Summary

A medRxiv preprint compared matched blood and bone-marrow cell-free DNA from 74 people with multiple myeloma. Peripheral-blood cfDNA matched or exceeded marrow-plasma cfDNA for several genomic signals, while fragment patterns varied by sample compartment and handling.

A matched-sample study in a medRxiv preprint found that peripheral-blood cell-free DNA (cfDNA) captured several multiple-myeloma genomic signals at least as well as cfDNA obtained from bone-marrow plasma. The result challenges the assumption that a liquid biopsy collected closer to a bone-marrow tumour must always provide a stronger molecular readout.

The analysis included 74 patients and 372 cfDNA sample time points. It also showed that the most informative sample compartment depends on the type of biomarker being measured, the amount of tumour-derived DNA present and how the sample is processed.

Contents

Why compare blood with marrow?

Multiple myeloma is a cancer of abnormal plasma cells that is predominantly located in the bone marrow. A liquid biopsy analyses DNA fragments released by cells into a body fluid instead of relying only on a tissue sample. In this setting, marrow plasma cfDNA is an attractive option because it may be physically closer to the disease than DNA circulating through the peripheral blood.

The researchers profiled matched marrow plasma cells, marrow-plasma cfDNA and peripheral-blood cfDNA from the same patients. This allowed them to compare the two cfDNA compartments against molecular features defined from the marrow rather than comparing unrelated patient groups.

The fraction of cfDNA originating from tumour cells, known as the tumour fraction, is important because a higher fraction generally makes tumour-specific alterations easier to detect. The study found that tumour fractions in the two cfDNA compartments were strongly correlated, with a correlation coefficient of r=0.84.

What the matched analysis found

Peripheral-blood cfDNA was not inferior across the main genomic measurements reported in the abstract. For marrow-defined copy-number alterations (CNAs), which are gains or losses of sections of the cancer genome, recovery was 77.4% in peripheral blood compared with 72.0% in marrow plasma. For mutations, recovery was 69.2% in peripheral blood compared with 65.4% in marrow plasma. The source reports high specificity for these findings.

Disease-associated features were detected in 69 of 74 peripheral-blood profiles and 68 of 74 marrow-plasma profiles. B-cell-receptor (BCR) clonotypes—the sequence signatures of related immune-cell clones—had the highest recovery among the features assessed. Bone-marrow-dominant clonotypes were recovered in 31 of 34 callable profiles in each compartment.

The difference between the compartments also narrowed when the tumour fraction was high. For adverse lesions, the reported sensitivity reached 100% in both cfDNA compartments at high tumour fraction. This suggests that the quantity of tumour-derived DNA can be as important as the location from which the sample is collected.

Why the signal depends on the feature

The researchers additionally identified a distinct fragmentomic architecture in marrow-plasma cfDNA. Fragmentomics examines patterns in cfDNA fragments, such as their distribution and other structural characteristics. The marrow-plasma pattern was described as sensitive to sample handling, and the analysis reported a leave-one-out cross-validation area under the curve (LOOCV AUC) of 0.910.

An AUC summarises how well a measured signal separates two analytical conditions across thresholds; a value of 0.910 indicates strong separation in the reported analysis. Here, the result is relevant to assay development because it indicates that fragment-based measurements can carry information about the sample compartment, but can also be affected by processing context.

Taken together, the findings support a more specific view of liquid biopsy in multiple myeloma. Peripheral blood may provide useful access to marrow-associated genomic features, while some fragmentomic signals may depend strongly on whether the DNA came from marrow plasma or circulating blood and on how the specimen was handled.

The evidence is analytical rather than a treatment trial: the study compares molecular signal recovery and sensitivity between matched specimens. It is presented as a medRxiv preprint, and prospective work would be needed to determine how these measurements perform in routine disease monitoring or treatment decisions.

Sources