Can AI Change Cancer Care?

Artificial Intelligence (AI) is rapidly transforming the world of cancer care. From early detection to personalized treatment and drug discovery, AI is helping researchers and clinicians make faster, smarter, and more accurate decisions. By leveraging Real-World Data (RWD) from clinical trials, Electronic Health Records (EHRs), medical imaging, genomic databases, and insurance claims, AI systems can identify patterns that would be nearly impossible for humans to detect alone. The result is a new era of more precise, data-driven oncology care.

Screening and early detection

One of the most promising areas of AI in oncology is cancer screening and early detection. AI-powered imaging tools are already improving the sensitivity and efficiency of cancer detection. In large-scale mammography studies, AI-assisted screening increased cancer detection rates by nearly 18% while also reducing radiologist workload. This represents a major leap forward, not only in improving diagnostic accuracy but also in helping address physician burnout and increasing efficiency in healthcare systems.

AI is also reshaping pathology and radiology through the use of Deep Learning (DL) models. These systems can analyze medical images to detect tumors, segment cancerous tissue, and even grade disease severity with impressive accuracy. Pathologists can now identify metastases in lymph nodes more quickly and consistently, helping clinicians make faster treatment decisions for patients.

Precision oncology

Perhaps the most exciting development is in precision oncology. AI can integrate and analyze multi-omics data — including genomics, proteomics, and radiomics — to identify tumor driver mutations and predict how patients may respond to specific therapies. This marks a major shift away from the traditional “one-size-fits-all” model of cancer treatment toward highly personalized medicine. By understanding the molecular characteristics of an individual patient’s tumor, clinicians can select targeted therapies that are more likely to produce better outcomes and fewer side effects.

Drug discovery and clinical trials

Drug discovery and development is another area where AI is having a significant impact. Traditionally, developing a new cancer drug can take years and cost billions of dollars. AI is helping accelerate the identification of therapeutic targets and promising drug candidates, potentially shortening the development timeline considerably. Recent initiatives by the U.S. Food and Drug Administration to support real-time clinical trials (RTCTs) highlight how regulatory agencies are also embracing data-driven innovation. In these models, clinical trial endpoints and safety signals can be reported and analyzed in real time, improving efficiency, reducing costs, and enhancing patient safety.

Clinical trials themselves are becoming smarter through AI. Recruiting eligible patients has historically been one of the biggest bottlenecks in oncology research. AI can analyze EHRs, genomic databases, and clinical registries to identify patients who meet complex eligibility criteria much faster than traditional methods. In addition, computational models can simulate treatment responses and optimize trial design before studies even begin. Recent research has shown that AI models incorporating protein structure analysis and functional genomic data outperform traditional methods in identifying cancer-driving mutations. AI has also helped interpret variants of unknown significance (VUSs) in genes such as KEAP1 and SMARCA4, which have been associated with poorer outcomes in non-small cell lung cancer patients.

Language models and multimodal AI

Behind many of these advances is the growing use of Natural Language Processing (NLP) and Large Language Models (LLMs). These technologies can extract meaningful clinical insights from unstructured medical data such as pathology reports, physician notes, and radiology summaries. At the same time, multimodal AI systems are beginning to combine imaging, genomic, and clinical data into unified predictive models that can estimate patient outcomes more accurately than any single data source alone.

Where this still falls short

Despite its enormous promise, AI in cancer care still faces important challenges. Real-world healthcare data is often incomplete, inconsistent, and fragmented, requiring extensive curation before it can be used effectively. Privacy and security concerns remain critical, particularly when dealing with sensitive patient information across institutions. Researchers are increasingly exploring privacy-preserving approaches such as federated learning, which allows AI models to train