Genomewide Copy Number Alteration Screening of Circulating Plasma DNA

Potential for the Detection of Incipient Tumors

L. Lenaerts; P. Vandenberghe; N. Brison; H. Che; M. Neofytou; M. Verheecke; L. Leemans; C. Maggen; B. Dewaele; L. Dehaspe; S. Vanderschueren; D. Dierickx; V. Vandecaveye; F. Amant; J. R. Vermeesch


Ann Oncol. 2019;30(1):85-95. 

In This Article


We here report unique data on the potential of genomewide CNA detection in plasma cfDNA for cancer screening in a large cohort of individuals without a cancer history. Five out of the 1002 screened participants were diagnosed with a hematological cancer (3 NHL, 1 HL and 1 MDS-EB1), with the CNAs in cfDNA originating, at least partially, from malignant cells. This frequency of cancer detection (0.5%) is three to four times lower than the expected prevalence of all cancers together, but is approximately four times higher than reported incidence rates for NHL, HL and MDS combined, in people aged 65 years and above.[13–15] Importantly, for two of our participants, the malignancy was not yet reflected in abnormal blood parameters, pointing to the potential of GIPseq for presymptomatic detection of these hematological cancers. Whereas current incidence numbers are based on the usual clinical presentations and conventional diagnostic criteria for hematologic malignancies, genetic screening approaches for clonal DNA mosaicisms in blood may result in a higher detection rate: especially in indolent disorders, the observation of clonality can lead to diagnosis of a malignancy before clinical signs or symptoms bring it to attention. The preponderance of hematological diagnoses might be plausible when assuming that cfDNA is largely derived from hematopoietic cells.[1] Three of the four incidental cancer diagnoses, following a normal GIPseq profile, were solid tumors and the genomes of these cancers were characterized by CNAs. Either these tumors did not shed DNA into the circulation or the load of ctDNA species was too low to be detected via GIPseq. Also, plasma cfDNA concentrations did not differ between participants with and without CNAs in cfDNA, suggesting that, even with a higher turnover of malignant cells, the ctDNA fraction remains limited.

Besides detecting hematological malignancies, GIPseq profiling might allow detecting premalignant conditions. Firstly, one high-count MBL diagnosis was made, being recognized as a precursor lesion of CLL.[16] Secondly, in 9 of the 24 cases (38%) without a cancer diagnosis, the isolated anomalies in cfDNA were originating from a low-grade mosaicism in peripheral blood cells, some of these aberrations being typically associated with MDS and myeloproliferative neoplasms.[17,18] Two of our cases had, respectively, a mild anemia and monocytosis, potentially pointing to an underlying, developing myeloid neoplasm. Previous studies on cellular DNA from peripheral blood found that clonal mosaicisms for these chromosomal anomalies in blood cell DNA are present in ~2.5% of aged people and, importantly, that they predict a 5- to 10-fold increased risk of a subsequent cancer.[19,20]

For about two-third of study participants (15 of 24 cases), with anomalies in cfDNA but no cancer diagnosis, the chromosomal aberrations were not detectable in peripheral blood DNA. This does not necessarily preclude that the observed abnormalities were from hematopoietic origin. It might be possible that aberrant cfDNA species originated from neoplastic clones in bone marrow without resulting in aberrant peripheral blood cell descendants. Such an observation was previously reported by Yeh et al. in the context of MDS.[21] Similarly, the cases identified with MDS and MALT in our study had a normal molecular karyotype in peripheral blood cells. Molecular testing of bone marrow biopsy DNA in cases with detectable anomalies in cfDNA but not in peripheral blood could help identifying the origin of aberrant cfDNA species. However, this is not ethically justifiable in otherwise healthy individuals. Nevertheless, there remains the possibility that the cfDNA clones might be derived from other than hematopoietic cells and reflect other non-neoplastic conditions.[22] Future analyses, exploiting supramolecular information contained in cfDNA, may help tracing the tissue from which the detected aberrant cfDNA species originated.[23]

To allow exploiting the full potential of GIPseq for cancer screening, a longer follow-up of our study population should be carried out to determine the clinical and biological significance of chromosomal imbalances in cfDNA of apparently healthy individuals. Either these chromosomal events represent malignant precursor clones or they reflect the normal clonal variation in aged people. Both scenarios impact future cancer screening programs searching for cancer-specific CNAs in liquid biopsies. Furthermore, long-term follow-up will allow determining the false negative rate in our population. Finally, further investigations are necessary to understand why, in about one-third of cases, the initial deviating GIPseq profile could not be reproduced upon subsequent plasma sampling, and whether this irregularity has a technical and/or a biological nature.

Together, these data underscore the potential of genomewide cfDNA profiling, as an unbiased screening approach for chronic hematological malignancies and premalignant conditions, such as HL, NHL, MDS or MBL. Such a secondary prevention strategy could be particularly valuable for patient groups at risk, e.g. after genotoxic cancer therapy. As shown above, plasma cfDNA analysis might allow detecting a broader range of chromosomal abnormalities than do examinations restricted to peripheral blood cells. Identifying individuals with an indolent hematological malignancy or a clonal mosaicism at risk of subsequent malignant evolution would allow a close follow-up and the introduction of specific measures for the management of the early-stage disease.[24] Finally, potential consequences for the patient's mental health need to be taken into account: the discovery of clonal mosaicism in cfDNA and/or an indolent hematological malignancy together with uncertainty about its future evolution may cause anxiety, akin to that described in patients diagnosed with CLL that were approached with a strategy of active surveillance and 'watchful waiting'.[25,26] Lastly, this study also provides a baseline of CNAs expected to be present in plasma cfDNA.