Same Mutation, Different Fate

Two patients with acute myeloid leukemia (AML) achieve remission. Both test positive for measurable residual disease (MRD). Both show the same mutation at a similar variant allele frequency. By conventional metrics, they appear to carry the same risk. Six months later, one remains in remission. The other relapses. This divergence is not random. It reflects biology that current MRD tools are not equipped to measure.

The hidden assumption inside MRD

Most AML MRD testing approaches are built on an implicit assumption: residual cancer cells are interchangeable and that the presence of a mutation alone is sufficient to define relapse risk. In practice, MRD is treated as a quantitative problem. How many malignant cells remain? How deep is the remission?

But residual cells are not static remnants of disease. They are survivors. They persist because they have acquired the biological programs required to evade therapy, resist apoptosis, or remain dormant until conditions favor expansion. Counting cells tells us that disease may be present; it does not tell us whether those cells are capable of driving relapse.

This distinction is critical. Relapse is not determined by cell quantity alone, but by clonal identity, mutational cooperation, and cellular state. A small population of resistant leukemic progenitors can pose a far greater risk than a larger population of biologically inert cells.

Why current MRD tools leave clinicians guessing

In AML, MRD monitoring most commonly relies on bulk next-generation sequencing (NGS) and multiparameter flow cytometry. Both are analytically sensitive, but both leave important biological gaps.

Bulk NGS reports population-averaged variant allele frequencies. While this approach can detect mutations at very low levels, it obscures clonal architecture. It cannot determine whether multiple mutations coexist within the same malignant cell or are distributed across unrelated populations. Critically, it cannot distinguish a mutation arising from age-associated clonal hematopoiesis from the same mutation embedded within a relapse-competent leukemic clone.

Flow cytometry addresses a different dimension of disease by interrogating cell surface markers, but it is vulnerable to phenotypic drift. Under therapeutic pressure, leukemic blasts frequently alter antigen expression. This can blur the distinction between residual leukemic stem or progenitor cells and regenerating normal hematopoietic populations, particularly in the post-treatment setting.

When bulk sequencing and flow cytometry disagree, clinicians are left without a clear adjudicator. The result is a test report that identifies a residual signal but cannot reliably answer the clinical question that matters most: Does this patient require additional therapy, or are we at risk of overtreatment?

From detection to definition: What single-cell MRD changes

Single-cell multi-omic analysis addresses these limitations by linking genomic and immunophenotypic data at the level of individual cells. Rather than averaging signals across millions of cells, this approach asks a more precise question: Which mutations exist together in the same cell, and what biological state does that cell exhibit?

By simultaneously profiling somatic mutations and surface protein expression, single-cell MRD enables direct genotype–phenotype correlation. This allows for the definitive identification of residual malignant cells while filtering out biological noise, such as benign clonal hematopoiesis.

In AML, this distinction is not theoretical. Mutations such as DNMT3A or TET2 are frequently detected after treatment, yet they may persist in otherwise normal hematopoietic cells and carry little prognostic significance. Bulk assays often flag these mutations as MRD-positive, triggering anxiety and potentially unnecessary intervention. Single-cell resolution reveals whether these mutations reside within a leukemic clone that also harbors cooperating driver lesions, or within a biologically quiescent, non-malignant population.

Importantly, single-cell analysis can also identify rare, multi-mutant subclones that are invisible to bulk approaches yet central to clonal evolution and therapeutic resistance. These subclones often represent the true seeds of relapse.

Seeing relapse before it declares itself

Drug resistance does not emerge suddenly. It develops through the selection and expansion of small cellular populations that have already activated survival and resistance pathways. Conventional MRD assays often detect relapse only after these populations have expanded to detectable levels.

When applied at sufficient cellular depth, single-cell MRD analysis can detect resistant populations at very low frequencies. More importantly, it identifies them based on biological behavior, not just presence. This enables more precise risk stratification and creates a window for intervention before overt clinical relapse occurs.

This is more than earlier detection; it is about earlier understanding.

Three decisions this changes

A more biologically grounded MRD assessment reshapes several critical decisions in AML care.

First, when to intensify treatment. A patient in morphological remission may test MRD-positive by conventional assays. Single-cell analysis can reveal whether residual cells exhibit clonal architecture and phenotypic features associated with relapse, supporting timely therapeutic escalation.

Second, when to hold back. Another patient may show persistent mutations over time, yet single-cell profiling demonstrates that these mutations reside in biologically inert or non-leukemic populations. In this context, observation rather than additional chemotherapy may be the most appropriate course.

Third, who belongs in a clinical trial. Many trials stratify patients using binary MRD cutoffs that group individuals with fundamentally different disease biology. Defining patients by clonal composition and cellular state enables more precise enrollment, smaller trials, and clearer interpretation of therapeutic effect.

From sensitivity to significance

For more than two decades, MRD innovation has focused on sensitivity, pushing detection limits from 1% to one cell in a million. But sensitivity was never the primary barrier; the challenge has always been clinical significance.

Single-cell MRD does not replace existing approaches; it complements them. Counting residual cells remains important, but without biological context, it is insufficient. When MRD testing evolves from a numerical readout into a definition of clonal behavior, it becomes what clinicians have needed all along: a tool that informs action, not just detection.

The question is no longer whether we can measure residual disease. It is whether we are measuring the right biology.

Photo: ST.art, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.