AI X-ray for tuberculosis screening in remote Nepal: Benefits and challenges from a doctor's perspective

Area of Science:

• Public Health
• Medical Imaging
• Infectious Disease Control

Background:

• Community-based pulmonary tuberculosis (TB) screening traditionally relied on symptom assessment.
• Subclinical TB cases, though asymptomatic, can be infectious and exhibit radiographic abnormalities.
• Advancements in medical technology are crucial for improving early detection rates.

Purpose of the Study:

• To evaluate the effectiveness of AI-enhanced ultraportable X-ray machines for community TB screening.
• To assess a parallel screening strategy combining chest X-rays and symptom assessment for TB detection.
• To determine the role of AI in identifying both clinical and subclinical TB cases in the community.

Main Methods:

• Introduction of AI-enhanced ultraportable X-ray machines for community screening.
• Implementation of a parallel screening strategy: chest X-ray plus symptom assessment.
• Confirmatory sputum testing for positive cases identified through screening.

Main Results:

• Chest X-ray screening is more efficient and practical than symptom-only assessment for TB.
• The combined strategy effectively detects both symptomatic and asymptomatic (subclinical) TB cases.
• AI-assisted portable X-ray devices represent a significant advancement in community TB screening.

Conclusions:

• AI-enhanced portable X-ray screening, coupled with symptom assessment and sputum testing, is an effective strategy for community TB detection.
• This integrated approach enhances the ability to identify infectious individuals, including those with subclinical disease.
• While promising, further evaluation of limitations within this AI-driven screening model is warranted.