Stories From the New Frontier in Cancer Research

When Dr. Kunle Odunsi first began treating ovarian cancer patients decades ago, he was haunted by a sobering reality: more than 70 percent would initially respond to treatment, sometimes achieving complete remission. He would tell these patients there was no evidence of disease, and sometimes they would hug him. But at the back of his mind was the knowledge that roughly 70 percent would relapse, often with devastating consequences.

"That became one of the driving forces for me initially," recalls Dr. Odunsi, who now directs the University of Chicago Comprehensive Cancer Center and holds membership in the National Academy of Medicine. "I began to ask several questions: How can we extend remission rates in ovarian cancer patients? Can we use the immune system to prevent relapse of cancer, similar to how we use vaccines for seasonal flu, when a patient is in remission?"

Today, Dr. Odunsi leads a groundbreaking $15 million collaboration with Argonne National Laboratory, using artificial intelligence to revolutionize cancer drug discovery. This partnership exemplifies the profound transformation occurring in cancer medicine. Medical science is changing at a phenomenal rate, with AI serving as the primary driver of this revolution. The new drugs, surgical procedures, and diagnostic processes emerging from AI-powered research are vastly superior to those they are replacing. For anyone facing a life-threatening disease, understanding these advancements could mean the difference between a terminal diagnosis and effective treatment.

Predicting Whether a Patient's Cancer Will Respond to Immunotherapy

At the forefront of this revolution, pioneering physicians are developing AI tools that fundamentally change how cancer is detected, understood, and treated. Dr. Eytan Ruppin, who leads the Cancer Data Science Laboratory at the National Cancer Institute, exemplifies this new generation of physician-scientists combining clinical expertise with computational power.

Dr. Ruppin and his collaborator Dr. Luc Morris at Memorial Sloan Kettering Cancer Center recently developed LORIS (Logistic Regression-Based Immunotherapy-Response Score), an AI tool that uses routine clinical data to predict whether a patient's cancer will respond to immunotherapy. The tool analyzes just six variables commonly measured in clinical settings—including the patient's age, cancer type, blood albumin levels, and neutrophil-to-lymphocyte ratio—to calculate a score indicating response likelihood.

"We were able to develop a new predictive model for immunotherapy response across many different cancer types using only six simple variables," Dr. Morris explains. "In contrast to prior models, some of which are very complex, this model is very accessible to clinicians."

Meanwhile, at Harvard Medical School, Dr. Kun-Hsing Yu has developed CHIEF (Clinical Histopathology Imaging Evaluation Foundation), a ChatGPT-like AI platform that performs multiple cancer evaluation tasks simultaneously. "Our ambition was to create a nimble, versatile ChatGPT-like AI platform that can perform a broad range of cancer evaluation tasks," Dr. Yu explains. "Our model turned out to be very useful across multiple tasks related to cancer detection, prognosis, and treatment response across multiple cancers."

CHIEF successfully predicted patient survival based on tumor histopathology images obtained at the time of initial diagnosis. In all cancer types and all patient groups under study, CHIEF distinguished patients with longer-term survival from those with shorter-term survival, outperforming other models by 8 percent overall and by 10 percent in patients with more advanced cancers.

AI's Diagnostic Revolution

These physician-led advances are delivering unprecedented diagnostic accuracy. AI systems now surpass human diagnostic precision across multiple cancer types. FDA-cleared systems reach 98.7% sensitivity for stroke detection. AI software trained on large datasets of brain scans has proven "twice as accurate" as human professionals in interpreting CT scans, MRIs, and X-rays.

In surgical interventions, AI-assisted robotic procedures reduce operative times by 25% and intra-operative complications by 30%, while improving surgical precision by 40%. AI-powered tools are revolutionizing the detection of musculoskeletal injuries, spotting more bone fractures than humans—crucial given that urgent care doctors reportedly miss broken bones in up to 10% of cases.

The technology extends far beyond imaging. In pilot programs, AI triage systems reduce missed incidental findings in abdominal CTs by over 60%. An AI tool has successfully detected 64% of epilepsy brain lesions that radiologists had previously missed. Most remarkably, AI can detect diseases years in advance with high confidence for over 1,000 conditions, often before symptoms manifest clinically—identifying acute kidney injury risk up to 48 hours before clinical signs appear.

The Precision Medicine Revolution

A primary characteristic of the new AI-driven therapies is drugs custom-designed to an individual patient's disease and genetic characteristics. Advanced treatments now target specific diseases based on precise molecular targets rather than applying broad chemotherapy or radiation that might work for one person but fail for another with the same diagnosis.

AI assesses both the cancer's genetic makeup and the individual patient's genetics, maximizing drug effectiveness while minimizing adverse reactions. Integrated with wearable devices, AI can continuously monitor patient health, detect abnormalities, and even adjust dosages automatically.

One targeted therapy category, monoclonal antibodies, represented 22% of FDA approvals in 2023. These laboratory-designed proteins stimulate our immune system to treat cancer, autoimmune diseases, and infections like COVID-19. These "designer antibodies" are highly specific, binding to single targets for precise immune responses.

In another breakthrough, AI recently enabled the design of new "binders" that can attach to shape-shifting amorphous proteins for the first time, including some implicated in cancer and Alzheimer's. Previously, designing such binders took drug companies months or years. A new AI tool from the University of Washington can design binders for many previously "undruggable" proteins.

This AI software recognizes disordered proteins and generates binders by understanding the target's overall shape, recombining binding pockets in different configurations to create libraries of templates. Using diffusion AI techniques, it generates roughly a thousand pockets allowing for trillions of combinations that can grab onto these challenging proteins.

Global Treatment Advances

Important developments emerge from research centers worldwide. In 2024, the FDA approved 50 novel drugs alongside nine new cellular and gene therapy products, totaling 59 new medical therapies. In Europe, this number reached 114 new medicines, including 46 novel compounds.

In October 2024, researchers announced the biggest breakthrough in cervical cancer treatment in two decades through the INTERLACE trial conducted at 32 medical centers across Brazil, India, Italy, Mexico, and the UK. Results showed that giving cervical cancer patients a short course of chemotherapy before standard treatment reduced death risk by 40% and reduced cancer recurrence risk by 35%.

Some breakthrough drugs don't complete the FDA approval process due to funding limitations rather than efficacy concerns. The pharmaceutical industry spends an estimated $2 billion on average to bring a single drug to market. Companies behind worthwhile drugs sometimes exhaust resources trying to meet U.S. regulatory requirements. At Emerging Cures, we actively seek drugs championed by respected physicians and backed by strong data that didn't complete approval due to funding constraints rather than safety or efficacy issues.

Toward Prognostication

Pattern recognition is only the beginning. Once AI learns to see “what is there,” it can begin to predict “what will happen.” Mayo Clinic researchers recently demonstrated AI's predictive power by examining medical records of pancreatic cancer patients, analyzing abdominal CT scans and blood tests taken up to three years before diagnosis. Using advanced algorithms measuring changes in body fat, muscle, bone, and blood biomarkers like cholesterol and glucose, they identified subtle yet significant body changes occurring as pancreatic cancer develops. In 2023, these models were tested against millions of patient health records. By spotting obscure patterns in disease codes — many not directly tied to the pancreas — the system outperformed population-based estimates and even some genetic sequencing tests.

At Stanford Medicine, researchers developed MUSK (Multimodal Transformer with Unified Mask Modeling), an AI model combining visual and text data to predict cancer outcomes with remarkable accuracy. Trained on 50 million pathology slides and a billion pathology-related texts, MUSK integrated visual and textual data to predict prognosis across 16 cancers. In validation, it correctly forecast disease-specific survival 75% of the time, compared with 64% for conventional staging.

For non-small cell lung cancer, MUSK correctly identified patients who benefited from immunotherapy treatment about 77% of the time, while the standard method based on PD-L1 expression was correct only about 61% of the time.

Looking Forward: The Next Twelve Months

When I asked an AI platform about cancer medicine's trajectory over the next six to twelve months, the response revealed the extraordinary pace of change ahead. By mid-2026, AI's presence in cancer research will shift from innovative tool to indispensable partner, driving increasingly precise, personalized, and rapid advancements.

More AI-powered diagnostic and prognostic tools will gain regulatory approvals and transition from research to widespread clinical use. The European Federation for Cancer Images (EUCAIM) project aims to establish pan-European digital infrastructure for cancer images with at least 50 AI tools by 2026.

AI models will become more sophisticated at integrating vast, diverse datasets—combining genomics, proteomics, metabolomics, spatial transcriptomics, and high-resolution 3D imaging data. This will provide unprecedented insights into tumor microenvironments and immune responses, leading to more effective targeted therapies and immunotherapies.

Multi-cancer early detection tests will see further refinements in sensitivity and specificity, potentially leading to additional regulatory breakthroughs and expanded trials for population-wide screening. AI will play critical roles in distinguishing true cancer signals from background noise, improving test reliability.

Generative AI will move beyond predicting molecule properties to autonomously designing and optimizing entirely novel compounds with specific desired characteristics. This will further shorten drug discovery timelines and reduce R&D costs, with higher probabilities of successful candidates entering clinical stages.

AI-driven adaptive clinical trials will become standard practice, using real-time adaptation to emerging safety and efficacy signals for faster treatment evaluation. AI's role in patient stratification, outcome prediction, and trial automation will become central, making trials more efficient, cost-effective, and inclusive.

There will be increasing emphasis on developing interpretable and "explainable AI" models, allowing researchers and clinicians to understand how AI makes predictions. This will foster greater trust, address ethical concerns around bias, and facilitate regulatory approval and widespread adoption.

The Bottom Line for Patients

AI is dramatically accelerating drug discovery, enhancing diagnostic accuracy, enabling personalized treatments, and boosting medical research efficiency. The transformation is happening now, not in some distant future. Physicians like Dr. Ruppin, Dr. Odunsi, Dr. Kontos, and Dr. Litzow are leading this revolution, developing tools that can identify treatment opportunities traditional approaches might miss.

"I believe this 'dream team' has the potential to revolutionize the cancer drug discovery timeline and change the paradigm for patients that currently have a poor prognosis and little hope for recovery," Dr. Odunsi reflects. The revolution is underway, and it might save your life.

If you face a life-threatening disease, you need to understand what's happening in clinical trials and emerging therapies relevant to your condition. The AI-driven advances emerging from leading research centers worldwide may offer options that weren't available even months ago. Tools like LORIS are already publicly available at https://loris.ccr.cancer.gov for assessing immunotherapy response likelihood. These tools are life-saving, but unfortunately not yet widely deployed.