Candidate retinal vascular metrics in patients with newly diagnosed cancer: a prospective pilot cohort study

Abstract

Purpose : To investigate the difference between retinal vascular measurements in a cohort of newly diagnosed cancer patients and matched cancer-free controls.

Methods : Participants newly diagnosed with cancer were recruited through the Cancer Centre of the Thunder Bay Regional Health Sciences Centre, Thunder Bay, Ontario, Canada. Age and sex matched cancer-free controls were recruited from the wider hospital community. All participants had digital retinal images taken using an OPTOS ultra-widefield scanning laser ophthalmoscope in an outpatient ambulatory eye clinic setting. Retinal images were analyzed using the VAMPIRE (Vascular Assessment and Measurement Platform for Images of the REtina) software application which characterises the retinal vasculature. Differences in the numeric values between retinal vascular measurements in the cancer and cancer-free cohorts were assessed using Analysis of Variance (ANOVA) and effect sizes calculated using the ETA-squared statistic with a P-Value of <0.05 deemed to be statistically significant.

Results : A total of 160 participants were recruited, including 59 controls and 101 cases for which readable digital retinal images were available for analysis. The mean age of controls and cases was 57.5 and 62.2 years of age, respectively (P <0.038) and 57 % of all participants were female. The main types of newly diagnosed cancer included: breast (n=31); hematological/leukemia (n=21); prostate (n=11); thyroid (n=10); lung (n=7); all others (n=21), respectively. Approximately 17% (40/234) of all VAMPIRE retinal measurements were found to be significantly different between cases and controls. A small to medium effect size for differences in fractal dimension values for retinal vascular measures in the peripheral retinal arteries; vessels and veins was observed with ETA-squared values of 0.012-0.014; 0.012-0.018 and 0.021-0.029, respectively.

Conclusions : Retinal vascular metrics, specifically fractal dimension values in the peripheral retina displayed a small to medium effect size between persons with and without cancer. Further research is required to explore the capture of retinal images and refinements in the use of appropriate retinal vascular metrics as a possible adjunct tool for further stratifying persons at an elevated risk for developing cancer, thereby potentially improving early detection and overall cancer survival.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.