The right testing can make a significant difference in AML treatment, personalizing care for patients. Dr. Sangeetha Venugopal shares an overview of the essential tests performed following an AML diagnosis, how molecular testing helps to determine treatment options, and why repeat testing may be needed as the disease changes over time.
Dr. Sangeetha Venugopal is Assistant Professor in the Department of Internal Medicine and Division of Hematology at the University of Miami. Learn more about Dr. Venugopal.
Related Resources
Transcript
Katherine Banwell:
Dr. Venugopal, how is artificial intelligence beginning to influence AML diagnosis and treatment planning?
Dr. Sangeetha Venugopal:
I think this is a fantastic question. We are still trying to find our way where it fits in the paradigm of diagnosis and treatment.
AI, in particular, I would say the strongest near-term influence is on diagnosis speed, risk stratification, and treatment planning—not fully autonomous decision-making. AI is beginning to improve how AML is diagnosed and subclassified from routine clinical materials. One of the things that excited me in the last year was faster molecularly informed diagnosis. Investigators at Dana-Farber developed this system called the MARLIN system, and this uses DNA methylation with machine learning to classify leukemia subtypes in under two hours from the biopsy received in real-time testing, compared with days to weeks for conventional combined molecular and cytogenetic workflows.
I do think this is important because this will definitely shorten the time between diagnosis and the treatment plan, especially when subtype or hidden genomic features would otherwise take longer to confirm.
One of the other things about AI is better risk stratification. Machine learning models combining clinical, cytogenetic, and molecular data can definitely outperform conventional guideline-based risk systems in prognostic accuracy. As for the treatment selection. This is an emerging field, especially because the treatment response is biologically heterogeneous. There is some data about SMART trials, which can return ex vivo drug response reports within seven days for most participants and improve risk stratification. This is particularly important because this will lead to more individualized therapy rather than just choosing one regimen that fits all. The next role, which is emerging, would be to see transplant and intensive treatment planning.
AI models can help predict outcomes, such as mortality after transplant, and to refine who may need alternative collection or mobilization strategies. I would say AML is a disease where timing definitely matters, molecular heterogeneity matters, and relapse biology can change very quickly. That makes it a particularly good fit for AI in a few ways. AI can compress turnaround time from raw pathology and genomics into a usable classification. It can surface mutation-linked treatment implications. It can definitely update risk dynamically as new labs, response data, or residual disease information come in. However, it’s not going to replace hematopathologists, molecular diagnostics, or leukemia specialists in AML care. Definitely, the current role is augmented decision support.
Katherine Banwell:
Yeah, that’s a good point. There are still going to be humans taking care of other humans.
Dr. Sangeetha Venugopal:
For sure. Yeah.
Katherine Banwell:
Dr. Venugopal, is there anything you’d like to add about the evolution of AML care and treatment? What are you excited about?
Dr. Sangeetha Venugopal:
I’m actually excited about how we are doing so many novel things. For example, I talked about the Marlin system. I want to know how this will change current clinical practice because this is still a system in testing. I would want to see broad-scale application of this particular system in diagnosis. Definitely, I, as a leukemia researcher and a treating physician, would want to know something about my patient’s leukemia genetic makeup immediately rather than waiting for about a week. This will help me in treatment planning with regard to clinical care and how I’m going to approach this patient in particular.
That’s something I’m excited about. Again, the next one I’m excited about is the menin inhibitor combination therapies because, right now, as a menin inhibitor monotherapy, the duration of response is not spectacular.
Nevertheless, we have another medication to treat our patients with relapsed leukemia. I’m quite certain that combination therapy will definitely improve the outcomes of patients with AML whose leukemia is sensitive to menin inhibition. And all-oral therapy. This is actually a treatment-changing paradigm. The first treatment-changing paradigm was treating older adults with a hypomethylating agent and venetoclax (Venclexta) combination, which happened six years ago. Six years later, now in 2026, we have an all-oral therapy, which is oral decitabine and venetoclax, to treat our patients with AML who are not eligible to receive intensive chemotherapy.
This means that your patient will spend less time in the clinic and probably have less financial toxicity by driving back and forth and all that stuff.