Diagnosis of the future by facial analysis on the computer

Researchers are testing computer-aided facial analysis for diagnosis

Digitization will revolutionize healthcare, both in terms of diagnostic options and treatment options. Scientists from the Berlin Charité and the University Hospital Bonn have already successfully tested computer-aided facial analysis as a diagnostic tool.

"In rare diseases, the computer-aided image analysis of patient portraits can facilitate diagnosis and significantly improve it," reports the Rheinische Friedrich-Wilhelms-Universität Bonn from the study results. On the basis of the clinical picture of the GPI anchor disorders (also GPI Ankersynthese disturbances) the scientists could show that the computer-aided evaluation of portraits offers a good starting point for the diagnosis. The researchers published their results in the journal "Genome Medicine".

Computerized facial analysis could significantly improve the diagnosis of disease in the future. (Image: babsi_w / fotolia.com)

Improvements in diagnostic options investigated

The international research team led by Dr. Ing. Alexej Knaus and Prof. Krawitz from the Institute for Genomic Statistics and Bioinformatics of the University Hospital Bonn and Prof. Dr. med. Denise Horn from the Charité Institute of Medical Genetics and Human Genetics has investigated in this study how computerized facial analysis can help improve the diagnosis of GPI anchor destruction. It was also investigated whether modern, particularly fast DNA sequencing methods and surface analyzes of cells can improve the diagnostic possibilities.

What are GPI anchor failures??

An example from the group of GPI anchor disorders is the so-called Mabry syndrome. This rare disease leads to mental retardation and is triggered by the change in a single gene, the scientists report. The syndrome belongs "to a group that we describe as GPI anchor damage and which includes more than 30 genes", explains Prof. Dr. med. Peter Krawitz from the Institute for Genomic Statistics and Bioinformatics of the University Hospital Bonn.

Signal transmission and transport of the cells disturbed

The "anchors" of glycosylphosphatidylinositol (GPI) adhere to the outside of the cell specific proteins and if this does not work properly by a gene mutation, the signal transmission and transport within the cell and between the cells are disturbed, the researchers report. This also has an influence on the outward appearance of those affected. However, the spectrum of the external appearance of the GPI anchor disruptions is wide-ranging, which makes it difficult to identify those affected.

Effects on the appearance

"The effects of a mutation in a particular gene can be very mild to very strong"; the researchers report. This also applies to the abnormalities on the face. For example, in the Mabry syndrome "a narrow, sometimes tent-shaped upper lip, a broad nose root and a large eye relief with long eyelids" are known as classical features. However, these can be more pronounced or weaker, which often complicates the diagnosis of the rare disease.

DNA sequencing, cell surface analysis, computer aided image analysis

In addition, with Mabry's syndrome, the eponymous elevation of alkaline phosphatase in the blood can not be detected in every patient, "patients and their loved ones often go through a yearlong odyssey until the right diagnosis is made," Krawitz says. New DNA sequencing methods, cell surface analysis and computer-aided image analysis could significantly improve the diagnostic capabilities in the future.

Typical facial features recognizable

In their current studies, the researchers used the images of the faces of a total of 91 patients. The analysis of the cell surfaces in the subjects also showed that some of the participants had the typical changes for GPI anchor destruction. Also in the analysis of the genetic material gene mutations were identified that are typical of this rare disease group, the researchers report. The researchers then used "artificial intelligence techniques to simulate models of disease from genetic data, cell surface texture, and facial features."

Diagnostic options significantly improved

The artificial modeling of gene-typical faces based on the available datasets has, according to first author Dr. "The computer-assisted evaluation of patients' portraits clearly facilitates and improves the diagnosis of GPI anchor damage." The study's findings are "great progress," according to the researchers, and hope to use combined data from the lab and the computer to better understand also molecular processes of such diseases.

Facial analysis also applicable to other diseases?

For example, an increased value for the alkaline phosphatase in the blood with conspicuous results in the image analysis is a reliable indication for the classification of a new mutation in a GPI anchor destruction, explain the experts. The novel combination of cell and genome analyzes and computer-aided image analysis is trend-setting in their view. "Foreseeable, these methods can also be applied to other diseases. This would be a major leap forward in terms of diagnosis, "says Prof. Krawitz. (Fp)