Eye exams could enable earlier detection of systemic disease

There is increasing recognition that the eye is a good source of information about systemic diseases. Two recent journal articles underscore the organ’s importance in identifying systemic disease processes. This is clinically relevant in that the eye provides an unobstructed and noninvasive view of the body’s vascular and neurologic systems and gives eye care clinicians the opportunity for early detection of diseases.

National study of retinopathy of prematurity (ROP) screening

In this diagnostic study,¹ the investigators used artificial intelligence (AI) to analyze images obtained during ROP screening. identified bronchopulmonary dysplasia and pulmonary hypertension in infants, according to first author Praveer Singh, PhD, from the Department of Ophthalmology, University of Colorado School of Medicine, Aurora.

Singh and colleagues emphasized that bronchopulmonary dysplasia is a leading cause of morbidity in premature infants, affecting one-third of very preterm infants, or approximately 18,000 infants annually in the US.² Pulmonary hypertension, a condition separate from bronchopulmonary dysplasia but related, causes increased pulmonary vascular pressure.^3-5 Both diseases can be challenging to diagnose and are causes of morbidity and mortality in premature infants, according to the investigators.

They wanted to determine if using the ROP images obtained during screening would show features of the two diseases and if a multimodal model integrating imaging features with demographic risk factors could outperform a model based on demographic risk alone, they explained.

A deep learning model was used to study the images obtained from the Imaging and Informatics in Retinopathy of Prematurity study.⁶ The infants included were at high risk of developing ROP and were undergoing routine screening. All of the images were obtained when the infants were 34 weeks or less postmenstrual age to precede the clinical diagnosis of bronchopulmonary dysplasia or pulmonary hypertension.

What did the AI model show?

The study included 493 infants (mean ± standard deviation, gestational age, BPD, 25.7 ± 1.8 weeks; normal, 27.3 ± 1.8 weeks).

“For bronchopulmonary dysplasia, the multimodal model showed higher accuracy (area under the curve [AUC], 0.82; 95% confidence interval [CI], 0.72-0.90) than demographics only (0.72; ∆AUC, 0.1; 95% CI, −0.008 to 0.21; P = 0.07) or imaging only (0.72; ∆AUC, 0.1; 95% CI, 0.04-0.16; P = 0 .002) models. For pulmonary hypertension, the multimodal AUC was 0.91 vs demographics only 0.68 (∆AUC, 0.14; 95% CI, 0.006-0.27; P = 0 .04) and imaging only 0.91 (∆AUC, −0.09; 95% CI, −0.3 to 0.12; P = 0.40) models. The results persisted when trained on images lacking clinical ROP signs.”

Singh and colleagues concluded, “The results suggested that retinal images obtained during ROP screening may be used to predict the diagnosis of bronchopulmonary dysplasia and pulmonary hypertension in preterm infants, which may lead to earlier diagnosis and avoid the need for invasive diagnostic testing in the future.”

Along those same lines…

A large Canadian study⁷ evaluated the short- and long-term risks of cardiovascular disease associated with a first diagnosis of transient visual loss. The investigators, led by Tasha Miller, MD, found that major adverse cardiovascular events, stroke, myocardial infarction, arrhythmia, and hospitalization are indeed associated with transient visual loss. She is from the Temerty Faculty of Medicine, University of Toronto.

To reach that conclusion, they retrospectively identified patients with an incident diagnosis of transient visual loss and matched the patients 1:1 to control patients with dry eye. The primary outcomes included major adverse cardiovascular events, stroke, myocardial infarction, ventricular arrhythmias, venous thromboembolism, hospitalization, and all-cause mortality, they explained.

Did the study underscore the importance of transient vision loss?

A total of 37,750 patients were included in both the patient and control groups. The mean age was 56.8 years in the group with transient visual loss and 56.6 years in the control group.

Miller and colleagues reported, “Within 14 days, the stroke risk increased over 21-fold (hazard ratio [HR], 21.7; 95% confidence interval [CI], 13.4 to 37.4), the major adverse cardiovascular events nearly 10-fold (HR, 9.80; 95% CI, 7.19 to 13.34), arrhythmia over 4-fold (HR, 4.01; 95% CI, 2.72 to 5.90), myocardial infarction 5-fold (HR, 5.00; 95% CI, 1.92 to 12.06), and hospitalization nearly 4-fold (HR, 3.83; 95% CI, 3.52 to 4.17) compared with controls.”

The risk of venous thromboembolism was modest and transient, they added.

For the patients who were event-free at 90 days or 1 year, the investigators identified an elevated long-term risk that persisted up to 10 years for a major adverse cardiovascular event, stroke, arrhythmia, and hospitalization.

The authors concluded, “TVL is associated with increased short- and long-term risks of a major adverse cardiovascular event, stroke, myocardial infarction, arrhythmia, and hospitalization, which warrant prompt systemic evaluation and long-term monitoring.”

References:

1. Singh P, Kumar S, Tyagi R, et al. Deep learning–based prediction of cardiopulmonary disease in retinal images of premature infants. JAMA Ophthalmol. Published online January 22, 2026. doi:10.1001/jamaophthalmol.2025.5814

2. Thébaud B, Goss KN, Laughon M, et al. Bronchopulmonary dysplasia. Nat Rev Dis Primers. 2019;5:78. doi:10.1038/s41572-019-0127-7

3. 3.Clark RH, Kueser TJ, Walker MW, et al. Clinical Inhaled Nitric Oxide Research Group. Low-dose nitric oxide therapy for persistent pulmonary hypertension of the newborn. N Engl J Med. 2000;342:469-74. doi:10.1056/NEJM200002173420704

4. 4.Kinsella JP, Abman SH. Recent developments in inhaled nitric oxide therapy of the newborn. Curr Opin Pediatr. 1999;11:121-125. doi:10.1097/00008480-199904000-00004

5. 5.The Neonatal Inhaled Nitric Oxide Study Group. Inhaled nitric oxide in term and near-term infants: neurodevelopmental follow-up of the neonatal inhaled nitric oxide study group (NINOS). J Pediatr. 2000;136:611-617. doi:10.1067/mpd.2000.104826

6. Brown JM, Campbell JP, Beers A, et al. Imaging and Informatics in Retinopathy of Prematurity (i-ROP) Research Consortium. Automated diagnosis of plus disease in retinopathy of prematurity using deep convolutional neural networks. JAMA Ophthalmol. 2018;136:803-10. doi:10.1001/jamaophthalmol.2018.1934

7. Miller T, Xie JS, Qureshi AR, Tao BK, Margolin E. Cardiovascular risk following transient vision loss. Br J Ophthalmol. 2026; published online first. https://bjo.bmj.com/content/early/2025/12/09/bjo-2025-328605