Artificial Intelligence (AI)-powered Chest X-ray in Primary Care Accelerates Time to Lung Cancer Diagnosis

Abstract

Abstract Introduction: Lung cancer is the leading cause of cancer-related deaths worldwide. Timely evaluation and referral in primary care are difficult due to its non-specific symptoms. In high tuberculosis (TB) burden areas, misdiagnosis of cancer as TB is a serious issue. Interpreting Chest X-Rays (CXR) in primary care is also challenging. Although recent advances in Artificial intelligence (AI)-based software have shown promise in flagging lung nodules, its impact on reducing diagnostic delays in primary care has not been widely reported. Case presentation: We describe the diagnostic approach to a patient attending a rural primary care centre that is managed by a single family physician. A 38 year old male presented with a three-day history of productive cough and dyspnoea (mMRC scale grade II). He reported no haemoptysis, fever, weight loss, or appetite loss and had no family history of malignancy or TB exposure. He smoked cigarettes infrequently (0.45 pack-years). Vital signs were stable, and physical examination findings were unremarkable. AI (qXR, Qure.ai) software flagged a suspicious nodule on the CXR in the right lower zone (Figure) and AI result for TB was negative. Sputum for acid-fast bacilli was negative. Patient was recommended a contrast computed tomography (CT), which revealed an enhancing lesion with micro-lobulated margins in the right lower lobe. Subsequent ultrasound-guided biopsy confirmed a diagnosis of adenocarcinoma of the lung. The patient commenced targeted therapy at a cancer treatment center. The duration from CXR acquisition to cancer diagnosis was 5 days. Novelty and importance of the case: The case report highlights the real-world impact of CXR AI in enabling access to radiology services in primary care, especially in rural and remote settings. The AI flagged a suspicious nodule that prompted early referral for CT imaging and biopsy. In India, 80% of patients first consult their primary care physician (PCP) and there is a significant delay (1-6 months) before specialist referral. AI CXR can alert PCPs and non-specialists to possible lung cancer. In high TB burden countries, AI integration in a combined TB and lung cancer screening model may help prevent misdiagnosis. AI CXR can also help detect malignant nodules in younger, low-risk populations (like this patient) who may not qualify for low-dose CT (LDCT) screening under current guidelines. This case demonstrates the utility of AI-powered CXR in accelerating lung cancer detection in resource-limited settings and improve patient outcomes.

Ähnliche Arbeiten