AI Technology to Identify Genetic Abnormalities of the Face

In addition to being able to detect disturbances on facial skin, AI technology can also identify genetic abnormalities from photographs of a person’s face.

Detect genetic abnormalities commonly known through ultrasound examination while the baby is still in the womb or by blood tests. Because, reported from various sources, genetic disorders – such as Down syndrome – can be detected from the age of 10 weeks of gestation. Now, quoted from CNN, there is already AI technology (artificial intelligence) or artificial intelligence that can accurately identify several types of rare genetic disorders using only photos of the patient’s face.

AI technology helps detect genetic disorders
A study noted, 8 percent of the world’s population has a disease with a major genetic component, and most have similar facial features. For this reason, Al’s technology developed DeepGestalt, a technology that can identify genetic disorders, such as Angelman’s syndrome.

Angelman Syndrome is a disorder that affects the nervous system with special characteristics, such as a wide mouth with sparse tooth spacing, squinting eyes and protruding tongue.

This AI (artificial intelligence) technology is the result of applying sophisticated algorithms that can identify something in a person’s body, even though there is not much patient data.

Meanwhile, researchers conducted tests on the DeepGestalt screening tool using 17,000 images of patients’ faces taken from a database of patients who have been diagnosed with more than 200 different genetic syndromes. Then, from 17,000 samples, 502 images were taken to be identified.

The result, the research team found, AI technology successfully outperformed doctors in two separate test sets to identify the syndrome contained in the hundreds of images selected earlier. In fact, in each test, AI managed to propose a list of syndromes that are potentially affected by each patient. Amazingly, the technology correctly identifies the patient’s syndrome within a certain time.

To be more convincing, another test was also carried out, namely by identifying different genetic subtypes in Noonan syndrome. The syndrome carries various characteristics and health problems such as heart defects.

Well here, the algorithm in AI technology has a 64 percent success rate for detecting genetic subtypes in Noonan syndrome. As for doctors who see images of patients with Noonan syndrome only able to identify 20 percent. The level of technological success that produced the heat map visualization on this face was also considered quite high.

AI technology still has shortcomings
With the discovery of this technology, researchers have opened the door to find more effective ways to identify genetic disorders in the future. However, even though it seems very sophisticated, it does not mean this technology has no shortcomings.

These deficiencies include the easy access to a person’s face image to increase discriminatory analysis of the patient’s health condition.

In addition, without a doctor’s diagnosis (and without a complete patient health database), this tool will only display many possibilities or a few lists of gene disorders that are quite confusing.

In the end, to determine which gene abnormalities are the most dominant and accurate, a doctor’s analysis is still needed.

However, this breakthrough in AI technology has the potential to change the lives of many people who have genetic disorders. This screening tool can help doctors or other medical personnel in detecting genetic abnormalities early. With this tool, the doctor can also determine the best therapy for the patient more accurately.