Researchers are increasingly turning to artificial intelligence (AI) to customize cancer therapies and explore new applications for existing medications. Some systems can analyze microscopic images or detect subtle biological indicators that might go unnoticed by humans.
While many of these technologies have shown promise in assisting patients, others are still in clinical trials or research phases, highlighting the importance of distinguishing between what’s experimental and what’s currently available as treatment.
What researchers are achieving today would have seemed almost impossible a few years back. Here’s a look at how AI is making strides in medicine and some considerations you should bear in mind before relying on it for your health.
AI is Personalizing Experimental Melanoma Treatments
A noteworthy advancement has come from Moderna and Merck, which recently shared positive initial results from a Phase 3 melanoma trial involving a treatment called intismeran autogene (also referred to as V940 or mRNA-4157) used with Keytruda. The trial engaged 1,137 participants with high-risk melanoma that had been fully excised through surgery. Notably, it achieved its main goal related to recurrence-free survival, in addition to a secondary goal focused on avoiding distant metastasis.
This represents the first positive Phase 3 results for an individualized neoantigen therapy as well as mRNA-based cancer treatment. The concept, while complex, is intriguing: researchers start by analyzing a patient’s tumor sample for unique mutations. Using algorithms, they identify targets that may help the immune system recognize cancer cells. This tailored treatment can encode up to 34 neoantigens, with integrated AI algorithms playing a role during development.
Understanding Melanoma Trial Results
There’s a lot of excitement surrounding these findings, but it’s essential to note some limitations. Currently, only topline results have been revealed, and the full data presentation is awaited at an international medical conference, along with necessary regulatory discussions. The overall survival rates of participants are still under review.
Earlier, a smaller Phase 2b trial indicated that the combination of intismeran and Keytruda led to a 49% reduction in recurrence or death risk compared to Keytruda alone, alongside a 59% decrease in the risk for distant metastasis or death. Those earlier findings were from a smaller participant pool, making the Phase 3 results significantly important. Yet, it’s crucial to remember that intismeran is still in the investigational phase and not yet approved by the FDA for treating melanoma.
AI Identifies New Uses for Existing Drugs
Developing new medicines is a lengthy process, so another research group is exploring whether effective treatments may already be out there. Dr. David Fajgenbaum founded the nonprofit Every Cure to investigate this potential. According to their 2025 report, around 18,000 diseases are recognized globally, but only about 4,000 have FDA-sanctioned medications, leaving a large number of diseases with few treatments available.
Every Cure employs AI to sift through biomedical knowledge, searching for connections between current drugs and alternative diseases they might help. Their AI system claims to produce millions of potential predictions within a day. Promising avenues are then further investigated by researchers.
The federal Advanced Research Projects Agency for Health (ARPA-H) is supporting this initiative through a project named MATRIX, which uses machine learning to pinpoint existing FDA-approved drugs that could treat other diseases. While AI doesn’t substantiate that a drug will effectively treat another condition, it can refine the search process significantly.
Fajgenbaum has witnessed the life-changing impact of repurposing drugs. For example, Kaila Mabus was diagnosed with multicentric Castleman disease at age 13 and became gravely ill even after chemotherapy. In 2020, her doctors used ruxolitinib, a drug meant for blood disorders but not sanctioned for Castleman disease, leading to her improvement and remission within months. Although AI didn’t discover her treatment, her experience underscores the urgency behind Every Cure’s goal of rapidly revealing new drug-disease connections.
AI Detects Sperm Cells Often Missed by Conventional Methods
Shifting to a different application, Columbia University Fertility Center has developed the Sperm Tracking and Recovery system (STAR), which combines speed imaging with AI models and microfluidics.
Designed for patients suffering from azoospermia or cryptozoospermia—conditions where sperm may either appear absent or exist in very limited quantities—STAR analyzes semen samples much more thoroughly than traditional manual methods allow. The system can capture and process about 1.1 million images each hour to identify sperm cells.
Upon confirming a sperm cell, a microfluidic mechanism isolates it for potential fertility treatment or future use. In one instance, embryologists searched for sperm for two days without success, while STAR managed to find 44 sperm in just one hour, showcasing the efficacy of AI in tedious tasks where human oversight can falter.
A Blood Test Using AI Could Indicate Heart Risks Early
At the University of Hong Kong, researchers are looking into a different avenue. Their AI tool, called CardiOmicScore, analyzes blood molecular data by examining 2,920 circulating proteins and 168 metabolites along with genomic information sourced from the UK Biobank.
This tool uses deep learning methods to predict the risk of six cardiovascular diseases, including coronary artery disease and stroke. Notably, it can identify increased risk as much as 15 years prior to any symptoms surfacing, which could transform how we manage cardiovascular diseases in the future.
However, as with many of these technologies, CardiOmicScore is still in research development and isn’t available in routine screenings yet.
Lab-Grown Tumors Aim to Guide Treatment Choices
Researchers at UCLA are investigating a different method for tailoring cancer treatment through the creation of tiny lab-grown tumor replicas known as organoids. By exposing these organoids to various drugs and observing the reactions, they hope to better understand how individual patient tumors might respond to different therapies.
The platform integrates 3D bioprinting, advanced imaging, and AI, which aids in processing the extensive data generated. This capability allows for a thorough examination of how tumor segments react to drugs, which could prove invaluable since cancer can behave differently from one patient to another. While the technology is still being developed, the goal is to improve treatment personalization significantly.
Your Voice May Provide Health Insights
AI’s potential stretches beyond physical samples; researchers are also exploring voice analysis. A recent perspective in npj Digital Medicine focused on voice biomarkers for diseases like ALS and Parkinson’s, which can lead to measurable speech alterations.
The belief is that AI can analyze these changes, thereby helping to monitor disease evolution. For ALS, specific changes affecting speech and swallowing are of particular interest. The research is still nascent though, and while an ALS speech analytics platform has received FDA Breakthrough Device designation, further validation through the regulatory process is still needed.
Final Thoughts on AI in Healthcare
AI is becoming more integrated into healthcare settings, often without patients even realizing it. Hospitals might employ it when analyzing tissues. Fertility clinics might use it to discover what manual methods miss. However, as incredible as this advancement is, it’s crucial to consider the supporting evidence for each specific AI application.
Research initiatives differ considerably from approved medical devices, and it’s essential to recognize the extent of human oversight in the process. AI can aid healthcare professionals by offering insights and identifying patterns, but ultimately, personal medical judgments remain vital.
As AI becomes part of your healthcare, it can be helpful to ask a few questions to clarify its role.
Key Questions to Ask About AI in Your Healthcare
AI can certainly enhance medical practices, but understanding its implications on your care is essential.
1) What role does the AI play in my treatment?
It’s important to comprehend how the technology fits into your care. Is it analyzing data for your doctor? Is it bringing attention to something that needs further evaluation? “AI-powered” can mean many things, so a clear explanation is crucial.
2) Who verifies the results?
Determine whether a medical professional checks the AI’s findings prior to any decision-making, especially if the results could impact treatment or diagnosis.
3) What’s the regulatory status of the technology?
Inquire if the FDA has approved the AI technology and what research backs it. Initial research might suggest potential but could still leave significant questions unanswered.
4) How is my health data handled?
Given that medical AI often requires sensitive health information, it’s wise to understand how your data will be stored and who has access to it. Also, is there a possibility that your information could be used to enhance the AI system? More transparency around AI in healthcare is crucial.
Final Thoughts
It’s fascinating to see how AI can assist medical professionals in uncovering insights that may otherwise slip through the cracks. A technology capable of examining a million images for a single sperm cell or identifying potential new uses for existing drugs reflects remarkable advancements. Yet, amidst the enthusiasm for AI integration, it’s vital to remember the importance of waiting for substantial scientific evidence. With new therapy results like those from the melanoma trial, caution is still necessary.
If you were presented with a treatment recommended by AI that your physician hadn’t considered, how much data would you feel you require before trying it? Feel free to share your thoughts!
