Multiple myeloma begins in plasma cells, the cells in bone marrow that normally make antibodies. When those cells become cancerous, they multiply, push out healthy blood cells, thin the bones and can harm the kidneys. Many people live with the disease for years on modern treatment. Getting to a diagnosis, though, often takes far longer than it should, and most AI research in this area targets that gap.
For families in the Gulf, specialist care no longer means flying abroad. Anyone looking into multiple myeloma treatment UAE choices will come across Cleveland Clinic Abu Dhabi, which says its Multiple Myeloma and Amyloidosis Program is the only center in the country dedicated to these related conditions. Below is a look at what AI is and isn’t doing for myeloma patients today, plus some questions to take to your next appointment.
Why doctors often diagnose myeloma late
The early signs are easy to explain away. People put a sore back down to age or a bad mattress, and they blame tiredness on a busy month. Repeated chest infections or a dip in kidney function each have plenty of likelier causes.
Researchers who reviewed the problem in 2025 found the same pattern: symptoms are usually vague, the disease can sit quietly for a long time before it causes trouble, and doctors often don’t think of it. They also pointed to hard-to-access diagnostic tests, uneven follow-up for people with precursor conditions, and departments that don’t share information well.
Some of the warning signs may already be sitting in a patient’s file. One study cited primary-care records showing that, before their diagnosis, 83% of myeloma patients had a kidney-function test, 80% a full blood count, and 52% a calcium test. In many cases, doctors had already ordered the right blood tests. Nobody joined the dots.
How researchers are using AI, and how far along it is
Almost all of this work is still in research, and patients should know that up front. Here’s a rough map:
| Area | What the AI is meant to do | How mature it is |
|---|---|---|
| Routine blood tests | Spot results that hint at undiagnosed myeloma | Mostly retrospective, single-hospital studies |
| Protein electrophoresis | Help interpret a core diagnostic lab test | Early but growing |
| CT, MRI and PET/CT scans | Find and measure bone lesions more consistently | Active research, some automated tools |
| Prognosis and treatment | Estimate relapse risk and likely drug response | Exploratory |
Reading patterns in everyday blood tests
This is arguably the idea with the most practical promise. A team at Shengjing Hospital in China built several machine-learning models from 4,187 routine blood and biochemistry records the hospital collected between 2010 and 2020, of which 1,741 came from people with myeloma. With carefully chosen inputs, the models did well at telling myeloma apart from look-alike conditions such as kidney, liver and rheumatic disease.
A more recent study applied the same idea to patients whose doctors suspected myeloma and sent them to a tertiary hospital. Its authors noted that blood counts, biochemistry and inflammatory markers are available almost everywhere, while specialist myeloma tests aren’t. If a tool like this works in practice, it could nudge a family doctor to order the right follow-up test months earlier.
There’s an important catch. Researchers have only checked most of these models against data from the hospital that built them. Nobody yet knows how well they perform on other populations, and that includes patients in the Gulf.
Making scan reading more consistent
Scans carry a lot of weight in myeloma, since 70 to 80% of patients already have lytic bone lesions when doctors diagnose them. PET/CT works very well for tracking how treatment is going, but the marrow patterns it shows can be complicated. Readings can therefore vary from one radiologist to the next.
Researchers are developing AI tools for CT, PET/CT and MRI partly to reduce that variation. The aim is less about replacing the radiologist and more about getting the same measurement from the same scan, every time.
Helping with lab interpretation
Serum protein electrophoresis and immunofixation sit at the center of a myeloma diagnosis, and reading them well takes training. A 2026 scoping review gathered machine-learning studies from 2015 to 2025 that try to automate parts of this, with the stated goal of helping non-specialist labs catch myeloma sooner.
Predicting relapse
Myeloma often returns, sometimes in a form that no longer responds to drugs that worked before. Reviewers think machine learning could eventually help find new markers and better predict relapse and survival. For now this is the least developed area, and no algorithm is picking anyone’s treatment.
What hasn’t changed
Specialists still make the treatment decisions. The big gains in myeloma survival over the past two decades came from new drugs, not software. A typical modern plan draws on some mix of proteasome inhibitors, immunomodulatory drugs, monoclonal antibodies and other immunotherapies. For patients who are fit enough, it may also include an autologous stem cell transplant. Doctors collect the patient’s own stem cells, give high-dose chemotherapy, and then return the cells so the marrow can recover.
Myeloma also overlaps with AL amyloidosis, a condition in which misfolded proteins collect in organs such as the heart, kidneys and nerves. Treating the two together needs doctors who know both well, which is a big part of why dedicated programs exist. AI doesn’t do that job.
Questions to take to your care team
It’s fine to ask about technology at an appointment, and more and more patients do. Some questions worth raising:
- What stage and subtype is my myeloma, and which tests showed that?
- Why do you recommend this particular treatment for me?
- Which blood markers will you track, and how often?
- Do you use any AI or computer-aided tools on my scans or lab results? Who checks what they produce?
- Am I eligible for any clinical trials?
- Which side effects should I report right away?
It also helps to bring every record you have to the first visit, including biopsy reports, scans and lab results. A specialist with the full history can plan sooner.
Recovery, and the family’s part in it
Recovery after intensive treatment, particularly a transplant, can take months. Fatigue, a higher risk of infection, and swings in appetite and mood are all common. Follow-up usually means regular blood checks and, for many people, maintenance therapy.
Relatives often end up carrying a lot of this: keeping track of appointments, dealing with the pharmacy, noticing when something seems off. Some of that coordination is getting easier. AI agents are changing how families coordinate care for aging parents, and similar reminder and scheduling tools can help with the daily routine of supporting someone through myeloma treatment.
FAQs
Q. Can AI diagnose multiple myeloma?
No, not by itself. Research tools can flag blood-test or scan patterns that point toward myeloma, but a diagnosis still needs specialist tests and a hematologist.
Q. Do UAE hospitals use AI in myeloma care?
It differs from hospital to hospital and is changing quickly. Your care team can tell you whether they use any AI-assisted tools to review your scans or lab work.
Q. Is AI going to replace oncologists?No. Current myeloma AI handles narrow jobs like measuring
lesions or screening blood results, and clinicians review what it produces.
Related: Orthodontic Technology in Australia: What AI Can—and Can’t—Predict
This article is for general information and isn’t medical advice. Please talk to a qualified hematologist or oncologist about your own diagnosis and treatment.
