AI-Driven Chest X-Ray Tool Boosts Lung Cancer Detection

AstraZeneca presented new data today at the European Lung Cancer Congress (ELCC) 2025, demonstrating the potential of artificial intelligence (AI) in improving lung cancer risk detection, especially within resource-limited healthcare settings. The CREATE study utilised Qure.ai’s qXR-LNMS tool to provide real-world data demonstrating its effectiveness to predict the risk of malignancy of Incidental Pulmonary Nodules on a chest X-ray compared to a radiologist’s assessment using the gold standard imaging modality, low dose CT scan. Separately, the CREATE Budget Impact Model illustrates that integrating qXR® into a clinical workflow could be cost-neutral by year 5, specifically utilising Vietnam as the model country.

These findings suggest that AI-integrated CXR could be used as a preliminary step before LDCT in resource-limited settings, to identify individuals at the highest risk and optimise the cost-effectiveness of screening programs. 

Lung cancer is the leading cause of cancer-related deaths worldwide, it is the leading cause of cancer related death. Low-dose computed tomography (LDCT) is the gold standard for lung cancer screening, but its cost and accessibility limit its use in many countries. Chest X-rays (CXR), a helpful preliminary step to LDCT, are the most common, affordable and easily assessable triaging tool in resource-limited settings. 

“To transform patient care and eliminate cancer as a cause of death, we must address the significant and disproportionate burden in low-and-middle-income countries and resource-limited settings via tailored and scalable solutions,” said Iskra Reic, Executive Vice President (EVP), International, AstraZeneca.  

“At Qure.ai, we harness AI to transform routine chest X-rays—especially in resource-limited settings—into an opportunity for earlier lung cancer detection,” said Prashant Warier, CEO and Founder, Qure.ai. “We’re proud to support AstraZeneca’s efforts to enable a more timely diagnosis of lung cancer.” 

The CREATE study, conducted in five countries – Egypt, India, Indonesia, Mexico, and Turkey – involved over 700 individuals and assessed qXR®'s ability to predict the risk of ‘incidental pulmonary nodules’ (IPNs) on CXR being benign or malignant.  

CREATE achieved its primary outcome, demonstrating that qXR® correctly categorises IPNs on CXR as high- and low-risk of malignancy compared with radiologist assessment of LDCT. The positive predictive value (PPV) of qXR® was 54.1%, achieving the pre-defined threshold of success of 20%. The negative predictive value (NPV) was 93.5%, achieving the pre-defined threshold of 70%. 

The results remained consistent across diverse patient groups, including people who have never smoked and individuals under 55 years old, who are typically not eligible to participate in traditional lung cancer screening programs. 

A CREATE Budget Impact Model study assessed the clinical and financial implications of implementing qXR® for lung nodule detection across the entire population of Vietnam. The model predicted that qXR® implementation would lead to 3,155 additional lung cancer diagnoses (at an early stage, when curative treatment is possible) and help prevent 4,742 premature deaths over five years.   

The study concluded that by year five, the implementation of qXR® in Vietnam could reduce premature deaths significantly and achieve cost-neutrality, as earlier cancer detection leads to more efficient treatment and reduced overall healthcare expenses. 

AstraZeneca partners with Qure.ai to explore the utility of AI solutions to improve early-stage disease detection.