irregularity also indicated an increased risk for stroke, with a total of eight parameters identified. Specifically, each SD decrease in artery and vessel fractal dimension was associated with a 12.4–16.0% increased stroke risk. Similarly, each SD reduction in the number of arterial bifurcation points, branching points, nonterminal points, and segments, and the number of vessels in the macular region increased the risk for stroke by 10.4–17.5%.
Among tortuosity-related parameters, only arterial inflection count was significantly associated with stroke risk, with each SD decrease raising the stroke risk by 10.1%.
The study also demonstrated that incorporating retinal parameters into models assessing the traditional risk factors for stroke, the measurement of which can be expensive or may require invasive blood tests, led to a statistically significant increase in the area under the receiver operating characteristic curve (AUC) from 0.738 to 0.752. The authors highlight that using only age, sex, and retinal parameters achieved a comparable AUC to that using traditional risk factors alone of 0.739, suggesting "a practical and easily implementable approach for incident stroke risk assessment, particularly for primary healthcare and low-resource settings.”
They add: “This approach could inform future policy regarding stroke prevention strategies potentially leading to earlier intervention and improved patient outcomes.”
The authors acknowledge, however, that the predominantly White ethnicity (91.2%) of the UK Biobank cohort limits the generalizability of their findings, while its observational nature prevents confirmation of a causal link between retinal vascular parameters and stroke risk.
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2025 Springer Healthcare Ltd, part of the Springer Nature Group