How minimal residual disease is tracked in cancer research

After treatment, cancer can persist at levels that standard assessments cannot detect. Following cancer-associated markers offers a way to study that residual disease over time and relate changes in the measured signal to treatment response and recurrence. The challenge is to choose markers, samples and measurement methods that turn a scarce signal into useful evidence about the disease that remains.

On this page

• What is minimal residual disease?
• Samples for measuring residual disease: Tumor cells and cell-free DNA
• How minimal residual disease is measured
• What a positive or negative result can show
• Following residual disease over time in research
• QIAGEN products for residual-disease research

 

What is minimal residual disease?

Minimal residual disease (MRD) refers to cancer cells remaining after treatment that are no longer detectable by standard methods such as imaging, blood counts or morphological examination of bone marrow. (1–3) Blood-cancer guidelines now use the term measurable residual disease, and the term molecular residual disease is also used in studies of circulating tumor DNA (ctDNA). The terms describe the same idea from different fields. (1, 2, 4)

Being in remission by these standard assessments does not necessarily mean that every cancer cell has been eliminated. In leukemias such as acute myeloid leukemia (AML) and chronic lymphocytic leukemia (CLL), MRD assays can detect small populations of cancer cells or their molecular markers in blood and bone marrow that conventional assessments miss. These markers include abnormal patterns of proteins on cell surfaces and genetic changes in the cancer cells. (1, 2) A result is called positive or negative relative to the assay's threshold. (1)

MRD assessment is already used clinically in AML to help assess relapse risk after treatment. In patients in remission after standard chemotherapy, detecting MRD is associated with a higher risk of relapse. (1) Its clinical role varies with the cancer, treatment and assay. (1, 2)

Two questions guide MRD research: how reliably can an assay detect the remaining cancer, and what does that signal tell us about response to treatment or later recurrence? Assay development and evaluation address the measurement, while comparisons with patient outcomes test its interpretation. In solid tumors, for example, postoperative ctDNA studies investigate whether residual disease can be detected before recurrence appears on imaging. (4, 5)

Samples for measuring residual disease: Tumor cells and cell-free DNA

Residual disease can be studied through markers in cancer cells or in DNA released by cancer cells into plasma. Cell-containing blood and bone marrow samples are used to study residual disease in leukemias. (1, 2) Plasma ctDNA provides a way to investigate residual disease in solid tumors. (4, 5)

Tumor cells in bone marrow and blood

In CLL and AML, residual leukemia cells can be present in bone marrow and blood. (1, 2) Samples from these sites allow flow cytometry to identify leukemia cells among normal leukocytes by their abnormal marker patterns. Molecular assays can instead examine genetic material extracted from the sampled cells. (1, 2)

The sample determines which part of the disease is measured. CLL can involve blood, bone marrow and lymphoid tissues, and blood and bone marrow MRD results can differ after treatment. (2) In AML, molecular MRD assays can also use either blood or bone marrow, but sensitivity may be five- to 10-fold lower in blood. (1) The ease of collecting blood therefore has to be considered alongside what the chosen assay can detect in that sample.

Collection can also affect what a bone marrow sample reveals. In AML, dilution of a bone marrow aspirate with peripheral blood can lower the measured frequency of leukemia cells and increase the chance of a false-negative result. Using a small volume from the first aspirate draw helps limit that dilution. (6)

Solid tumors also shed circulating tumor cells (CTCs) into the bloodstream. (3)

Cell-free DNA in plasma and cerebrospinal fluid

Plasma contains cell-free DNA (cfDNA), short DNA fragments that mostly come from leukocytes. ctDNA is the part released by cancer cells; it can be a tiny fraction of the total, and how much a tumor releases varies by tumor type and stage. (7–9) Because a plasma sample contains no intact cancer cells, what it can detect is set by the number of tumor-derived DNA fragments in the blood draw, in solid tumors and blood cancers alike.

In primary brain tumors, plasma ctDNA is found in fewer than half of patients. (7, 9) Cerebrospinal fluid (CSF), which surrounds the brain and spinal cord, shows a higher ratio of ctDNA to cfDNA than plasma. In medulloblastoma research, copy number changes in CSF that matched the primary tumor have been used as an MRD marker. (7) However, collecting CSF usually requires a lumbar puncture, which is more burdensome for patients than a blood draw. (7)

Blood cancers also release ctDNA. (7) In CLL, disease can remain in lymph nodes, spleen and liver, compartments that blood and bone marrow cell measurements do not directly sample. This raises a research question: could cfDNA in plasma help detect residual disease in those compartments? (2)

Sample	Cancer examples	What carries the marker	What affects residual-disease detection?
Bone marrow cells	AML and CLL	Cancer cells and their DNA or RNA	Number of cells analyzed; in AML, dilution of the aspirate with blood (1, 2, 6)
Blood cells	AML and CLL	Circulating cancer cells and their DNA or RNA	In AML, molecular assays can be less sensitive than in bone marrow; in CLL, blood and bone marrow results can differ after treatment (1, 2)
Plasma cfDNA	Solid tumors; also explored in blood cancers	DNA fragments released by cancer cells	Number of tumor-derived fragments in the draw relative to background cfDNA (7)
CSF cfDNA	Brain tumors	DNA fragments released into CSF	Sample quality, including blood contamination during collection (7)

 

How minimal residual disease is measured

Choosing markers: Tumor-informed and tumor-agnostic approaches

To identify residual disease in cells or cell-free DNA, the measurement needs a marker that distinguishes cancer-associated material from the normal background. In tumor-informed work, markers are selected from a sample taken before treatment, ideally at diagnosis. These can be variants found by sequencing the tumor or diagnostic sample, an immunoglobulin gene rearrangement such as IGH in CLL, or a pattern of surface proteins identified by flow cytometry. (1, 2)

A variant found in a diagnostic sample does not necessarily identify residual cancer after treatment. In AML, mutations in the genes DNMT3A, TET2 and ASXL1 can be present in leukemia cells, but also in non-leukemic blood-forming cells. (1, 10) Such mutated blood-forming cells can expand with age, a process called clonal hematopoiesis, and persist during remission. Detecting one of these mutations after treatment therefore does not, by itself, establish that leukemia remains. (1) Inherited (germline) variants are also unsuitable because normal cells carry them too. (1)

The marker must also continue to represent the disease after treatment. In AML, some mutations are present only in a subclone, a subset of the leukemia cells. A negative result for one of these mutations can miss disease in cells that lack it, which is why such mutations are best followed alongside other MRD markers. Broader sequencing panels can also reveal variants not found at diagnosis. (1)

Tumor-agnostic assays use a predefined set of cancer-associated markers rather than markers selected from an individual tumor. Methylation-based ctDNA assays, for example, measure methylation patterns that recur in a cancer type. The same marker panel can therefore be applied without first profiling each person's tumor to select targets. (11)

Measuring cellular and molecular markers

Once the markers are known, the next step is to detect them among the many normal cells or DNA fragments in a sample. In AML and CLL, flow cytometry distinguishes leukemia cells from normal cells by the combinations of proteins on their surfaces, then counts the cells with the leukemia-associated pattern. (1, 2) In AML, for example, cells can show a myeloid marker such as CD13 together with CD7, a combination not expected in normal bone marrow. (6)

Molecular measurements look for genetic targets instead. These can include a tumor mutation in DNA extracted from bone marrow cells or in plasma cfDNA. (1, 4) RNA extracted from leukemia cells can also provide targets, such as transcripts from a mutated nucleophosmin 1 (NPM1) gene or from leukemia-associated fusion genes. (1) Molecular MRD methods include real-time quantitative PCR (qPCR), digital PCR (dPCR) and error-corrected sequencing with unique molecular identifiers. (1, 7)

For one or a few selected mutations, dPCR can detect and quantify rare mutant sequences among abundant wild-type DNA. Assays that distinguish the two sequences estimate their absolute copy numbers and the fraction carrying the mutation. (4) These readouts can be followed in repeated samples to study whether a marker persists, declines or becomes detectable again after treatment. (10) The counts are obtained by dividing the reaction into thousands of partitions and using the distribution of positive and negative partitions to estimate target copy numbers with Poisson statistics. (4)

To follow many variants in one analysis, targeted next-generation sequencing (NGS) can use panels designed around variants found in an individual tumor or predefined panels. (1, 7) Tracking several mutations by PCR or sequencing makes monitoring less dependent on any single marker: if one mutation is absent from part of the cancer, another may still identify DNA from that cell population. (1)

Examples of MRD measurement in bone marrow and plasma

In AML, Winters and colleagues used diagnostic sequencing to identify mutations, then followed them with mutation-specific dPCR assays in DNA extracted from bone marrow cells. (10)

In a three-patient gynecological cancer study, Balasan and colleagues selected mutations by sequencing tumor exomes, then measured those mutations in plasma cfDNA using QIAGEN dPCR LNA Mutation Assays and custom dPCR LNA assays on QIAcuity. Each reaction was divided into approximately 26,000 partitions in a QIAcuity Nanoplate well. Tumor mutations were detectable in postoperative plasma in two of the three patients. (4) The tumor profile supplied the targets; the later plasma measurements tested whether those targets remained detectable after treatment.

What a positive or negative result can show

In an MRD assay, a positive result means the selected marker was detected above the assay's reporting threshold in that sample. A negative result means the marker did not reach that threshold in the material tested. It does not establish that the disease has been eradicated, and relapse can still occur. (1, 2) In AML, an estimated 25–30% of patients who test MRD-negative by flow cytometry still relapse. (6)

When selecting MRD markers or evaluating an assay, a negative result needs to be considered in light of three limits:

• The amount tested. The detection limit depends on the assay and on the number of cells or amount of DNA analyzed. (2) With little DNA available, a mutation at low variant allele frequency may be represented by very few target molecules. In plasma, the volume collected can also limit how many mutations can be tested. (4)
• What reaches the sample. Disease can remain in the body without enough of its cells or DNA reaching the sample being analyzed. In plasma, a tumor that releases little DNA can produce a false negative result. (12) In AML, blood samples may be less sensitive than bone marrow. (1, 6) In brain tumors, little ctDNA may reach plasma, and CSF can be the richer sample. (7, 9)
• Whether the tracked marker remains in the cancer. The marker chosen for MRD measurement may not be present in the cancer that returns. In AML, relapses without the NPM1 mutation present at diagnosis have been reported. (1) An assay directed only at that mutation would not identify leukemia cells that lack it.

These limits are evident in the Balasan study: one patient had recurrence although both tracked mutations remained undetected in plasma. The study did not establish why those targets were undetected. (4)

A single negative MRD result therefore describes one sample measured with one assay at one time point.

Following residual disease over time in research

Repeated sampling for MRD analysis follows the measured marker level at several times after treatment. The signal may become undetectable, persist or become detectable again. Comparing these patterns with later relapse tests how well the measurements reflect the subsequent course of disease. (5, 10)

A change in measured marker level or detection status needs to be considered alongside any changes in the sample type, marker, assay or detection limit. Those differences can alter the readout, so a rise or fall in signal may reflect measurement conditions as well as a change in disease. Consistent sampling and measurement help distinguish these possibilities. (1, 2)

Winters and colleagues measured a median of five remission bone marrow samples per patient. In three of 29 MRD-negative patients, mutations became detectable again, and all three relapsed. (10) After surgery for early-stage lung cancer, Abbosh and colleagues detected ctDNA before or at clinical relapse in 13 of 14 patients who relapsed. Excluding one patient without a postoperative plasma sample before relapse, the median lead time over CT-confirmed relapse was 70 days. (5)

These observations link molecular measurements to later relapse, helping assess whether a marker reflects the course of the disease. Whether changing treatment in response to that signal improves outcomes is a separate research question; these studies do not establish that benefit.

QIAGEN products for residual-disease research

Counting known targets precisely is where QIAGEN products support residual-disease research. When tumor-derived molecules make up only a small part of the DNA tested, the measurement needs to distinguish the target from normal background and quantify it reliably.

For one or a few known mutations, dPCR LNA Mutation Assays with the QIAcuity Digital PCR System use separate probes for mutant and wild-type sequences and provide absolute quantification. This supports following a selected mutation within a much larger background of normal DNA. The Balasan study illustrates this use with commercial assays and custom assays matched to tumor mutations. (4)

For a broader examination of predefined regions in cfDNA, QIAseq Targeted cfDNA Ultra Panels enrich selected regions and attach unique molecular identifiers (UMIs) to DNA molecules before amplification. UMIs correct for PCR amplification bias and, together with high-fidelity chemistry, help identify rare variants with high confidence. What the assay can detect still depends on the amount and quality of the DNA tested. (2)

References

1. Heuser M, et al. 2021 Update on MRD in acute myeloid leukemia: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2021;138(26):2753–2767. doi.org/10.1182/blood.2021013626
2. Wierda WG, et al. Measurable residual disease in chronic lymphocytic leukemia: expert review and consensus recommendations. Leukemia. 2021;35(11):3059–3072. doi.org/10.1038/s41375-021-01241-1
3. Lin D, et al. Circulating tumor cells: biology and clinical significance. Signal Transduct Target Ther. 2021;6(1):404. doi.org/10.1038/s41392-021-00817-8
4. Balasan N, et al. Sensitive Detection of Gynecological Cancer Recurrence Using Circulating Tumor DNA and Digital PCR: A Comparative Study with Serum Biochemical Markers. Int J Mol Sci. 2024;25(22):11997. doi.org/10.3390/ijms252211997
5. Abbosh C, et al. Phylogenetic ctDNA analysis depicts early-stage lung cancer evolution. Nature. 2017;545(7655):446–451. doi.org/10.1038/nature22364
6. Tettero JM, et al. Technical Aspects of Flow Cytometry-based Measurable Residual Disease Quantification in Acute Myeloid Leukemia: Experience of the European LeukemiaNet MRD Working Party. Hemasphere. 2022;6(1):e676. doi.org/10.1097/hs9.0000000000000676
7. Tivey A, Church M, Rothwell D, Dive C, Cook N. Circulating tumour DNA - looking beyond the blood. Nat Rev Clin Oncol. 2022;19(9):600–612. doi.org/10.1038/s41571-022-00660-y
8. Diehl F, et al. Circulating mutant DNA to assess tumor dynamics. Nat Med. 2008;14(9):985–990. doi.org/10.1038/nm.1789
9. Bettegowda C, et al. Detection of circulating tumor DNA in early- and late-stage human malignancies. Sci Transl Med. 2014;6(224):224ra24. doi.org/10.1126/scitranslmed.3007094
10. Winters AC, et al. Multi-gene measurable residual disease assessed by digital polymerase chain reaction has clinical and biological utility in acute myeloid leukemia patients receiving venetoclax/azacitidine. Haematologica. 2024;109(6):1766–1778. doi.org/10.3324/haematol.2023.283790
11. Eriksen SV, Zedan AH, Kahns S, Osther PJS, Timm S, Hansen TF. Evaluation of a Methylated Circulating Tumour DNA Panel for Detection and Disease Stratification in Prostate Cancer. Int J Mol Sci. 2026;27(13):6081. doi.org/10.3390/ijms27136081
12. Husain H, et al. Tumor Fraction Correlates With Detection of Actionable Variants Across > 23,000 Circulating Tumor DNA Samples. JCO Precis Oncol. 2022;6:e2200261. doi.org/10.1200/po.22.00261