As in other fields, artificial intelligence (AI) and machine learning (ML) are changing the oncology and hematology practice.
In medicine, the terms AI and ML are used interchangeably, but they actually have distinct roles within computer science and technology. “Artificial intelligence encompasses a diverse range of technologies and methodologies dedicated to creating systems that can emulate humanlike intelligence and decision making processes,” Mohamed Elhadary and colleagues wrote in Blood Reviews. “On the other hand, machine learning is a specialized subset within the AI domain, concentrating on the development of models known as function approximators. These models are able to autonomously make informed decisions and draw conclusions by identifying patterns and extracting meaningful insights from raw data.”
One industry-sponsored study that harnessed AI and ML to expand the understanding of chronic lymphocytic leukemia (CLL) was presented in a poster at the 2024 ASCO Annual Meeting by Bessi Qorri, PhD, MSc, and her colleagues at NetraMark.
CLL, a lymphoproliferative disorder characterized by monoclonal B cell proliferation, is the most common adult leukemia in Western countries and accounts for 25% to 30% of leukemias in the U.S. Although the exact etiology of CLL is unknown, genetic factors play a role. CLL is rarely seen children, and in adults, the incidence increases with age.
Diagnosis entails blood tests that show increased lymphocytes, peripheral blood smears, and flow cytometry. In recent years, advances in AI and ML have generated various models and algorithms to support the diagnosis and classification of CLL. [https://www.sciencedirect.com/science/article/pii/S0268960X23000954]
Advanced Therapeutic Decision-Making
As the researchers wrote in their abstract, they applied the NetraAI proprietary AI and ML system to a CLL dataset and used it to refine the analysis and interpretation of the corresponding transcriptomic data (GSE39411).
NetraAI is a “ML system that provides an intuitive interface for scientists to interact with multimodal datasets to uncover connections” involving efficacy, toxicity, and placebo response in small datasets, according to the NetraMark website. The system generates hypotheses in the form of interactive representations of patient populations that show heterogeneity and statistically significant driving factors. An automated agent interacts with the representations, accounts for all variables, subpopulations of samples, and reports what the dataset has found.
Using this approach, the authors identified significantly lower levels of FADS3, GSDME, LPL, IMMT, NMB, and AEBP1 and higher expressions of COBLL1, P2RY1R, PDE8A, SYNE2, and FCRL3 as hallmarks of indolent CLL, suggesting that these markers could inform less aggressive disease management strategies.
In 104 CLL samples, researchers detected two distinct subpopulations within aggressive CLL that are delineated by unique genetic markers. One group, consisting of 31 aggressive CLL samples, is mainly characterized by lipoprotein lipase expression, while the other group, comprising 22/23 aggressive samples, is characterized by ZBTB20 and SYNE2.
“Our results . . . . underscore the heterogeneity within aggressive CLL . . . . and highlight the potential for these markers to guide prognostic assessments and therapeutic decisions,” the authors note. “This study advances our understanding of CLL heterogeneity, but also sets the stage for the development of more personalized and effective treatment modalities. By identifying drivers of disease aggression and indolence, we pave the way for targeted therapies that could significantly improve patient outcomes in CLL.”