Molecular fingerprint shows potential to spot Fabry in either sex
Initial diagnosis often relies on enzyme tests that can be inaccurate
A machine learning algorithm to find patterns in data from infrared spectroscopy — a technique that can provide a molecular fingerprint by looking at how molecules in a sample take up light and reflect it — correctly distinguished people with Fabry disease from healthy adults in a pilot study, regardless of sex.
“It is a fast, inexpensive, and minimally invasive technique applicable to both sexes,” the researchers wrote in “Infrared spectroscopy as a new approach for early Fabry disease screening: a pilot study,” which was published in the Orphanet Journal of Rare Diseases.
Fabry disease is caused by genetic mutations, leading to a deficiency in the production of the enzyme alpha-galactosidase A (alpha-Gal A). Without this enzyme, fatty molecules called glycosphingolipids build to toxic levels in cells, causing damage to organs and resulting in disease symptoms.
A Fabry diagnosis usually begins with alpha-Gal A enzyme activity tests
Diagnosing Fabry disease usually involves testing the activity levels of the alpha-Gal A enzyme in a small sample of blood. However, this enzyme activity test has limitations, especially for female patients, whose levels may appear normal despite having the disease.
Researchers in Brazil used a type of infrared spectroscopy called attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy to reveal disease-specific patterns in the molecules present in blood samples from 47 people with classic Fabry disease and compared with samples from 52 matched healthy individuals.
This technique works by shining infrared light onto a sample that is placed on a crystal surface. The infrared light interacts with the sample’s molecules, causing them to vibrate. As the infrared light exits the crystal, it carries information about these molecular vibrations, which is then analyzed using a detector and converted into a spectrum.
Both groups had a similar percentage of females (59.6% vs. 65.4%) and a similar average age (39.2 vs. 36.7 years). Male patients had more severe symptoms than females, but even within the same family, symptoms varied widely. The kidneys were the most affected organ.
“A computational algorithm capable of identifying a specific biomolecular signature for a disease with multifactorial clinical conditions has the potential to achieve greater sensitivity than a single biomarker,” the researchers wrote.
About two-thirds of patients (65.9%) were on enzyme replacement therapy (ERT), which works by providing the body with a working version of the enzyme that is missing. But some were not using ERT, due to either refusing it or being newly diagnosed.
Molecular fingerprint technique showed 100% sensitivity, specificity in study
A machine learning tool was used to compare the ATR-FTIR spectra — the graphs that show the intensity of light being reflected — between patients and healthy adults. The most important difference was in a region of the spectrum that detects fatty molecules, consistent with the buildup of glycosphingolipids.
The technique was highly accurate, achieving 100% sensitivity and specificity in distinguishing between the two groups, regardless of sex, the researchers reported. This means that it identified all patients, while excluding all individuals who did not have Fabry disease.
“ATR-FTIR spectroscopy harnessed to pattern recognition algorithms … [offers] the potential of a fast and inexpensive screening test,” the team concluded. However, “while the results reported herein are promising, future research is required for proof-of-concept.”
About the Author
