This new ‘atlas’ of lung cells draws on AI to help scientists detect and treat lung diseases better

Every year 2.2 million people are diagnosed with lung cancer and it claims more lives than any other form of cancer: 1.8 million annually. By 2030, this number could rise to nearly 2.4 million – 30% above today’s incidence. Most patients are detected at the late stages, but if found and treated early, lung cancer is survivable, according to the experts behind a recent report from the World Economic Forum.

Doctors treating lung cancer can now draw on a new resource for understanding the lung, the Human Lung Cell Atlas. Created with the help of AI, the atlas’ authors describe it as the largest and most comprehensive cell map of the human lung. They hope it will be instrumental in the development of new diagnostic tests and therapies.

To create the atlas, researchers merged 49 datasets from nearly 40 separate studies. Using machine learning, more than 2.4 million cells from close to 500 individuals were analyzed as part of the project.

The initiative has already revealed both rare cell types and lung cell variations as well as uncovered commonalities between diseases such as lung fibrosis, cancer and COVID-19. The learnings from the atlas suggest that therapies working for one disease might also apply to the others.

The Human Lung Cell Atlas is part of the Human Cell Atlas (HCA) project, which aims to create a comprehensive map of all human cells. It enables scientists to understand cell types, their roles and their interactions. The initiative has already made significant contributions, including in COVID-19 research.

Applications of AI in healthcare are increasing. In a 2022 survey of healthcare executives in the US, 59% said that deploying AI and machine learning was “very effective” or “often effective” in improving clinical outcomes.

While AI can be instrumental in automating data analysis and generating new insights, as in the case of the Human Lung Cell Atlas, it has many more clinical applications.

For example, it is increasingly being used to detect diseases such as cancer more correctly and at earlier stages, according to a PwC report. There is evidence that AI can speed the review of mammograms by a factor of 30 with 99% accuracy. This means fewer biopsies are needed, benefitting both the patient and stretched healthcare budgets. AI is also supporting doctors in diagnostics, clinical decision-making and treatment.

However, the Human Lung Cell Atlas relied not only on AI and machine learning but on nearly 100 partners from more than 60 academic departments for integrating disparate data. This included researchers from renowned institutes such as Germany’s Helmholtz Munich research centre for health and environment, the University Medical Center Groningen in the Netherlands and Northwestern University in the US.

The results achieved by this large-scale collaboration underscore the need to leverage new technologies, and for international action around lung cancer and in global health at large.

The World Economic Forum is aiming to further progress towards achieving SDG 3.4 (reducing non-communicable diseases and promoting mental health) through platforms such as its Lung Cancer Collaboration initiative, the Centre for Health and Healthcare, and regular reports such as the Global Health and Healthcare Strategic Outlook.