Age related macular degeneration (AMD) is the most common cause of blindness in the developed world. Treatment of selected patients can significantly improve their prognosis, but often requires intensive long-term monitoring. In this thesis we establish that the number of patients in need of treatment will increase by 50% in the coming decade, driven by the aging population. To effectively follow-up these patients in an already burdened public health care environment, novel strategies will be required. We therefore designed an artificial intelligence (AI) based system to autonomously follow AMD patients, training it on 105.000 examinations. The system demonstrated safety on par with regular follow-ups at our clinic during prospective validation. An estimated 60% of patients with the disease could be followed without human intervention, potentially easing the pressure on ophthalmology services in the coming years and helping maintain the current quality of care. Potential for deployment in the primary sector has the additional benefits of shorter travel and waiting times. In the future, research must be directed towards confirming efficacy of the AI system in a larger live validation setup and establishing solid ethical and legal frameworks for use of AI in medicine, which are currently lacking.

Ivan Potapenko​, cand.med.

4th February 2022, Department of Ophthalmology, Rigshospitalet, Glostrup

Wiley Online Library