Researchers in UABC created a computer tool that analyzes chest X-rays, allowing medical personnel to diagnose more accurately.

A student and his professors at UABC developed an AI-based computational tool for analyzing chest X-rays to make a quick, timely, and accurate diagnosis of tuberculosis (TB). This ailment continues to affect the most vulnerable people in underdeveloped countries. The computer tool was registered with the Mexican Institute of Intellectual Property under MIA-TB-Rx.

Tuberculosis (TB) is an infectious disease caused by bacteria of the Mycobacterium tuberculosis complex. It is transmitted mainly through the air, from one person to another via droplets expelled by TB patients, and it manifests primarily in the lungs. Still, it can also be present in the nervous system, bones, skin, intestines, genitalia, ganglia, and other organs.

One of the pillars in the fight against TB is timely diagnosis. However, many countries, including Mexico, base the diagnosis of TB on smear microscopy, an inexpensive technique used for more than 100 years, with the disadvantage that it presents many limitations in sensitivity and specificity levels that do not exceed 60%-70%, which makes cases of “false negative” very common.

The problem is complicated when the person with a false negative, not considered sick, does not receive the appropriate drug treatment, dispersing the pathogen in the community. Studies have shown that one person with untreated tuberculosis can infect 15 other people in a year.

On the other hand, chest radiography (Rx) is an established diagnostic test for detecting and screening pulmonary TB, which is fast and cheap compared to other tests. Its primary disadvantage –low specificity when analyzed by health personnel– was solved by implementing a computer-assisted diagnosis system, thus achieving high sensitivity and specificity values.

MIA-TB-Rx is the product of the master’s thesis of Miguel-Angel Guerrero-Chevannier. The co-directors of this thesis and responsible for the research project are Dr. Dora Luz Flores-Gutiérrez from the School of Engineering, Architecture and Design, and Dr. Raquel Muñiz-Salazar from the School of Health Sciences.

The collaborators are professors Rafael Laniado-Laborín, Ricardo Perea-Jacobo, Rogelio Zapata-Garibay, J. Eduardo González-Fagoaga, Monserrat Luna-Valderrabano, bioengineer Guillermo R. Paredes-Gutiérrez and student Irvin Eduardo Zavala-Román.
According to WHO, in 2021, 10.6 million people fell ill with tuberculosis, and an estimated 1.6 million died from this disease, equivalent to ten deaths per day worldwide.