AI-based chest X-ray (CXR) screening for tuberculosis (TB) is gaining traction but faces significant real-world challenges in implementation, such as software setup, integration into workflows, network connectivity issues, and a shortage of healthcare personnel. To tackle these issues, PATH initiated a TB screening program in Nagpur, India, utilizing qXR, an AI software tool, across eight private CXR labs that met specific technical prerequisites. The program revealed valuable insights into the operational feasibility of AI deployment in resource-limited settings, highlighting the strategies needed to overcome logistical hurdles.
Back
Throughout the program, 10,481 individuals were screened based on clinical history, resulting in 2,303 flagged as presumptive TB cases by either AI or radiologists. Thanks to AI, there was a 15.8% increase in TB case detection compared to traditional screening methods alone. This increase underscored the effectiveness of AI as a complementary diagnostic tool, especially in settings with limited healthcare resources.
TB continues to be a major public health issue in India, which reports nearly a third of the world’s TB cases. With India’s goal to eliminate TB by 2025, innovative approaches to diagnostics and patient care are crucial. Although CXRs are a recommended screening tool in India, limitations in resources—such as a shortage of radiologists—make consistent implementation difficult. AI solutions like qXR from Qure.ai, designed to assist in identifying TB signs in CXR images, offer promise but must overcome practical barriers in integration and use in diverse healthcare settings.
The Nagpur screening program encountered several technical and operational challenges that were addressed through strategic adaptations. Basic infrastructure, including internet access and compatible systems, proved critical for smooth AI software installation and operation. Initial reluctance from local radiology associations, driven by unfamiliarity with AI, was overcome through educational sessions, transparency, and consistent engagement with stakeholders, which ultimately built trust and acceptance of the new technology. Additionally, flexible technical setups allowed labs to address privacy concerns while accommodating AI functionality. By working closely with local partners, PATH was able to implement qXR in a way that respected the needs and limitations of each lab, leading to improved screening outcomes.
Overall, the increased TB detection rate achieved through the use of AI screening underscores the potential of AI-based tools to transform TB screening in India and other high-burden areas. However, broader deployment of these tools will require carefully designed solutions to address the real-world challenges of workflow integration, infrastructure compatibility, and stakeholder acceptance. This experience shows that with the right approach, AI can be a powerful asset in advancing TB care, particularly in resource-constrained environments.
