Integrating artificial intelligence with circulating tumor DNA for non-small cell lung cancer: opportunities, challenges, and future directions

Abstract

Non-small cell lung cancer (NSCLC) remains a leading cause of cancer mortality, with late-stage diagnosis contributing to poor survival. Circulating tumor DNA (ctDNA) has emerged as a non-invasive biomarker for screening, diagnosis, and monitoring, with limitations about sensitivity and specificity challenges. The integration of artificial intelligence (AI) offers a promising avenue to enhance ctDNA applications in NSCLC by improving mutation detection rates and sensitivities, refining minimal residual disease (MRD) predictions, enabling earlier detection of relapse, sometimes earlier than imaging, differentiating tumor vs. non-tumor derived signals to improve specificities. AI achieves 0.002% mutant allelic fraction detection, 94% relapse detection sensitivity, and 5.2-month lead time over imaging. This narrative review explores the role of ctDNA in NSCLC management, highlighting how AI amplifies its utility across screening, diagnosis, treatment evaluation, MRD detection, and disease surveillance while outlining key opportunities, challenges, and future directions.

Keywords: artificial intelligence; circulating tumor DNA; lung cancer; minimal residual disease; screening.

Figure 1

Blood is collected from patients and ctDNA is extracted from blood plasma and mutations can be analyzed by next generation sequencing involving a few steps, including DNA extraction, DNA library preparation, sequencing, sequence alignment, mutation annotation, and so on (58).