- Research article
- Open Access
Abnormal plasma DNA profiles in early ovarian cancer using a non-invasive prenatal testing platform: implications for cancer screening
BMC Medicine volume 14, Article number: 126 (2016)
Non-invasive prenatal testing (NIPT) identifies fetal aneuploidy by sequencing cell-free DNA in the maternal plasma. Pre-symptomatic maternal malignancies have been incidentally detected during NIPT based on abnormal genomic profiles. This low coverage sequencing approach could have potential for ovarian cancer screening in the non-pregnant population. Our objective was to investigate whether plasma DNA sequencing with a clinical whole genome NIPT platform can detect early- and late-stage high-grade serous ovarian carcinomas (HGSOC).
This is a case control study of prospectively-collected biobank samples comprising preoperative plasma from 32 women with HGSOC (16 ‘early cancer’ (FIGO I–II) and 16 ‘advanced cancer’ (FIGO III–IV)) and 32 benign controls. Plasma DNA from cases and controls were sequenced using a commercial NIPT platform and chromosome dosage measured.
Sequencing data were blindly analyzed with two methods: (1) Subchromosomal changes were called using an open source algorithm WISECONDOR (WIthin-SamplE COpy Number aberration DetectOR). Genomic gains or losses ≥ 15 Mb were prespecified as “screen positive” calls, and mapped to recurrent copy number variations reported in an ovarian cancer genome atlas. (2) Selected whole chromosome gains or losses were reported using the routine NIPT pipeline for fetal aneuploidy.
We detected 13/32 cancer cases using the subchromosomal analysis (sensitivity 40.6 %, 95 % CI, 23.7–59.4 %), including 6/16 early and 7/16 advanced HGSOC cases. Two of 32 benign controls had subchromosomal gains ≥ 15 Mb (specificity 93.8 %, 95 % CI, 79.2–99.2 %). Twelve of the 13 true positive cancer cases exhibited specific recurrent changes reported in HGSOC tumors. The NIPT pipeline resulted in one “monosomy 18” call from the cancer group, and two “monosomy X” calls in the controls.
Low coverage plasma DNA sequencing used for prenatal testing detected 40.6 % of all HGSOC, including 38 % of early stage cases. Our findings demonstrate the potential of a high throughput sequencing platform to screen for early HGSOC in plasma based on characteristic multiple segmental chromosome gains and losses. The performance of this approach may be further improved by refining bioinformatics algorithms and targeting selected cancer copy number variations.
The detection and monitoring of specific cancer mutations by sequencing circulating DNA holds much promise, but has yet to be widely translated into clinical care. In contrast, sequencing plasma DNA during pregnancy to detect fetal chromosomal abnormalities (non-invasive prenatal testing, NIPT) has been rapidly implemented globally due to its high accuracy and proven clinical validity .
Circulating DNA of tumor origin can interfere with NIPT performance and produce abnormal genomic profiles that suggest occult malignancy in pregnant women . Amant et al.  recently reported the pre-symptomatic identification of cancer in three pregnant women undergoing NIPT, suggesting that genomic profiling for copy number variations (CNVs) may be a feasible approach for cancer screening. However, the sensitivity and specificity of clinical NIPT platforms for cancer remains unknown.
Ovarian cancer is the leading cause of gynecologic cancer-related deaths in developed countries  and there is a pressing need for an effective screening test [5, 6]. High-grade serous ovarian cancer (HGSOC) accounts for most deaths from the disease  and demonstrates a marked chromosomal instability . We hypothesized that these tumor-derived chromosome abnormalities would be detectable in the plasma of HGSOC patients collected prior to primary surgery. The aims of this study were to investigate whether a clinical NIPT platform could detect HGSOC in the non-pregnant population based on an abnormal plasma DNA profile, and to compare the detection rates for early and advanced stage HGSOC.
We performed a case control study of 64 plasma samples obtained from the Western Australia Gynecologic Oncology Biospecimen Bank. These were prospectively collected between January 2013 and August 2015 with informed consent from patients prior to undergoing surgery. Ethical approval was granted for this study.
The 32 cancer cases comprised 16 women with International Federation of Gynecology and Obstetrics (FIGO) stage I and II HGSOC (‘early cancer’), and 16 women with FIGO stages III and IV HGSOC (‘advanced cancer’). The control group included women with benign gynecologic disease undergoing surgery (n = 24), or germline BRCA1 and BRCA2 mutation carriers without evidence of malignancy who were undergoing risk-reduction surgery (n = 8).
DNA libraries, prepared from cell-free DNA extracted from plasma, were sequenced on a commercial whole genome NIPT platform using the standard workflow employed for aneuploidy screening (percept™ prenatal test, Victorian Clinical Genetics Services, Parkville VIC Australia, based on Illumina’s verifi™ NIPT methodology ). Each research sample was sequenced alongside 14 clinical samples, with 36-cycle single-end sequencing on an Illumina NextSeq500. The read depth was low coverage at 0.2× to 0.3× based on 18–28 M × 36 bp single end reads. Laboratory and analysis staff were blinded to the case/control allocation of samples. Two types of data analyses were performed.
We used the open source algorithm WISECONDOR (WIthin-SamplE COpy Number aberration DetectOR) to detect whole chromosome and subchromosomal abnormalities not identifiable by the standard NIPT pipeline . Segmental changes > 15 Mb were prespecified as abnormal calls (“positive cancer screen”).
We also analyzed the sequence data using the routine clinical percept™ pipeline, developed to detect fetal aneuploidy for chromosomes 21, 18, 13, X, and Y.
Paired tumor DNA was unavailable to correlate with plasma sequencing data. We therefore compared the results of the WISECONDOR analysis with somatic CNVs reported in the Integrated Genomic Analyses of Ovarian Carcinoma (IGAOC) derived from 489 HGSOC tumor genomes by The Cancer Genome Atlas Research Network . Our data were examined for recurrent regional aberrations affecting extended chromosome regions that were reported as statistically significant by the IGAOC (8 gains and 22 losses).
We detected 6/16 early stage and 7/16 advanced stage HGSOC cases using the WISECONDOR analysis, giving an overall detection rate of 13/32 (sensitivity 40.6 %, 95 % CI, 23.7–59.4 %). There were two false positive calls in the control group (specificity 93.8 %, 95 % CI, 79.2–99.2 %) (Table 1).
Table 2 presents the specific CNVs detected in the 13 true positive cancer cases and the two false positive controls. Twelve of the 13 true positive cancer calls had a CNV that was reported in The Cancer Genome Atlas Network as statistically significant (FDR q value < 0.25) at high frequency (>50 % of tumors). The most common DNA amplifications observed in the 13 true positive calls affected chromosome arms 3q (n = 5), 8q (n = 7), 20q (n = 4), and 12p (n = 3). The most common DNA losses were seen on chromosome arms 5q (n = 3), 8p (n = 3), 13q (n = 4), and 15q (n = 3). Figure 1 shows the WISECONDOR plots of sequenced cfDNA showing copy number variations of chromosome 3 in the plasma of five subjects with high-grade serous ovarian carcinomas.
The percept™ pipeline resulted in one “monosomy 18” call from the cancer group, and two “monosomy X” calls in the controls (Table 2). In five cancer cases and one control case, the pipeline failed to produce a result because of unexpected profiles on normalizing chromosomes.
A post hoc analysis of our results showed that many smaller focal aberrations identified by the IGAOC were also present in the “screen positive” cancer cases. Most of the cancer cases had multiple focal changes, whereas none of the benign controls, including the two false positive calls, had more than one focal change (Additional file 1).
The two false positives in the control groups in the WISECONDOR analysis had single segmental gains on 20q. The clinical history of these controls included a benign fallopian tube cyst in a patient with endometriosis and a hemorrhagic follicular cyst in a patient with a prior history of breast ductal carcinoma in situ which had been completely excised prior to plasma collection. Both patients were alive with no clinical evidence of malignant or systemic disease at the time of writing.
In this proof of concept study, low coverage plasma DNA sequencing and analysis for chromosomal CNVs ≥ 15 Mb detected 40 % of HGSOC. Surprisingly, we detected similar proportions of early and advanced stage HGSOC cancers with this approach. This finding was unexpected because one would assume a higher detection rate in the advanced stage cases, given the lower tumor bulk of early disease. This suggests that the detection of ovarian tumor CNVs in plasma is not directly related to cancer stage; other biological factors such as fractional concentration of tumor DNA in plasma, tumor genetic heterogeneity, vascularity, and cell turnover may also be important influences on detection rates.
A limitation of our study was the inability to correlate the plasma sequencing data with paired tumor DNA due to the absence of suitable archived specimens. However, the principle that tumor DNA is detectable in plasma using NIPT sequencing platforms has been previously established [2, 3]. Furthermore, the majority of genomic aberrations detected in our cases included common imbalances previously reported in a cohort of 489 HGSOC specimens , supporting our assumption that the DNA aberrations detected in plasma originated from ovarian tumors.
Prior “liquid biopsy” studies in ovarian cancer have relied on the identification of tumor-specific mutations in advanced disease and the postoperative monitoring of patient-specific mutations in plasma via deep sequencing [10, 11]. Our results are notable for demonstrating that it is possible to detect early stage ovarian cancer in the absence of patient-specific tumor DNA using an existing low coverage sequencing platform. Thus, high throughput whole genome plasma sequencing, with or without the addition of other biomarkers, is an exciting avenue for future studies of cancer screening. It may have utility as a cost-effective method of monitoring high risk patients for whom tumor tissue is unavailable, such as presymptomatic BRCA1/2 mutation carriers, or to assess the preoperative risk of malignancy in patients presenting with ovarian masses.
Potential reasons for the false positive WISECONDOR results in the two controls include technical issues with the archived plasma samples or reference chromosome set. The two “monosomy X” calls in the NIPT pipeline in the controls (aged 43 and 54 years) might be explained by normal age-related X chromosome loss  or low grade mosaicism . It is plausible that, with larger cohorts, algorithms could be devised that increase test specificity. Further work is also required to assess the technical issues with archived plasma samples and to develop the clinical potential of this approach.
A low coverage plasma DNA sequencing protocol used in a high throughput prenatal screening platform detected more than one in three women with early stage ovarian cancer based on common segmental chromosome gains and losses. Further refinement of this approach may have utility for future studies of ovarian cancer screening.
copy number variation
high grade serous ovarian carcinoma
non-invasive prenatal testing
within sample copy number aberration detector
Taylor-Phillips S, et al. Accuracy of non-invasive prenatal testing using cell-free DNA for detection of Down, Edwards and Patau syndromes: a systematic review and meta-analysis. BMJ Open. 2016;6(1):e010002.
Bianchi DW, et al. Noninvasive prenatal testing and incidental detection of occult maternal malignancies. JAMA. 2015;314(2):162–9.
Amant F, et al. Presymptomatic identification of cancers in pregnant women during noninvasive prenatal testing. JAMA Oncol. 2015;1(6):814–9.
Ferlay J, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136(5):E359–86.
Jacobs IJ, et al. Ovarian cancer screening and mortality in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a randomised controlled trial. Lancet. 2016;387(10022):945–56.
Narod SA, Sopik V, Giannakeas V. Should we screen for ovarian cancer? A commentary on the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) randomized trial. Gynecol Oncol. 2016;141(2):191–4.
Bowtell DD, et al. Rethinking ovarian cancer II: reducing mortality from high-grade serous ovarian cancer. Nat Rev Cancer. 2015;15(11):668–79.
Cancer Genome Atlas Research Network. Integrated genomic analyses of ovarian carcinoma. Nature. 2011;474(7353):609–15.
Straver R, et al. WISECONDOR: detection of fetal aberrations from shallow sequencing maternal plasma based on a within-sample comparison scheme. Nucleic Acids Res. 2014;42(5):e31.
Bettegowda C, et al. Detection of circulating tumor DNA in early- and late-stage human malignancies. Sci Transl Med. 2014;6(224):224ra24.
Pereira E, et al. Personalized circulating tumor DNA biomarkers dynamically predict treatment response and survival in gynecologic cancers. PLoS One. 2015;10(12):e0145754.
Russell LM, et al. X chromosome loss and ageing. Cytogenet Genome Res. 2007;116(3):181–5.
Wang Y, et al. Maternal mosaicism is a significant contributor to discordant sex chromosomal aneuploidies associated with noninvasive prenatal testing. Clin Chem. 2014;60(1):251–9.
The National Health and Medical Research Council provided salary support to ST (#1050765) and LH (#1105603). PC is supported by the Jakovich Family and the St John of God Foundation. NJH is supported by a University of Melbourne CR Roper Fellowship. None of the funders had any involvement in the study design, data collection, data analysis, manuscript preparation or publication decisions. The authors would like to thank Professor Yee Leung, Drs. Stuart Salfinger, Jason Tan, and Ganendra Raj Kader Ali Mohan at the Western Australia Gynecologic Cancer Service, and Dr Nik Zeps and Sanela Bilic at St John of God Subiaco Hospital and the Western Australia Gynecologic Oncology Biospecimen Bank. We gratefully acknowledge the patients who participated in this study.
Funding for this study was provided by the Norman Beischer Medical Research Foundation.
Availability of data and materials
The datasets during the current study are available from the corresponding author on reasonable request.
PC contributed to the conception and study design, acquisition, analysis, and interpretation of data, drafting, and final approval of the manuscript, and is accountable for all aspects of the work. NF contributed to the acquisition, analysis, and interpretation of data, drafting and final approval of the manuscript, and is accountable for all aspects of the work. ST contributed to the conception and study design, acquisition, analysis, and interpretation of data, drafting and final approval of the manuscript, and is accountable for all aspects of the work. NH contributed to the acquisition, analysis and interpretation of data, drafting and final approval of the manuscript, and is accountable for all aspects of the work. MP contributed to the study design, acquisition, analysis, and interpretation of data, drafting and final approval of the manuscript, and is accountable for all aspects of the work. LH contributed to the conception and study design, acquisition, analysis, and interpretation of data, drafting and final approval of the manuscript, and is accountable for all aspects of the work. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Ethical approval and consent to participate
Ethical approval for the study was granted by the Human Research Ethics Committees at the Mercy Hospital for Women, 163 Studley Road, Heidelberg, Victoria 3084 Australia (ref R15/41) and St John of God Hospital, 12 Salvado Road Subiaco, Perth, Western Australia 6008 (ref 916). All plasma samples used in this study were obtained from the Western Australia Gynecologic Oncology Biospecimen Bank. These were prospectively collected between January 2013 and August 2015 with informed consent from patients prior to undergoing surgery.
Additional file 1:
CNVs >15 Mb in the Western Australia Biospecimen Bank Dataset, CNVs reported in an ovarian cancer genome atlas, and coordinates of overlapping regions. (XLS 52 kb)
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
About this article
Cite this article
Cohen, P.A., Flowers, N., Tong, S. et al. Abnormal plasma DNA profiles in early ovarian cancer using a non-invasive prenatal testing platform: implications for cancer screening. BMC Med 14, 126 (2016). https://doi.org/10.1186/s12916-016-0667-6
- Non-invasive prenatal testing
- High-grade serous carcinoma
- Ovarian cancer screening
- Circulating tumor DNA
- Low coverage sequencing
- Copy number variations
- Genomic profiling
- Liquid biopsy