WCLC 2025 - Posters & ePosters-P1.17.31 Feasibility of Integrating Ai-Assisted Lung Cancer Screening in High Tuberculosis Burden Settings

Pdf Summary

This multi-country feasibility study evaluated integrating AI-assisted lung cancer (LC) screening into existing tuberculosis (TB) screening programs using chest X-rays (CXRs) in high TB burden settings in Ethiopia, the Philippines, and Vietnam. Given the clinical and radiological similarities of TB and lung cancer, which often lead to misdiagnosis and delayed lung cancer treatment, the study explored leveraging AI software (qXR Lung Nodule Malignancy Score by Qure.ai) already used for TB detection to opportunistically flag high-risk incidental pulmonary nodules indicative of lung cancer during routine TB screening.<br /><br />From a total of 198,271 adults screened with CXRs across the three countries, the AI flagged 3,518 individuals (1.8%) as having potentially malignant lung nodules. Of these, 63.5% had further clinical evaluation data available, including 725 (32.5%) who underwent low-dose computed tomography (LDCT) scans. Radiologists reviewed LDCT images using the Lung-RADS scoring system to assess malignancy risk, guiding follow-up and treatment according to local clinical guidelines.<br /><br />The findings demonstrate the operational feasibility and potential value of integrating AI-enabled lung cancer screening within existing TB screening workflows without requiring substantial additional resources. This integration presents a cost-efficient strategy to maximize diagnostic yield, reduce lung cancer underdiagnosis, and improve early detection in settings lacking dedicated lung cancer screening programs. The overlapping TB and lung cancer symptoms, such as cough and weight loss, make this dual screening particularly relevant.<br /><br />Future research priorities include longitudinal follow-up to measure impact on early-stage lung cancer diagnosis and mortality, and thorough evaluation of cost-effectiveness and health system implications to inform policy decisions and potential scale-up. Funding for AI software and LDCT referrals was supported by Qure.ai and AstraZeneca grants. Overall, the pilot provides promising evidence for pragmatic, resource-optimized lung health screening strategies in high TB burden countries.

Asset Subtitle

Saniya Pawar

Meta Tag

Speaker Saniya Pawar

Topic Global Health, Health Services, and Health Economics

Keywords

AI-assisted lung cancer screening

tuberculosis screening programs

chest X-rays (CXRs)

high TB burden settings

qXR Lung Nodule Malignancy Score

incidental pulmonary nodules

low-dose computed tomography (LDCT)

Lung-RADS scoring system

integrated screening workflow

cost-effective lung health screening