2018 ISPD Conference

Sensitivity of non-invasive prenatal testing for cancer detection and treatment monitoring in pregnant women
Liesbeth Lenaerts1, Joris Vermeesch2, Nathalie Brison3, Maria Neofytou4, Magali Verheecke5, Luc Dehaspe6, Hans Wildiers7, Sigrid Hatse8, Frédéric Amant5

1Catholic University Leuven, Leuven, Belgium
2Center for Medical genetics, Katholieke Universiteit Leuven, Leuven, Belgium
3Center for Human Genetics, University Hospital Leuven, Leuven, Belgium
4The Cyprus Institute of Neurology and Genetics, Nicosia, Cyprus
5UZ/KU Leuven, Leuven, Belgium
6Center for Human Genetics, KU Leuven, Leuven, Bahamas
7Medical Oncology, University Hospitals Leuven, Leuven, Belgium
8Department of Oncology, KU Leuven, Leuven, Belgium
Objectives
We developed and implemented a genomewide cell-free DNA analysis pipeline, coined GIPSeq (Genomic Imbalance Profiling from cell-free DNA SEQuencing) for non-invasive prenatal testing (NIPT), allowing genome-wide detection of foetal and maternal chromosomal imbalances.  Analysis of about 40.000 asymptomatic pregnant women identified 7 profiles reminiscent of cancer-related copy number variations (CNVs). All seven were referred to the oncology unit and whole-body diffusion-weighted magnetic resonance imaging uncovered a malignant process in six of them. Based on these incidental findings, we set out to explore the sensitivity of GIPSeq for cancer detection and treatment monitoring in pregnant women.
Methods
Pregnant women diagnosed with invasive breast or cervical cancer (i.e. the most prevalent solid cancer types encountered during pregnancy) were recruited in University Hospitals Leuven or via the International Network on Cancer, Infertility and Pregnancy (INCIP). A plasma sample was taken for GIPSeq. Non-pregnant premenopausal breast or cervical cancer patients were included as controls. In case of an aberrant GIPSeq-profile, consecutive plasma samples were taken to assess treatment response. Tumour biopsy specimens were subjected to shallow sequencing to investigate the specificity of CNVs detected in cell-free DNA.
Results
Pregnant breast (n=17) and cervical (n=4) cancer patients, and non-pregnant breast (n=17) and cervical (n=11) cancer controls were enrolled. For breast cancer, detection of abnormal GIPSeq-profiles significantly correlated with tumour stage. Furthermore, sensitivity was 3,3 times higher in pregnant (47,1%) than in non-pregnant (14,3%) breast cancers; the underlying reasons being explored. No aberrant GIPSeq-profiles were detected in pregnant cervical cancers, whereas sensitivity was 18% in non-pregnant controls and not depending on cancer stage. For true positive cases, the CNV-profile of tumor biopsy DNA showed comparable genomic imbalances as seen in circulating DNA. Furthermore, GIPSeq-profiling was able to follow treatment response.
Conclusions
These preliminary results indicate that GIPSeq of circulating cell-free DNA is able to detect cancer-specific CNV-profiles in a proportion of cancer cases, and that detection might depend on the pregnancy status of the women. Further recruitment is ongoing and supplementary analyses are planned to improve the sensitivity of the pipeline for cancer detection. Optimizing a cancer detection and treatment monitoring in pregnant women may ultimately lead to improved prognosis.