As artificial intelligence reshapes healthcare, the fight against cancer is witnessing unprecedented breakthroughs, from microscopic nanorobots to virtual treatment twins. This acceleration in drug discovery is just one example of AI’s impact on oncology, now and in the future.
According to the British Journal of Cancer, cancer research is one of the most heavily-funded areas of science. Billions are poured into research every year but the main ways in which we treat cancer – surgery, chemotherapy and radiotherapy – haven’t changed in the past 80 years.
But now artificial intelligence (AI) is transforming this landscape entirely, revolutionizing not just how doctors diagnose and treat cancer, but how we discover new treatments. “I think we are really in the golden age of scientific discovery, powered by the cloud and AI at scale,” says Dr Rowland Illing, Amazon Web Services’ Chief Medical Officer and Director of Global Healthcare and Nonprofits. At its core, cancer presents a fundamental challenge: cells that go rogue, growing and dividing uncontrollably without the normal genetic ‘braking system’ that regulates cell growth.
“But finding a way to put the brakes on is where it gets interesting,” says Dr Divya Chander, a physician, neuroscientist and futurist. For Chander, there are three key areas where AI is making significant inroads: drug discovery, designing new therapeutics (like living medicines) and precision clinical trials and precision medicine.
From data to diagnosis
Loading...
Traditional drug discovery is expensive and time-consuming but AI models can accelerate the process by identifying shared genomic features across diseases to predict effectiveness. “It’s incredibly helpful because with older medications, we already know they’re safe,” says Chander. “When we find drugs that are useful for another indication, it’s really an extraordinary gift.” This is particularly important for rare and undiagnosed diseases, including certain cancers, that affect over 300 million people globally (but only 5%-7% have FDA-approved treatments).
A new open-source AI model called TxGNN, developed by Harvard Medical School researchers, aims to close this gap by identifying drug candidates for over 17,000 diseases, including rare cancers, using existing medicines. Then there’s Insilico Medicine, a Hong Kong-based biotechnology startup, who now has one of the first AI-discovered drugs to enter the clinical pipeline. The company’s own Pharma.AI platform, which includes tools like PandaOmics and Chemistry42, uses multiple AI models trained on millions of data samples to identify disease targets and design new drug compounds. What’s really exciting about this is the speed. Insilico reached the first phase of clinical trials in just two and a half years, compared to the traditional six-year timeline, and at a fraction of the usual $400 million cost. “It can cost over a billion dollars per drug and take over a decade and a lot of lives are lost in that time,” adds Chander.
This acceleration in drug discovery is just one example of AI’s impact on oncology. Working with Genomics England, Illing and his team faced an immense challenge: analysing thousands of scientific papers about genetic links to diseases. “They have a team of bio curators – experts who do searches, pull out papers, read them and work out the associations between disease types and genes,” explains Illing. But with so much key data on cancer and genetics being published, Genomics England’s manual review process couldn’t keep up with the sheer volume of research. Illing’s team worked on an AI solution – automated bio-curation – which transformed this landscape completely. “Over an eight to 10-week period, we built a generative AI platform that was able to look at all of those papers published and identify 20 new gene associations which previously were not known,” Illing says, adding that if they had manually looked at all the data, it would be out of date. “The role of AI is very much automation. It’s how do you take stuff which needs to be done at scale globally, make it faster and take off the burden of the care providers?”
The digital patient
This revolutionary approach isn’t simply fast-tracking research, it’s shifting how quickly scientific discoveries can become life-saving treatments. Among these innovations, AI is being utilized to create increasingly sophisticated simulations of disease progression in virtual environments called digital twins. In fact, Research and Markets have projected that the digital twin industry will reach $130.77 billion by 2029, driven by demand in healthcare and pharmacology. In a new research paper published in Nature Medicine on digital twins, Dr Eric Stahlberg, a Leidos scientist, explained that “if you look at the aerospace industry, they’re using digital twins extensively to simulate fluid dynamics and materials processing to improve engine and aircraft designs. There are many similarities between those models and newer models that simulate the human cardiovascular and circulatory systems,” he continues. “The complexity of the human body is so substantial that there will always be uncertainty in the models. But ultimately, digital twins will inform patients, help them rank various treatment options and increase their chances of survival.”
UK Biobank, a global biomedical database and research hub, currently has a project underway that’s using AI to create detailed digital twins that mirror how cancer develops and responds to treatment for individual patients. “Digital twins enable precision care in a way that could never be done before,” says Chander. These computer models are groundbreaking because they combine multi-modal data from genetic patterns and protein behavior to chemical changes in the body and patients’ full medical histories. “Most of the models that are coming out to date, like ChatGPT or Claude, have been language models, but we’re seeing a shift into large data models,” explains Illing. “Healthcare isn’t just language – it’s also genomics, proteins and imaging. By integrating all of those different data streams you get a much better idea, upfront, about what kind of treatment a patient can have.” By analyzing this rich mixture of biological and clinical information, UK Biobank’s AI system can spot subtle warning signs that traditional methods might overlook. First tackling pancreatic cancer, where the complex mixture of different cells in tumors makes early detection difficult, the team is hoping to look at lung, breast, colon, prostate, skin and bladder cancers in the future.
These AI breakthroughs are reshaping how doctors approach cancer treatment, turning an uncontrollable disease into one they can analyze, track and target with unprecedented precision. It’s a new era in medicine where cutting-edge technology can customize care. “The promise of personalized medicine is that when you come in with your specific problem, you will be treated as an individual, rather than just one of a ‘this sex, this age, in this environment’,” ends Illing. “AI is already having a massive impact on this kind of tailored understanding of what patients have got but also how things evolve over time.”
Will next-gen oncology be driven by nanorobots?
Remember those illustrations in the 1976 children’s book The Value of Believing in Yourself: The Story of Louis Pasteur? Tiny warriors drawn inside syringes, ready to battle disease. What once captured children’s imaginations is now a scientific reality. Today, AI is turning this vivid imagery into a radical new approach to fighting cancer: nanorobotics – sophisticated molecular machines that are redefining how we target and treat tumors. “These are kind of like rocket ships,” explains Dr Divya Chander, “your rocket ship could be a bacterial cell, or it could be a virus… or it can be a nanorobot that we create, but essentially it can hold a payload. And so, the shell, the main structure, is something that can actually be activated, like a heat seeking missile.” These tiny machines, typically measuring between 1 to 100 nanometers, are programmed to act as precision delivery systems. One breakthrough example comes from Arizona State University, where researchers have developed DNA nanorobots loaded with thrombin. These microscopic warriors seek out cancer cells and effectively starve tumors by cutting off their blood supply – a far cry from the broad, systemic effects of traditional chemotherapy. The development of nanorobots goes hand-in-hand with what Chander calls “living medicines”. Instead of traditional drugs, living medicine uses nature’s own tools (like bacteria or viruses) to fight diseases. It’s almost like having tiny helpers that can do things our bodies can’t do on their own. “A living medicine involves programming a life form to deliver a specific treatment, unlike traditional medicines, which are molecules that spread passively through the body based on physics and chemistry,” explains Chander. “These targeted therapies are extraordinary because they don’t come with the same side effects as systemic drugs… and chemotherapy is toxic.” What makes nanobots so exciting for cancer treatment is their ability in precisely attacking tumor cells. One new study found that nanorobots reduced tumor size by 65.2% compared to chemotherapy’s 40.5% reduction. They also delivered drugs with an impressive 92.7% accuracy, detected tumors nearly twice as fast and caused significantly fewer side effects. Most strikingly, patients treated with nanorobots achieved a 78% survival rate within 12 months, compared to 54% for those on chemotherapy.
Loading...