The need for standardising pre-analytical sample handling and integrating analytical workflows for ensuring liquid biopsies are ready for prime-time Karen Page1

The need for standardising pre-analytical sample handling and integrating analytical workflows for ensuring liquid biopsies are ready for prime-time
Karen Page1, PhD, Prof. Jacqueline A Shaw1, PhD, David S Guttery1,2, PhD
1The Leicester Cancer Research Centre, University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary, Leicester, LE2 7LX, United Kingdom.
2Corresponding author
The dawn of cancer genomics has heralded an unprecedented era of precision medicine, allowing the identification of genome-wide somatic driver alterations that can be used for early cancer diagnosis, prognosis, stratification to optimum therapies and monitoring of developing resistance, as well as predicting which patients are likely to relapse. Currently, clinicians and translational researchers are utilising our vastly improved understanding of the heterogeneous molecular landscape of cancer to stratify patients appropriately to carefully selected targeted treatments with the aim of ensuring patients receive the right treatment at the right time. Ultimately, there is hope that this will allow clinicians to either cure their patients disease (at the earliest stages) or manage their disease for the longest time possible, while still ensuring the highest possible quality of life (in advanced stages).
The current “gold standard” for diagnosing cancer and determining optimal therapeutic strategies is through surgical biopsy. However, this method has several limitations, not least of all its invasiveness. Surgical biopsies are also extremely limited in that they only offer a “snapshot” in space and time of the tumour depending on the disease stage and local area. Further, they are known to be unrepresentative of tumour heterogeneity, potentially resulting in predominance of resistant clones that are refractory to therapy and eventually disease progression/therapy resistance. Therefore, surgical biopsy cannot offer any indication of treatment efficacy and tumour evolution (especially in the metastatic setting) during a patient’s therapy. Compounding this, most biopsies are archived as formalin-fixed, paraffin embedded blocks that are used for routine pathology, necessitating highly-sensitive methods (such as next-generation sequencing (NGS) and digital polymerase chain reaction (dPCR)) for analysis of very limited amounts of poor quality nucleic acid. However, many of these limitations can be circumvented using liquid biopsies.
The liquid biopsy holds huge potential as a more cost-effective, easier, less-invasive method for diagnosing and monitoring cancer, as well as predicting response to many currently available therapies. The liquid biopsy is used as an “umbrella” term that encompasses different analytes that can be identified in blood samples including: circulating tumour cells (CTCs), circulating free DNA (cfDNA), circulating tumour DNA (ctDNA – the tumour-derived fraction of cfDNA), circulating RNA (including circulating microRNA), extracellular vesicles such as exosomes and more recently, platelets (in particular tumour-educated platelets or TEPs) (REF). Of these, CTCs and ctDNA are currently the most validated analytes and the most likely to translate into routine clinical use within the next 2-5 years due to their proven ability for early cancer diagnosis (REF), to dynamically monitor patient response to therapy (REF), to predict relapse (REF) and offer actionable somatic alterations for stratifying patients to the optimum therapies in real time (REFS). Overall, the liquid biopsy can complement a personalised medicine approach to cancer treatment as well as providing innovative methods towards patient selection in clinical trials.
However, despite the huge potential of the liquid biopsy for managing patient therapy, several technical and logistical challenges need to be overcome before it can be truly integrated into routine clinical use. Of note, there is still no widely accepted consensus regarding pre-analytical blood sample handling and technologies used for extracting cfDNA and CTCs, as well as no robust and reproducible workflow towards consistent molecular analysis of liquid biopsy analytes – all of which are urgently needed before liquid biopsies can become routinely used in the clinic. Here, we discuss the current status of liquid biopsy testing in patients with solid tumours and the variation in blood sample handling and molecular workflows used to analyse them, focusing on ctDNA and CTCs.
Circulating tumour DNA
There are many pre-analytical variables that are inherent in affecting the downstream data obtained from ctDNA analysis. The first is specimen type (i.e. plasma or serum). Several studies have demonstrated remarkably higher abundances of total cfDNA (both normal and ctDNA) derived from serum than from plasma (REFS); however, variability in cfDNA yields is far greater from serum. Plasma and serum constitute the non-cellular fraction of whole blood, but unlike plasma samples, serum is obtained by allowing whole blood to clot at room temperature for 15 to 30 mins prior to processing. This results in significant leukocyte and haematopoietic cell lysis, diluting the concentration of ctDNA present in the sample. Conversely, to obtain plasma whole blood is generally processed as soon as possible post-venepuncture, thereby reducing contamination by genomic DNA and hence, is currently considered the optimal specimen type for analysis of ctDNA.
The second – and arguably the most critical – pre-analytical variable is sample processing, which can be further stratified into 5 categories: 1) the location from which the sample is derived; 2) the type of tube used and time to centrifugation; 3) centrifugation of the blood; 4) storage conditions and 5) the cfDNA extraction method.
1) The vast majority of studies have acquired blood from peripheral veins; however, in certain instances this may not be the optimal strategy towards obtaining the most informative results. For instance, compared to plasma, ctDNA derived from cerebral spinal fluid (CSF) has been shown to provide a more comprehensive view of genomic alterations in patients with primary tumours of the brain and spinal cord (REFS). Studies comparing other sites (for instance pulmonary veins) are limited, but nonetheless warrant further investigation.
2) The type of collection tube is the most diverse pre-analytical variable. Many studies have reported an increase in total cfDNA yield over time prior to centrifugation when blood is drawn into EDTA-stabilising blood tubes, mainly due to leukocyte lysis (REF). Some studies have suggested that cfDNA yield is unaffected if EDTA-stabilised blood is centrifuged within 24 hours (REFS); however, it is currently recommended that time to processing is within 6 hours, preferably 2 hours to minimise the risk of genomic DNA contamination and subsequent dilution of ctDNA. To circumvent the need for immediate blood processing,