As many patients within the rare disease community know, the diagnostic odyssey can be long and arduous. But what if burgeoning technologies, like artificial intelligence (AI) and its branches (machine learning), could help to improve this process? According to Medical XPress, researchers from the University of Wisconsin – Madison’s Waisman Center are working to do just that. Through an evaluation of 1.7m patients and their e-health records, researchers discovered that machine learning algorithms could more easily help identify and diagnose patients with Fragile X syndrome (FXS). See the full study findings published in Genetics in Medicine.

Machine Learning

So what is machine learning? According to IBM:

Machine learning is a branch of artificial intelligence (AI) focused on building applications that learn from data and improve their accuracy over time without being programmed to do so. In data science, an algorithm is a sequence of statistical processing steps [which,] in machine learning, [are] ‘trained’ to find patterns and features in massive amounts of data in order to make decisions and predictions based on new data.

In particular, machine learning focuses on computers and computer programs which can not only access data but use that data to learn. Machine learning does not depend on human intervention; rather, the machine uses algorithms to develop a modeling system which can analyze and evaluate data.

This technology is incredibly helpful to researchers. In part, this is due to the fact that machine learning can more efficiently and quickly evaluate and analyze a ton of data. Thus, rather than researchers exploring all 1.7m records, the machine performs the analysis.

The Research

In this case, researchers sourced the 1.7m records, which spanned over a period of 40 years. The records focused solely on Wisconsin residents. Altogether, researchers discovered that:

• Patients with FXS are more likely than others to be diagnosed with digestive, respiratory, genital, urinary, circulatory, and metabolic conditions.
• It can take up to 2 years following initial symptoms for patients to be able to receive genetic testing.
• Machine learning algorithms and models were able to predict FXS diagnoses up to 5 years earlier than patients who received clinical diagnoses.
• Because of the amount of cardiac-related complications and comorbidities, patients with FXS should be frequently screened for circulatory diseases.
  • Note: In fact, the findings discovered that patients with FXS were 5x more likely to develop heart valve disorders than others.

Ultimately, researchers believe that earlier FXS diagnosis will allow families to receive better care for their children. Although there is no cure for FXS, early diagnosis can lead to better therapeutic interventions, opportunities for family planning, and genetic counseling. In the future, researchers would also like to hone their machine learning platform through the inclusion of other medical records from outside of Wisconsin.

Fragile X Syndrome (FXS)

FMR1 gene mutations cause Fragile X syndrome (FXS), a genetic disorder that causes learning and developmental delays. Normally, the FMR1 gene creates fragile X mental retardation 1 (FMR1) protein. This protein plays a role in normal nervous system function. As a result, gene mutations inhibit this function, causing cognitive impairment and other symptoms. Although patients with FXS have normal life spans, many are unable to live independently. FXS typically impacts males more severely than females. At first, children with FXS miss developmental milestones. However, later symptoms or characteristics include:

• Developmental and intellectual delays
• An elongated face with protruding ears, forehead, and chin
• Flat feet
• Stuttering
• Anxiety, depression, or hyperactivity
• Hand-biting or flapping
• Aggression
• Poor eye contact
• Hyperflexibility
• Frequent ear infections
• Difficulty with speech and language
• Seizures
  • Note: Seizures only occur in an estimated 5-15% of people with FXS.