Artificial intelligence helps fight acute myeloid leukemia

Artificial intelligence helps fight acute myeloid leukemia

This article has been reviewed in accordance with Science

Fact check

trusted source


From left, PhD graduate Erin Kraugi, recent PhD graduate Mauricio Ferrato, and Professor Sunita Chandrasekaran are joined in focusing on effective drug treatments for patients with acute myeloid leukemia (AML), an aggressive blood cancer. Credit: Jeffrey C. Chase/University of Delaware

× Close

From left, PhD graduate Erin Kraugi, recent PhD graduate Mauricio Ferrato, and Professor Sunita Chandrasekaran are joined in focusing on effective drug treatments for patients with acute myeloid leukemia (AML), an aggressive blood cancer. Credit: Jeffrey C. Chase/University of Delaware

When Mauricio Ferrato earned his doctorate in computer and information science from the University of Delaware a few months ago, he made his mark in more ways than one.

Ferrato played a pivotal role in a research collaboration involving UD and Nemours Children’s Health that used artificial intelligence to find the most effective drug treatments for patients with acute myeloid leukemia (AML), an aggressive blood cancer.

The work was published earlier this year in the journal Advances in bioinformaticsis another step forward in the trend towards precision medicine, where treatment will be customized according to the patient’s unique genetic profile, with greater effectiveness and fewer negative effects.

According to the Leukemia and Lymphoma Society, about 20,000 new cases of chronic myeloid leukemia appear each year in the United States, and more than 11,000 people die annually from the disease.

It affects both children and adults and occurs when the body produces too many immature blood cells, called myeloblasts, that cannot develop into normal white blood cells.

These abnormal cells grow out of control and crowd out healthy cells in the bone marrow. From there, it can spread to the lymph nodes, brain and other organs, causing a wide range of symptoms, from fatigue and shortness of breath, to joint pain, recurring infections and weight loss.

Leukemia develops rapidly, so early diagnosis is crucial. The five-year survival rate for patients after diagnosis is 31.7%, according to the National Cancer Institute.

Using genetic data from 451 patients made available through the BeatAML Initiative, Ferrato used machine learning, a form of artificial intelligence, to help determine whether a person with acute myeloid leukemia would be a “high responder” or a “low responder” to any of the following: 100 different cases. Pharmaceutical treatments. The team was then able to “reverse engineer” the results and trace the pathways back to a specific gene in the patient and determine whether that gene creates a protein that affects cancer or a protein that resists a particular drug.

This foundational research may help lay the foundation for more promising outcomes for patients. The researchers also hope to explore the impact of their strategy on other types of cancer datasets and drug treatments.

Putting machine learning to the test

Machine learning runs on algorithms — sets of instructions that allow computers to make predictions and make decisions based on data — without being explicitly programmed to do so. These algorithms help identify patterns and relationships from vast amounts of data and create computer models of the results.

This area of ​​AI was crucial to the AML project, which was co-led by Ferrato’s doctoral advisor, Sunita Chandrasekaran, assistant professor of computer and information sciences at UD, and Erin Krugi, former director of medical bioinformatics at Nemours Children’s Health, who currently holds the position of Associate Director of Bioinformatics at Incyte, a biopharmaceutical company headquartered in Wilmington, Delaware. Adam Marsh, associate professor in UD’s School of Marine Science and Policy, also participated, along with colleagues from Emory University and UC San Diego.

While Ferrato brought a lot of machine learning skills to the project, he wasn’t always drawn to computer science.

Originally from Venezuela, Ferrato came to Delaware when his parents moved to the state when he was 12 years old.

“UD was the best choice for me, it allowed me to live close to my family, the research had a strong reputation and the campus was beautiful,” he said. “I really wanted to go into sports journalism when I started, but I ended up working with Sunita as an undergrad, mostly in high-performance computing.”

Chandrasekaran had a major influence on Virato, and he stayed there to earn his master’s and doctoral degrees in UD’s Department of Computer and Information Science.

Ferrato became involved with the AML project when Krugi, a PhD graduate in bioinformatics, was working at Nemours Children’s Health and had received funding from the Lisa Dean Moseley Foundation to pursue research on pediatric patients with the disease.

“We obtained grant funding to bring on a doctoral student, and Mauricio was a perfect fit,” Krugi said. “Our goal was to answer the question: Can you predict before treatment a person’s response to a particular drug?”

Krugi compared the work to having a lot of marbles in a jar and knowing which marble is the most important.

“This is what feature selection in machine learning is all about,” she said. “Once you find that marble, it might be large or rectangular. How will it roll out? It’s a way to take a lot of data that a person can’t easily interpret and create an algorithm to extract what is meaningful from 20,000 genes in the genome, in this case, and show how the person will respond.” Patients with acute myelogenous leukemia for treatment.

Ferrato used SHAP (short for SHapley Additive exPlanations), a tool used in game theory, to map a particular feature back to its biological equivalent. So SHAP would select the top 30 traits, each representing a gene, and then the pathway analysis would show what that gene was affecting, such as creating a protein that resists an anti-cancer drug.

He spent many hours writing computer code in Python and running models on UD’s DARWIN high-performance computer.

“We looked at six different models for 100 different drugs, and then we had to run the models several times to validate them, checking to see if the results were consistent. We had to run 3,000 to 4,000 models to get the results, with each model taking about hour to run.

The promise of artificial intelligence

As a computer scientist, Ferrato said he doesn’t know all the biological terminology associated with the project, such as transcriptomes, gene expression, RNA, and the AML background.

Krugi guided him. In return, he helped her understand machine learning better.

Shortly after earning his PhD, Ferrato began working for NVIDIA as a solutions architect. He interned there during his graduate studies, using his computer science skills to find the perfect way to place wind turbines on a wind farm to generate as much energy as possible.

“I like work that is applied to a real-life problem that helps humanity in some way,” he said.

By working together, researchers from multiple disciplines can solve large-scale problems using AI that they could not address before. Krugi said team science is key.

When it comes to future applications of AI, the sky’s the limit.

“We need to be really smart about how we develop and implement these applications,” Krugi said. “We all use AI every day, but we don’t think about it that way. Your mobile phone has all kinds of cool AI. This work for patients with acute myeloid leukemia is powerful and impactful.”

Chandrasekaran is also a strong advocate of interdisciplinary problem solving, working with industrial and academic partners. It’s the hallmark of the new Artificial Intelligence Center of Excellence (AICoE), which she now co-directs at UD.

“Working with our collaborators, Mauricio and I have learned a lot about the impact machine learning can have in precision medicine,” Chandrasekaran said. “The results have been remarkable.”

“The tremendous growth we are seeing in generative AI tools underscores the need to ensure our next-generation workforce is prepared to use these tools,” she noted. “To this end, our AI Center of Excellence at UD, which works with researchers across disciplines to provide AI solutions, recently launched the Graduate Certificate in AI. It is open to UD students, as well as professionals outside of UD.”

Now that the results of the anti-money laundering research have been published, what happens next?

“This work lays the foundation, infrastructure and technology for the future,” Krugi said. “It will take society as a whole, bringing together academia, hospitals and the biopharmaceutical industry, to move precision medicine forward.”

more information:
Mauricio H. Ferrato et al., Machine learning classification methods for predicting response to RTK-type-III inhibitors show high accuracy using transcriptional signatures and ex vivo data, Advances in bioinformatics (2023). doi: 10.1093/bioadv/vbad034

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *