AI Tool Predicts Best Myeloma Treatment for Each Patient

Imagine if doctors could look at a simple test and know exactly which cancer treatment would work best for you, not just most patients.

Researchers at the University of Miami's Sylvester Comprehensive Cancer Center just made that vision real for people with multiple myeloma, a bone marrow cancer that affects over 35,000 Americans each year. Their new AI tool reads routine bone marrow slides and predicts which patients will respond to specific treatments.

The challenge has been huge. Multiple myeloma patients now have more treatment options than ever, from powerful immunotherapies to stem cell transplants. But these treatments come with serious side effects, and doctors have struggled to predict who needs the most intensive therapy and who can safely skip it.

Lead researcher Arjun Raj Rajanna and his team trained their AI model, called GigaTIME, to spot hidden immune signals in standard bone marrow biopsies. They tested it on 212 newly diagnosed patients and found something remarkable.

The AI could predict which patients would benefit from adding daratumumab, an immunotherapy drug, to their treatment. Patients with certain immune markers who received this combination stayed healthy much longer. At 18 months, 87% remained event-free compared to just 29% who received standard treatment alone.

The tool also identified patients who might safely avoid stem cell transplants, which can extend remission but temporarily weaken the immune system and increase infection risk. For patients with high levels of specific immune markers, outcomes were comparable with or without the transplant.

Dr. C. Ola Landgren, director of the Sylvester Myeloma Institute, emphasized the findings don't mean transplants are becoming obsolete. Instead, these decisions can now be driven by each patient's unique biology rather than a one-size-fits-all approach.

The Bright Side

This breakthrough represents something bigger than just cancer treatment. It shows AI moving beyond simple automation to genuine medical discovery, reading biological signals that human eyes miss in everyday tests doctors already perform.

The researchers presented their findings at the 2026 American Society of Clinical Oncology annual meeting, where the work generated excitement about AI-enabled digital pathology becoming standard practice. No expensive new tests or procedures required, just smarter use of what's already there.

For multiple myeloma patients, this could mean longer periods of disease control, fewer side effects, and less disruption to daily life. It transforms the question from "which drug works best overall?" to "which treatment fits this person's biology?"

The future of cancer care is getting personal, one AI-read slide at a time.