There is much debate concerning the uses of AI, but there is also something to be said about how it can be employed to make tuberculosis (TB) screening faster, more efficient, and more accessible, especially in the Philippines, which is flagged by the World Health Organization (WHO) as one of the 10 “high-burden” countries in TB incidence.

Today is World Tuberculosis Day. Last March 19, Siemens Healthineers and the Philippine Business for Social Progress (PBSP) held a media roundtable at The Westin Manila to introduce an AI-powered TB screening program intended to accelerate and streamline the normally long-winded screening process of diagnosing the illness.  

The Siemens-PBSP project is dubbed “Advancing Client-centered Care and Expanding Sustainable Services for TB” (or Access TB). It aims to deploy AI software qTrack—developed in partnership with Indian health tech company Qure AI—to assist in reading digital chest X-rays for potential TB findings in up to two minutes. Waiting time that normally takes two or three days would be condensed to a minute, thus doing away with the hassle of multiple hospital revisits for sputum tests.

Panel members Michael Schmermer, president and managing director of Siemens Healthineers Philippines, Tisha Boatman, Siemens Healthineers executive vice president for external affairs and healthcare access, and Elvin Uy, PBSP executive director, each delivered short remarks followed by a brief Q&A with media representatives. Select NGOs, patient advocacy groups, representatives from Qure AI, Valenzuela City Hospital, and the Department of Health were also present during the event.

Recommended by WHO

“TB is absolutely a number one global health priority, and we want to continue to reinforce that,” Boatman said. “The WHO changed their recommendation a few years ago and suggested that AI-assisted chest X-ray is the optimal way to screen for TB…This is the first disease in the world where the WHO recommends the use of AI in screening.”

The WHO recommendation represents a “significant opportunity” to move the screening process forward for an illness that has been “generally managed in the same way for many, many years,” Boatman said.

“The application of AI can significantly improve the quality of early case finding in early detection,” she added.

The Philippines is considered one of the biggest contributors to new tuberculosis cases in the world. The country’s TB incidence accounts for 6.8%, or about 739,000, of the total 10.8 million global count, and this number does not include “missing” patients, or those who are undiagnosed.

“We’re not doing well on tuberculosis,” PBSP executive director Uy said. “It’s high on the precedent…We have poor health outcomes in tuberculosis and other communicable diseases.”

‘Screen and treat’

Uy said 42% of patients experience “catastrophic costs.” “They are one bad disease episode away from financial or economic ruin.” he said. “That’s unacceptable for a country that will be upper-middle-income, mathematically, by the end of this year.”

Boatman also made note of the “economic and clinical” implications in the TB screening alone, where a patient will have to wait up to two days for test results from the initial screening, and if their results present symptoms of the illness, will have to return to the health facility for a sputum test. “Somebody who was suspected of having TB, has interacted with others, been in their community, potentially had lost wages for a second day,” Boatman said, adding:

“What’s absolutely critical with AI is being able to make a diagnosis on the spot, to see what we call in global health ‘screen and treat.’ Very quickly, a person who is presumptive of TB will be confirmed with a clinical sputum test, and the person can be started on their anti-retrovirals.”

The AI software will be deployed across 28 medical sites, including public hospitals and health centers, in the National Capital Region, Central Luzon, and Calabarzon, which collectively account for more than 60% of tuberculosis cases recorded last year.

Uy estimates that 126,000 people will be screened using the AI-assisted technology across those first three regions. “That’s a big number,” Boatman noted, pointing out that “1.5 million more people per year in the Philippines [would have] access to TB screening.”

Using the qTrack software, the healthcare worker can capture the patient’s risk roots and symptoms by accessing the data in the interactive patient channel. It will be able to detect and localize over 30 chest abnormalities from a chest X-ray, and generate a comprehensive report based on the findings. The healthcare worker can also use the app to follow up on the patient to schedule confirmatory lab tests.

“We expect these tools to help our local health frontliners do screening more accurately, more effectively,” Uy said.

Up front, this will narrow down the timeline for a diagnosed TB patient, who will have to be treated for six months, while also accelerating detection of transmissions within close contacts. 

“We’re talking about results in seconds, instead of manual reading of chest X-rays,” Uy said. “That’s why screening can be a game-changer.”