Brain -inspired technology can accelerate diagnosis of rare diseases

Identify rare diseases In Brazil and worldwide it is still a challenge for doctors and a constant anguish for thousands of families. Many of these conditions are genetic, complex and little known, which makes the correct diagnosis difficult and can take years of uncertainty.

However, innovative technology, inspired by the functioning of the human brain, begins to change this scenario and promises to accelerate the diagnostic process: it is graph databases .

A long journey until the diagnosis

Rare diseases are considered a group of 6,000 to 8,000 conditions that affect up to 65 people in every 100,000 individuals.

More than 13 million people live with one of these conditions in Brazil, according to the Ministry of Human Rights and Citizenship. In the world, it is estimated that at least 300 million have some kind of such disease.

Despite the significant number, the path to a correct diagnosis is usually Slow, confusing and exhaustive.

It is currently done through a combination of clinical evaluation, complementary examinations (image and biochemicals) and genetic tests.

“Many of the rare syndromes have varied symptoms that overlap with the clinical picture of common diseases, making clinical diagnosis even more difficult,” explains Vanessa Montaleone, a geneticist of the Syrian-Lebanese Hospital Genetics Center.

In Brazil, the challenges become even greater by the complexities intrinsic to the country. Antoine Daher, president of Casa Hunter and the Brazilian Federation of Rare Disease Associations (Febrararas), highlights the main difficulties faced by families:

• Delay in diagnosis (on average, it takes 5 to 7 years until it reaches a correct diagnosis);
• misinformation and lack of adequate knowledge on the subject (including health professionals responsible for the case);
• High cost of genetic exams, specialized consultations and support therapies;
• Inequal access to health services for families that depend exclusively on SUS.

“Although there are reference services in rare diseases, created by GM/MS Ordinance No. 199/2014they focus on large urban centers and are insufficient to meet national demand. System navigation, from basic to specialized service, is also flawed, which prevents timely access to diagnosis, ”says Daher.

Graph technology and the case of the German Children’s Hospital

Developed almost 20 years ago by Neo4J startup, graph database technology aims to identify and store the relationship between data – not just the data itself.

“The intention is to present a visual connection of these relationships, which act as the links promoted by the brain synapses. Thus, it is possible to discover hidden patterns and get highly connected insights in real time,” says VP Latam Paulo Farias da Neo4J.

Unlike traditional banks, which depend on fixed tables or schemes, spelling seats use us and connections to represent, connect and provide context to information more intuitively and flexible. “Being even 1,000 times faster And consuming less cloud resources, ”says Farias.

In the context of rare diseases, technology becomes a solution that allows researchers and doctors to connect symptoms, genetic variants, scientific and historical publications in a dynamic and consultable network. This contributes to the formulation of more accurate diagnostics and in less time.

“Taking into consideration the more than 3 billion DNA present in the human body, analyzing and mapping the relationship of each one with diagnoses, exams and medicines would be humanly impossible. Not to mention more than 6,000 to 8,000 rare diseases cataloged by the World Health Organization (WHO),” says Farias.

The appeal has been used at Dr. Von Hauner Children’s Hospital in Germany. For this, the clinical knowledge graph (CKG) was created, which combines graph technology, artificial intelligence (AI) and machine learning (ML), containing data from 2,500 pediatric patients in the country (so far).

Each is represented by a node in the system, being connected to other knots (such as symptoms, proteins and phenotypes).

With this, doctors at the hospital can sail for a set of about 16 million of us and 220 million relationships.

According to Daniel Weiss, Bioit Chief of Children’s Hospital Dr. Von Hauner, before the adoption of graph technology, diagnosing rare diseases was a manual process: when there was no known genetic variant, doctors and scientists needed to analyze dozens of possibilities individually – something that could take months or even years.

With technology (still in the proof of concept), it is possible to automate this process by connecting clinical, genetic data and scientific literature in a system that facilitates analysis by artificial intelligence.

“This automation deals with the work that previously fell on a group of experts (…) we have not yet celebrated outstanding successes for individual children. However, we are convinced that it is the way to follow: to automate specialized work, integrate more rich data layers (RNA, proteomic) and allow external researchers to contribute. genuinely personalized, ”says Weiss.

And in Brazil, is it possible to apply?

Despite the enthusiasm, there is no evidence that this technology crosses the ocean and landing in Brazil soon. In Daher’s view, some obstacles can make it difficult to adopt immediate adoption.
Limited infrastructure in public services, low standardization of health systems and low training to deal with artificial intelligence are some of the current barriers.

Still, he believes it is possible to change this scenario. For this, it would be necessary to invest in innovation, train professionals and bring research centers and hospitals closer to SUS.

For now, modern genetic tests already represent an important advance, highlights Montaleone. DNA sequencing techniques have been increasingly efficient and faster.

“The implementation of sequencing by NGS (Next Generation Seicncing) enables us to sequence multiple DNA fragments simultaneously and, in conjunction with advances in bioinformatics, allows the analysis of large data volumes faster, more accurately and at a lower cost,” says the doctor.

Regarding treatment, gene therapies for various genetic conditions are used or studied, as well as innovative techniques such as CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats).

“They bring the possibility of more advances in the field of gene therapies, allowing hope for families and patients suffering from rare genetic syndromes.”

Genetic testing points to ancestry and risk of disease; understand

This content was originally published in brain -inspired technology can speed up diagnosis of rare diseases on CNN Brazil.

Source: CNN Brasil