A new study suggests that artificial intelligence (AI) may help clinicians diagnose tuberculosis in regions of the globe where radiologists are scarce. As per US News, researchers reported at a European conference this week that AI software effectively identified tuberculosis (TB) from cellphone images of chest X-rays.
“We have demonstrated that AI software is at least as good at detecting tuberculosis as a trained radiologist and that a simple mobile phone photograph is sufficient for analysis,” said Dr. Frauke Rudolf of the infectious diseases department of Aarhus Hospital in Denmark.
Chest X-rays serve a crucial role in detecting tuberculosis in patients unable to produce sputum (phlegm) samples of sufficient quality for microbiological analysis. In areas with limited resources and few radiologists, using software to aid in the diagnosis of conditions based on X-rays could be beneficial.
The objective of Rudolf’s team was to ascertain the accuracy of this method. In evaluating chest X-rays, they compared the efficacy of AI software to that of two Ethiopian radiologists with varying levels of experience. The AI was provided with mobile phone images of non-digital chest X-rays. The authors of the study noted that 11% of 498 patients had been diagnosed with tuberculosis, 41 clinically and 16 through PCR assays.
In terms of identifying PCR-confirmed cases, the software was as effective as or better than a radiologist. It correctly detected approximately 86% of non-TB cases and 75% of PCR-confirmed cases. Comparatively, the less-experienced radiologist correctly identified approximately 63% of the PCR-confirmed cases and nearly 92% of those who did not have TB. The seasoned radiologist correctly identified 75% of PCR-confirmed cases and 82% of those without tuberculosis.
“With an estimated 3 million undiagnosed patients in 2021, there is an urgent need to develop novel strategies and technologies aimed at improving TB detection in low-resource, high-incidence settings,” said Rudolf in a press release from the European Congress of Clinical Microbiology & Infectious Diseases (ECCMID).
TB is a leading cause of death and disease worldwide, claiming the lives of 1.6 million individuals annually. It is the thirteenth highest cause of death worldwide and the second leading infectious cause of death after COVID-19.
Rudolf stated, “In low-resource regions with a high incidence of tuberculosis but a lack of radiologists, chest X-rays could be photographed with a mobile phone, and the image sent to be remotely analyzed by AI.” This would enable more chest X-rays to be properly interpreted and, most importantly, allow more TB cases to be diagnosed.
The study was slated to be presented at the ECCMID in Denmark on Monday. Before publication in a peer-reviewed journal, findings presented at medical conferences should be regarded as preliminary.