At the American Association for Cancer Research conference recently held in New Orleans, LA (April 16 - April 20, 2016) one theme that generated considerable interest was circulating tumor DNA detection, sometimes called liquid biopsy (although that term also encompasses circulating tumor cells). A few of the ctDNA presentations are highlighted below.
ctDNA at the Peter MacCallum Cancer Center, Victoria Australia
Dr. Sarah-Jane Dawson presented a ‘meet-the-expert’ early-morning workshop she called ‘Liquid Biopsies: Monitoring the Cancer Genome in Blood’. Her clinical focus is breast cancer, and she started her introduction with a Melbourne Australia physician named Thomas Ashworth who described a microscopic observation of circulating tumor cells (CTC) in the blood of a man with metastatic cancer in 1869, and in 1948 Mandel and Metais described circulating DNA in human plasma. She continued with a chart showing the balance between breadth of mutations covered (specifically digital PCR for a single locus with great sensitivity) versus to what sensitivity an assay can detect a circulating tumor mutation (i.e. up to whole-exome sequencing (WES) or whole-genome sequencing (WGS) which has been successfully performed from plasma).
As her clinical and research focus is breast cancer, she reviewed her prior work to monitor metastatic breast cancer (mBC), comparing the existing cancer antigen 15-3 (CA 15-3) biomarker against both CTCs and ctDNA in these patients. Her results showed ctDNA has greater dynamic range and greater correlation with changes in tumor burden in these patients than either CTCs or CA 15-3. In addition, she showed a figure that displayed, during treatment with Capecitabine and Carboplatin for metastatic breast cancer, ctDNA was detected to rise a full 5 months before radiology could show clinical progression of disease, reminiscent of other work reported last year regarding earlier detection of recurrent pancreatic cancer.
Her more recent research describes using circulating tumor DNA to track tumor metastatic heterogeneity in a single ER-positive and HER2-positive breast cancer patient who received two lines of targeted therapy over the course of three years. She presented three key patterns identified by their mutation type (and behavior over time), and implications for treatment given the changing picture of the circulating DNA mutations detected and evolving over the course of this individual’s treatment. As a case study, this is a forerunner of future studies examining the complex nature of tumor evolution in combination with therapeutic response.
Looking across tumor types, and in collaboration with Dr. Nitzan Rosenfeld of Cancer Research UK and Inivata Ltd., six patients with advanced breast, ovarian and lung cancers were followed over the course of 1-2 years and serial plasma samples were examined over the course of treatment via whole-exome sequencing. In these cases of advanced metastatic breast cancer, several resistance mutations acquired during therapy are being followed-up as a potentially useful screening tools.
TRAcking Cancer Evolution through therapy (TRACERx)
Dr. Charles Swanton (University College, London, UK) shared a plenary talk entitled “Tracking Cancer Evolution through Therapy (TRACERx): Implications for Immune Targeting and Surveillance”.This study (currently enrolling patients) seeks to examine over 800 non-small cell lung cancer (NSCLC) patients and examines the genetic drivers in these patients via spatial and temporal measurements, through not only examination of biopsy tissue specimens but also through the analysis of ctDNA and CTC’s over therapeutic treatment, and through to autopsy post-mortem.
Through this analysis, tumor heterogeneity, the origin of cancer subclones, and its relationship to ctDNA and CTC’s can be better understood at the level of founder events, later driver events, and clonal selection during the course of treatment. Even though this study is still recruiting participants, Dr. Swanton showed early (unpublished) data from the first 100 samples, which were skewed toward early-stage NSCLC from patients who smoked. Tell-tale signatures were spontaneous APOBEC deamination (C to T base changes, of which Sir Michael Stratton at the Wellcome Trust UK has studied extensively) as well as mitotic clock and genome doubling signatures. <http://www.ncbi.nlm.nih.gov/pubmed/24071852> He also showed impressive data of both positive and negative selective sweeps from biopsied tumor tissue, however such analyses are by their very invasiveness and complexity a limited snapshot in time.
By collaborating with Dr. Bernhard Zimmerman at Natera (San Carlos, CA), Dr. Swanton indicated that analysis of ctDNA down to a 0.1% mutation frequency, 34/50 of these early NSCLC patients (68%) detected these clonal and sub-clonal driver mutation types, and declared that these results “are certainly promising”. One individual, a never-smoker, was highlighted as an outlier in this early set of data, with a heavy mutational load (1,300 ‘trunk’ mutations) and homogeneous. This individual was a petrochemical worker, exposed to BPA and polycyclic hydrocarbons, and prior to the petrochemical work had worked in road construction with a long list of other environmental hazards.
Dr. Swanton concluded with work on neoantigens in collaboration with Memorial Sloan Kettering Cancer Center (New York, NY), comparing PD1 responders to progressive disease patients, and using whole-exome sequencing to identify neoantigens between these groups. Using this data, he is now able to predict and detect neoantigen reactive T-cells (NAR-T) in NSCLC.
10,000 clinical ctDNA profiles
Dr. Oliver Zill of Guardant Health (Redwood City, CA) presented “Comparison of over 10K clinical NGS circulating tumor DNA profiles to tissue-derived genomic compendia”, which is currently accepted for publication. Their aims are to look at ctDNA correlation to the genomic tumor biopsy, as a ‘global summary of tumor heterogeneity’.
Their assay (which Guardant Health calls Digital Sequencing™) has been recently published and has been shown to have ‘almost perfect’ analytical specificity of 99.9999% and analytic sensitivity down to 0.1% mutant allele fraction. The assay interrogates ‘critical exons’ in 70 genes (a total of about 150 kilobases of targeted sequencing), to an average read-depth of 15,000x. It combines a molecular barcoding technique (at a claimed 80%-90% efficiency) with a statistical filter, and also claim a detection of 3 mutant copies out of 4800 wild-type ones, for a allele frequency detection of 0.06%.
Their assay calls not only SNVs and indels, but also CNVs and DNA fusions. One data slide indicated a somatic alteration burden from their ctDNA measurements at 25%, dropping to 0.3% five months later, then rising again to 15% three months after that. Of interest was the observation that half of their reported variants detected occur at less than 0.4% minor allele frequency (a reported median at 0.39%, 25th percentile at 0.2%, and the 75% percentile at 1.9%)
Dr. Zill presented detection rates across cancer types; with n=5,240 samples, they achieved an 85% average detection rate. For glioblastoma multforme (GBM), their detection rate dropped to 57% likely due to the blood-brain barrier for circulating tumor DNA. By cross-referencing publicly-available TCGA (The Cancer Genome Atlas) data and using the cBio Cancer Genomics Portal (here’s the reference) to their Stage III-IV treated NSCLC data, they observed a high correlation of r=0.94 of about 4,000 TP53 mutations between their ctDNA data and the TCGA dataset. For the KRAS and PIK3CA genes a similar analysis showed similar results.
An exciting time for circulating tumor DNA
SeraCare’s Dr. Seth Harkins generated interest in the poster he presented at AACR, entitled “Methodological Considerations in the Preparation of Biomimetic Reference Materials for ctDNA Assays”. This poster is available online here. In addition, if you are developing circulating tumor DNA assays, SeraCare recently launched the Seraseq Circulating Tumor DNA-I Reference Material as a full-process, simulated plasma with fragmented and stabilized synthetic mutant DNA targets in a well-characterized background genomic DNA from a single individual. In addition, we have also made available Seraseq Circulating Tumor DNA-I Mutation Mix purified genomic DNA reference materials. More information about these products are available here.