Characterizing the levels of circulating tumour DNA in early stage prostate cancer

Jelena Belic ^1,2,3,11 Gahee Park ^1,2,3,11 Anne George Anne George ^4,5,6 Anne Babbage ^1,2,3 Wing Kit Leung ^3,7 ICGC Working Group UK ICGC Working Group UK ¹² Anne Warren ^4,8 Jonathan Wan Jonathan Wan ^3,7 Katrin Heider ^3,7 Christopher Smith ^3,7 Florent Mouliere ^3,7,9 Vincent Gnanapragasam ^3,4,5,6 Andrew Lynch ^7,10 Nitzan Rosenfeld ^3,7 Charlie Massie ^1,2,3,4,6

¹Department of Oncology, Hutchison/MRC Research Centre, Cambridge Biomedical Campus, University of Cambridge, Cambridge, UK
²Early Detection Programme, CRUK Cambridge Cancer Centre, Cambridge, UK
³Cancer Research UK Major Centre – Cambridge, Cancer Research UK Cambridge Institute, Li Ka Shing Centre, Robinson Way, Cambridge, UK
⁴Translational Prostate Cancer Group, Cancer Research UK (CRUK) Cambridge Cancer Centre, Cambridge, UK
⁵Academic Urology Group, Department of Surgery, University of Cambridge, Cambridge, UK
⁶Urological Malignancies Programme, CRUK Cambridge Cancer Centre, Cambridge, UK
⁷Cancer Research UK Cambridge Institute, Li Ka Shing Centre, Robinson Way, Cambridge, UK
⁸Department of Pathology, Addenbrooke's Hospital, Cambridge, UK
⁹Current affiliation: Department of Pathology, Cancer Center Amsterdam, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, Netherlands
¹⁰School of Medicine/School of Mathematics and Statistics, University of St Andrews, St Andrews, UK
¹¹*, Equal Contribution
¹²ICGC, Consortium, UK

Analysis of circulating tumor DNA (ctDNA) allows non-invasive analysis of the cancer genome and the assessment of tumor burden in most metastatic cancers. However, ctDNA detection rates are low in most early stage cancers, even with high-sensitivity single-locus assays or established targeted gene-panel sequencing methods. Early stage prostate cancer has a highly variable prognosis and current diagnostic biomarkers have low specificity and cannot discriminate aggressive from indolent tumours. To date there is limited evidence about the levels of ctDNA in prostate cancer patients with localized tumors, although data from small cohorts suggest allele fractions of ~0.1% which approaches sampling limits for single-target assays.

We sought to define the levels of ctDNA in clinically significant prostate cancer and establish the required analytical sensitivity of future prospective early detection assays. To achieve this, we applied an ultra-high sensitivity method developed by the Rosenfeld group that allows ctDNA detection below sampling limits. By leveraging thousands of patient-specific mutations, together with error suppressed deep sequencing, Integration of Variant Reads (INVAR) can achieve ctDNA detection at parts-per-million in limited input samples.

Our cohort includes 82 prostate cancer patients who underwent radical prostatectomy in Cambridge and whole genome sequencing (WGS) within the UK Prostate International Cancer Genome Consortium (ICGC). We combined patient-specific single-nucleotide variants (SNVs), structural variants (SVs) and insertions/deletions (INDELs) from tissue WGS, resulting in a median of ~3000 variants per patient for ctDNA analysis. We performed high-sensitivity custom capture and sequencing in pre-surgical plasma samples from each of the 82 patients.

Here, we will present the results our initial analysis, focussing on the levels of ctDNA in early stage prostate cancer and the implications for future early detection studies for clinically significant disease.