A recent literature review explored the diagnostic accuracy and feasibility of optical coherence tomography and optic disc photography in glaucoma screening, as well as the role artificial intelligence may have in enhancing effectiveness. The study also examined the cost-effectiveness of each method, the potential role of artificial intelligence in establishing sufficiently accurate and cost-effective screening tools, and the challenges associated with implementing these technologies in widespread screening programs.
Optical coherence tomography (OCT) provides high reproducibility and accuracy in quantifying structural damage in glaucoma, making it the standard method for diagnosing the disease. However, its high cost and lack of portability limit its widespread use in community-based screening, investigators noted in the Journal of Glaucoma.
Optic disc photography is more accessible and cost-effective, especially with advancements in portable and non-mydriatic fundus cameras. However, the subjective interpretation of images in this method can lead to variability in diagnostic accuracy. The integration of deep learning artificial intelligence (AI) models shows promise in improving this method’s accuracy by automating the detection of glaucomatous optic neuropathy. Still though, the effectiveness of these models depends heavily on the quality of training data. Current models may replicate the limitations of human graders if they are trained on subjective assessments.
Positive and negative predictive values based on specific clinical settings may also affect the screening model chosen. “These values are crucial as they directly reflect the effectiveness of the screening tool in accurately identifying disease presence or absence,” the investigators wrote. False negatives or false positives can lead to problems, including incorrect allocation of confirmatory testing and treatment and unnecessary stress on patients. “When deciding about the implementation of OCT in a screening effort … it is essential to select cutoff parameters that will target patients with clearly established disease.”
The investigators concluded that a combined approach leveraging the strengths of both methods could optimize glaucoma screening programs.
A full list of author disclosures can be found in the published research.
