The Role of AI in Lung Tuberculosis Screening for the Community

The quality of Chest X-ray (CXR) image interpretation is a crucial criterion for detecting active tuberculosis (TB) cases in both the community and healthcare facilities. AI contributes to an increased percentage of abnormal findings in suspected TB cases, significantly enhancing accuracy of CXR interpretation and functioning as an independent software, surpassing subjective factors that may influence results. The screening results through X-ray film evaluation are accurately assessed up to 95%, affirmed Dr. Pham Ngoc Thanh (Hai Phong Lung Hospital, Vietnam). 

According to USAID, AI software has assisted in detecting an additional 38 individuals with abnormal CXR showing signs suggestive of TB. This occurred during a screening event involving over 2,300 individuals at Tien Giang Hospital, Vietnam in July 2023 (Source: ttps://www.usaid.gov/vi/vietnam/news/jul-09-2023-usaid-utilizes-artificial-intelligence-improve-tuberculosis-tb-diagnosis-vietnam). 

The benefits are evident: saved time and reduced costs. AI software like DrAid™ for TB Screening sits at the heart of the 2X strategy, a nationwide initiative integrating chest X-rays with rapid diagnostic tests like Xpert MTB/RIF/Xpert Ultra/TrueNat. This innovative approach is streamlining tuberculosis screening across diverse settings.

Let's explore the flowchart of the tuberculosis screening process in the community through the 2X strategy.

Here is the flowchart of the TB screening process in the community through the 2X strategy. 

Step 1. Participants undergo CXR imaging (except for pregnant individuals). 

Step 2. AI will assist in diagnosis, classifying the CXR results into: 

• Abnormal cases with Suspected tuberculosis 
• Abnormal cases without suspicion of tuberculosis 
• Normal cases 

Step 3. Based on the CXR results: 

• Individuals with suspected tuberculosis will be directed for Xpert MTB/RIF or Truenat Smear test in tuberculosis diagnosis. 
• Those with abnormal results without suspicion of tuberculosis are recommended to visit a specialty clinic for further examination.  
• Individuals with normal CXR images are scheduled for follow-up examinations at the tuberculosis control unit for further diagnosis. 
• Contact people with TB-patients, sensitive to medications, aged 5 and above, at risk, are advised to undergo Tuberculin Skin Test (TST), a skin injection to determine latent tuberculosis. 

DrAid™ for TB Screening acts as a radiologist's AI sidekick, slashing case analysis time from 15 minutes to a few seconds during community screening. It saves costs by skipping unnecessary smear tests in clear X-rays, and its digital format reduces film waste, redefining healthcare and TB screening. 

Innovative technologies like DrAid™ for TB Screening will continue to strengthen the 2X strategy with the goal of detecting, treating, and preventing more TB cases. This approach aims to discover and treat more individuals with pulmonary TB, working towards the goal of ending tuberculosis by 2030.