Dr Ip on the Implications of NGS Plus Machine Learning for Diagnosing Hematologic Malignancies


In Partnership With:

Andrew Ip, MD, discusses the implications of using next-generation sequencing alongside machine learning in the diagnosis of hematologic malignancies.

Andrew Ip, MD, medical oncologist, John Theurer Cancer Center, Hackensack University Medical Center, discusses findings from a study assessing the feasibility of replacing flow cytometry analysis with next-generation sequencing (NGS) alongside machine learning to detect and diagnose hematologic malignancies, highlighting the clinical implications of this research.

At the 2023 ASH Annual Meeting, results from the study showed that the use of NGS to quantify RNA levels from 30 CD markers in combination with a machine learning algorithm may serve as a reliable and accurate screening method for differential diagnosis across various hematologic conditions. The study reported an area under the curve (AUC) consistently exceeding 90%, indicating a strong true positive rate with a 2% chance of false positivity. Notably, the data from this study were obtained from peripheral blood test samples.

The utilization of liquid biopsies has gained traction in recent years due to its ability to provide comprehensive information about cancer without the need for invasive tissue biopsies, Ip begins. This approach has been promising in hematologic malignancies, where NGS technologies can detect specific genetic alterations associated with various blood cancers, he notes.

Findings from this study build upon prior research, including a study published in the New England Journal of Medicine, which explored the feasibility of detecting acute myeloid leukemia (AML) using whole-exome sequencing, Ip states. The study found that whole-genome sequencing provided rapid and accurate genomic profiling in patients with AML or myelodysplastic syndrome, he reports.

Based on this study, Ip and his colleagues at the Genomics Testing Cooperative used a more targeted NGS panel focused on RNA and DNA mutations to assess the utility of NGS-based liquid biopsies across different hematologic conditions beyond AML, Ip says.

This research suggests that liquid biopsies using NGS and machine learning could improve the diagnosis and monitoring of hematologic neoplasms, Ip explains. By providing potentially more accurate, powerful, and automated methods of disease detection, this approach could enhance patient care, facilitate earlier interventions, and potentially reduce the need for more invasive diagnostic procedures like bone marrow biopsies or lymph node biopsies, Ip expands. Further research and validation will be needed to potentially integrate these technologies into routine clinical practice and optimize their utility across different hematologic cancers, Ip concludes.

Related Videos
Mike Lattanzi, MD, medical oncologist, Texas Oncology
Vikram M. Narayan, MD, assistant professor, Department of Urology, Emory University School of Medicine, Winship Cancer Institute; director, Urologic Oncology, Grady Memorial Hospital
Stephen V. Liu, MD
S. Vincent Rajkumar, MD
Pashtoon Murtaza Kasi, MD, MS
Naseema Gangat, MBBS
Samilia Obeng-Gyasi, MD, MPH,
Kian-Huat Lim, MD, PhD
Saurabh Dahiya, MD, FACP, associate professor, medicine (blood and marrow transplantation and cellular therapy), Stanford University School of Medicine, clinical director, Cancer Cell Therapy, Stanford BMT and Cell Therapy Division
Muhamed Baljevic, MD