Professor Kim Sang-jin, an ophthalmologist at Samsung Medical Center (SMC), has developed and verified the efficacy of severity index for retinopathy of prematurity (ROP) through joint research with Harvard University and Casey Eye Institute.

The international research team has improved the existing ROP severity scoring system, which was too simple, and classified the progressive changes into nine stages based on the plus disease that greatly affects prognosis.

Plus disease describes the abnormal changes in retinal blood vessels associated with ROP, SMC said in a news release. 

Patients with ROP, if untreated, can result in retinal detachment or bleeding due to abnormal development of retinal blood vessels, usually resulting in irreversible vision loss. ROP occurs in more than 50 percent of infants born before 30 weeks of pregnancy, and the rate soars to 90 percent before 24 weeks.

However, the existing three-stage classification had limits as it could not distinguish degeneration or improvement of the disease clearly, and doctors often have different diagnostic standards for the plus disease.

The research team divided ROP patients into 1 to 3 stages in the case of having no plus disease for vascular abnormalities, 4 to 6 when they are about to show the sign, and 7 to 9 if their plus disease is observable.

The artificial intelligence model, developed by the research team, learned about the clinical data of about 6,000 ROP patients and helped their diagnosis based on the results of its screening of the patients’ retinal posterior poles. 

As a result, the research team confirmed that its AI model showed consistent assessment to that of specialists, proving its effectiveness.

“If we use AI tools to diagnose diseases using a standardized severity index, we can identify diseases more accurately and quickly,” Professor Kim said.”

The study was published in Ophthalmology, the official journal of the American Academy of Ophthalmology.