AI tool accelerates leukemia diagnosis: What to know

A team of researchers from University of Florida Health has developed an open-access AI tool to accelerate the diagnosis of acute leukemias, according to a study published July 29 in Nature Communications. 

The Acute Leukemia Methylome Atlas, or ALMA, maps specific DNA tags called methylation patterns to determine specific leukemia subtypes. ALMA was trained on 3,300 leukemia samples and can match new samples to 27 leukemia subtypes as defined by the World Health Organization.

“Currently, clinicians often wait weeks for diagnostic lab results, but with our pipeline, this wait time can be reduced to 2-3 days for a preliminary diagnosis. This single-test assay, requiring only a laptop-sized sequencer, can be run in-house, lowering costs, widening access and improving long-term remission globally,” Jatinder Lamba, PhD, co-leader of the UF Health Cancer Center’s Cancer Targeting and Therapeutics research program and associate dean for research and graduate education, said in an Aug. 4 news release from UF Health.

Dr. Lamba led the development of ALMA and shared more about the tool’s potential with Becker’s.

Editor’s note: Responses have been lightly edited for clarity and length. 

Question: How do you envision a tool like ALMA changing the way hospitals and cancer centers approach diagnosing and treating acute leukemias?

Dr. Jatinder Lamba: ALMA has capability of utilizing both WGS and methylome data and return a unified report with WHO-2022 subtype call, epigenomic risk, and key variants/fusions within hours once data are uploaded. This not only can provide a preliminary diagnosis with short turnaround time but also has a potential to reduce dependency on multiple send-outs.

Further, the cloud model can be updated as new biomarkers emerge, so care benefits from a learning system rather than a static panel.

Q: What types of investments would hospitals and health systems need to make to integrate tools like ALMA into clinical workflows?

JL: In a simplistic way, a sequencing source that is capable of running WGS in-house or contracting a sequencing provider. Other tools needed will be: network and storage plumbing, secure upload to a HIPAA-configured cloud; plan for moving tens of GB per case and short-term storage pending report sign-out; SaaS tiering, which will vary by the volume of patients seen at a center; SOPs & validation. 

While initially labeled “Research Use Only,” sites should do parallel runs and local verification before using results to influence care, as well as select a pathologist and/or hematologist champion to own sign-out and IT support for smooth operations and legal compliance.

Q: In practice, what do you see as the most significant implications of ALMA for patient care and outcomes?

JL: Faster first decisions and the potential for cutting down decision time in a disease where days matter. 

Standardization and equity as community hospitals can deliver tertiary-center-grade genomic interpretation without building a bioinformatics shop, reducing referral delays and variability in care. 

Consolidating analysis into one SaaS pipeline can avoid serial send-outs and fragmented reporting, which will reduce cumulative cost and error opportunities.

The post AI tool accelerates leukemia diagnosis: What to know appeared first on Becker’s Hospital Review | Healthcare News & Analysis.