Rosa’s life had become an ordeal; no glasses worked for her. “But that was not the worst part. I couldn’t do anything in the house, if I was eating and went to pour myself a glass of water I would spill it on the table. Also, I did not go out alone for fear of having an accident. I was always glued to my son. It is not that I was blind; I saw shapes, shadows, but I could not distinguish faces. To know who I was talking to I had to listen to the person’s voice,” she told me.
Rosa was diagnosed with cataracts and underwent surgery at the Holguín Ophthalmological Center. After the operation, she immediately began distinguishing faces she had not seen for years.
However, not all patients are so lucky. Even when the surgery occurs without incident, an error in the lens calculation that is placed in the patient’s eye can lead to the need for a new operation. Therefore, the use of artificial intelligence in these procedures is seen as a possible solution.
A group of Cuban scientists has already applied it successfully. The work: “Inteligencia artificial y modelación numérica orientada al mejoramiento de la visión en la población cubana” (Artificial intelligence and numerical modeling aimed at improving vision in the Cuban population), by the authors Dr. Alfo José Batista Leyva and Dr. Iván Hernández López,1 was recognized in 2024 with the Prize of the Cuban Academy of Sciences.
How do we see?
The act of seeing depends on the perfect functioning and normal interaction between the eye, the structures that surround it, the optic nerve and the brain.
According to the American Academy of Ophthalmology (AAO), to direct our gaze in a certain direction, the eye has six muscles that act in a coordinated manner. However, this is not enough for us to see; we need the anterior portion of the eye to focus on the target, just as the lens of a camera would do. A structure called the crystalline lens plays a key role in this.
The light passes to the retina, where it becomes an electrical impulse and this, in turn, is transmitted through the millions of nerve fibers of the optic nerve to the brain.
What is cataract?
Cataract is the opacity of the crystalline lens due to the fact that its proteins begin to break down with age. It can be of the senile, presenile, traumatic, and metabolic type, as explained to OnCuba by Dr. Tareq Ziad Mustafa Toame, an ophthalmology specialist of Palestinian origin who works at the Ophthalmology Center of Holguín and who was in charge of Rosa’s surgery, mentioned at the beginning of this text.
According to Dr. Tania Milagros Fernández Domínguez, a Second-Degree Specialist in Ophthalmology who works at the same hospital, this prevents the correct entry of light and the quality of the formation of the image on the retina.
Refractive surgery is responsible for solving myopia, hyperopia and astigmatism, which are not normally considered diseases, but refractive defects. Cataract surgery is also considered a type of refractive surgery, according to Dr. Ziad Mustafa.
In this type of intervention, the patient’s opaque lens is replaced by a completely transparent one called an intraocular lens (IOL). There are different types of IOL. Factors such as the length of the eye and the lens, the curvature of the cornea, among others, are taken into account for their selection. Once this information is available, different formulas are applied depending on the size of the eye.
To obtain all these measurements, ultrasonic waves are used to suggest the type of lens to choose. In our field, we have a device called a diometer and the highly specialized work of our technicians, who are in charge of making the calculations depending on the characteristics of the eye and the objectives of the surgery, according to Dr. Fernández, who is also an Assistant Professor at the University of Medical Sciences of Holguín.
The exact calculation of the IOL is very important for the eye to reach a situation called emmetropia. This is the objective of cataract surgery, that is, for the eye to achieve normal vision without the use of lenses, in the absence of any other cause that prevents this from happening. The “refractive surprise,” on the other hand, is when the eye emerges from the intervention myopic or hyperopic, explains Dr. Tareq.
Artificial intelligence for lens calculation in Cuba
According to the article: “Impacto de la Inteligencia Artificial en la cirugía de cataratas desde una perspectiva social” Impact of Artificial Intelligence in cataract surgery from a social perspective), published in 2021 in the Cuban journal Revista Cubana de Oftalmología, the introduction of new technologies in Cuban medical sciences is a necessity.
Various institutions, including the Cuban Ramón Pando Ferrer Institute of Ophthalmology (ICO), have embraced this philosophy.
A team from that institution, led by Dr. Iván Hernández, worked together with a group of physicists, mathematicians, and computer scientists from the Higher Institute of Technology and Applied Sciences of the University of Havana (INSTEC) to develop an artificial intelligence tool that would make the calculation of intraocular lenses (IOL) more accurate. Its objective was to avoid the occurrence of refractive errors in the procedures due to inaccuracies in this calculation.
To better understand the work of these specialists, it is necessary to explain some concepts related to AI. To do so, OnCuba had the opportunity to talk about this with Computer Science Engineer Yandi Córdova Cobas, who currently works in Spain.
According to him, artificial intelligence is dedicated to developing systems and programs capable of imitating certain human capacities such as learning and reasoning. We can imagine AI as a “fast learner” to whom we provide data and tools and it becomes a master. Machine learning, on the other hand, is a process that allows computers to learn autonomously. Unlike traditional programs, where programmers must write specific instructions for each task, in machine learning it is not necessary to indicate exactly what to do. Instead, the computer is provided with a large amount of inputs and a model, so that it can identify patterns, draw conclusions, or make decisions.
An artificial neural network is a computer system that attempts to replicate the functioning of the human brain to solve various problems. It is made up of small interconnected units, known as “neurons,” that collaborate to learn from the input.
For Córdova, among the advantages of this technology are: the automation of repetitive tasks; greater speed and efficiency in their execution; algorithmic personalization (that is, they allow interfaces to make personalized recommendations, for example on platforms such as Netflix); improvement in decision-making based on data, which makes them faster and more accurate; continuous learning; innovation in different areas; cost reduction; access to new knowledge, among many others.
The work of the multidisciplinary team from ICO and INSTEC consisted of creating several artificial neural networks and training them to be able to predict the power of intraocular lenses and suggest the most suitable IOL for each patient. Their objective, as I mentioned, was to reduce refractive surprises as much as possible. When these occur, the patient must use lenses or undergo a new intervention.
To train the different neural networks, the researchers used data from more than 15,000 patients who had previously undergone surgery at Pando Ferrer. To calculate the IOL, countless formulas have been developed. These serve as a basis for choosing the lenses in each operation. Unlike how it is done conventionally, the artificial neural networks used in the experiment were not given any formula other than those normally used.
Using machine learning, patterns were identified and the power of the IOLs was predicted with an effectiveness never seen before. According to an article published in a specialized journal, this could reduce refractive surprises to less than 5%.
A step in the right direction
The first works in this field of intersection between artificial intelligence and cataract surgery are nearly 30 years old. It is encouraging that the joint work of a multidisciplinary team of Cuban researchers has obtained such relevant results in this medical specialty. However, in Cuba, these techniques are only being applied in the capital so far, and there are no new updates on their results in the specialized literature.
According to the authors of the article we are commenting on, refractive errors affect almost one in five patients operated on for cataracts in the world, and with the use of AI, this proportion decreases significantly. At the time of writing this text, there is no recent data on this indicator in our country.
The use of artificial intelligence is not a panacea in the field of medical sciences. These technologies can be expensive. In addition, they also include risks such as biases in algorithms from human biases that are amplified and perpetuated; lack of transparency in understanding the decision-making mechanisms of AI; job displacement (in this case it would make the work of the technical staff who perform the IOL calculation unnecessary); unethical use of technology, among others.
Artificial Intelligence is here to stay, also in medicine. Like all scientific advances, its use raises fears and enthusiasm. Assuming it with cautious optimism and getting the most out of it should be the goal.
________________________________________
Note:
1 The author tried to contact Dr. Iván Hernandez Silva but did not receive a response.
Author
Carlos Alberto González
