This article is more than 5 years old. Information may no longer be current.
Neural network guides keratoconus treatment, predicts outcomes with intracorneal rings
Click Here to Manage Email Alerts
COPENHAGEN — A neural network based on a large database of keratoconus cases implanted with intracorneal ring segments can guide selection and intelligent surgical planning, predicting outcomes and optimizing results, according to a poster presented at the European Society of Cataract and Refractive Surgeons meeting.
“Using the real-life data of 2,000 patients, we have selected the most representative, successful 400 cases. Through computer analysis of preoperative and postoperative data, a software selects, based on analogy, the cases that are most similar to the case you are going to treat as far as age, sex and all relevant corneal parameters. In this way we can plan our treatment and predict visual acuity, refraction and quality of vision,” Jorge Alió, MD, PhD, first author of the poster, said. “In simple words, the computer will tell you what to do in order to treat that patient successfully.”
Jorge Alió
According to Alió, neural networks have great potential in medicine.
“It’s prediction based on a large body of experience. It’s like giving a 10-year-old the experience-based criteria of a 40-year-old,” he said.
The poster presented a study in which the simulation postoperative results based on a commercial nomogram were compared with the simulation based on the decision-making pathway proposed by the neural network. Significantly better keratometry values were achieved in the second group. – by Michela Cimberle
Reference:
Alió J. Neural network to guide keratoconus treatment with ICRS. Presented at: 34th Congress of the European Society of Cataract and Refractive Surgeons; Sept. 10-14, 2016; Copenhagen, Denmark.
Disclosure: Alió reports no relevant financial disclosures.