AI in Lung Cancer Treatment

How AI Is Improving Lung Cancer Treatment: New Tools, Real Impact

See how AI is helping guide immunotherapy decisions, combine imaging for deeper insights, and support earlier detection—bringing more personalized lung cancer care.

Three Ways AI Is Moving Lung Cancer Care Forward

1) AI that predicts immunotherapy response in advanced NSCLC (Deep-IO)

A validated deep-learning model called Deep-IO may help doctors better predict which patients with advanced non-small cell lung cancer (NSCLC) will respond to immunotherapy—using routine pathology slides from tumor biopsies. In the study discussed in a Physician’s Weekly Q&A, researchers trained the model on hundreds of thousands of image “tiles” from 958 patients (US + EU cohorts, 2014–2022). The model showed meaningful accuracy for predicting response (reported AUCs varied by cohort) and, importantly, it may help identify non-responders more specifically—potentially sparing patients ineffective treatment and side effects. The article also notes that combining Deep-IO + PD-L1 performed better than either marker alone in that validation setting.¹

2) “Connecting the dots” across cancer images (Emory)

A separate Emory report highlights how AI is being used to combine different kinds of cancer images—from microscopic tissue slides to CT scans and even epigenetic data—so researchers can build a more complete picture of how a tumor behaves. In four studies (focused on head and neck cancers), one team used an AI platform called VISTA to transform standard H&E slides into “virtual IHC” to help identify tumor-associated macrophages that are otherwise difficult to see. Another approach adapted a swin transformer into a multimodal framework (called SMuRF) that helped integrate 2D pathology images with 3D radiology, supporting predictions about survival and even which patients might benefit from chemotherapy. A fourth effort linked slide images with epigenetic patterns using pathogenomic fingerprinting, aiming to improve risk assessment. Emory’s researchers also emphasize the need to be cautious and thoughtful before moving these tools fully into clinical practice.²

3) What this means for lung cancer patients right now

Together, these developments point to a near-future where AI can support:

More personalized treatment selection (who is most likely to benefit from immunotherapy)³
Smarter risk and prognosis tools by combining radiology + pathology + other data⁴
Earlier detection and faster pathways to care, especially as imaging AI expands in real-world settings⁵

It’s also important to keep the balance: experts continue to flag challenges like bias and fairness, the need for diverse multi-site datasets, and careful clinical validation before widespread adoption.⁶

More examples of AI in cancer research and care

NCI’s HistoTME model (NSCLC + immunotherapy): An NCI team described HistoTME, which analyzes digital pathology images to learn about the tumor microenvironment and improve prediction of immunotherapy response—supporting biomarker discovery and more personalized immunotherapy strategies.⁷
AI-driven early lung cancer detection (Bristol Myers + Microsoft, Jan 20, 2026): Reuters reports a collaboration using FDA-cleared radiology AI algorithms through Microsoft’s Precision Imaging Network to help clinicians detect lung nodules earlier and expand access in underserved communities.⁸
AI + breast cancer screening at scale (EDITH trial, UK): A UK government announcement describes a large NHS trial (~700,000 participants) evaluating AI to support mammogram reading and potentially reduce the need for a second specialist reader.⁹
Multimodal AI for recurrence risk (AACR, Dec 10, 2025): AACR describes an AI model combining digitized pathology slides with molecular + clinical data to improve long-term recurrence risk stratification in early breast cancer.¹⁰
AI to speed oncology drug research (AstraZeneca + Modella AI, Jan 13, 2026): Reuters reports AstraZeneca agreed to acquire Modella AI to boost quantitative pathology and biomarker discovery for oncology R&D.¹¹

AI is a supportive tool

AI isn’t replacing oncologists—it’s becoming a powerful support tool that can help doctors see patterns humans can’t easily spot, match patients to therapies more precisely, and potentially reduce trial-and-error in treatment. If you or someone you love is living with lung cancer, it may be worth asking your care team about biomarker testing (like PD-L1), available clinical trials, and how new imaging or pathology tools are shaping treatment decisions.¹²

Resources for Living With Stage 4 Cancer

Whether you’re navigating treatment options, seeking emotional support, or trying to make sense of a new diagnosis, Stage 4 Hope is here for you. Become a member of our community to gain access to trusted resources, as well as online support and more.

References: