Wafik S. El-Deiry, MD, PhD, FACP

Liquid biopsy allows from a blood sample the analysis of shed tumor DNA and other components found in blood, including microRNA, proteins, cytokines, metabolites, immune cells, platelets and other debris from tumors, such as exosomes. It is relatively noninvasive and can provide rapid feedback to clinicians on aspects of clinical importance. These include the amount of circulating mutated tumor DNA that can serve as a tumor marker; the presence of specific mutations, such as KRAS, that may steer patients with colorectal cancer away from anti-EGFR targeted therapies; or the presence of kinase-activating EGFR mutations in lung cancer that may direct therapy toward small molecule EGFR inhibitors.

Much more can be learned from liquid biopsies, including emerging drug resistance mechanisms over time and insight into intratumoral heterogeneity, both for treatment-naive tumors and in cases when resistance develops after treatment. Liquid biopsies also can inform about tumor mutation burden and may provide clues that a mismatch repairdeficient tumor is present, or give a clue to a repair defect or an inherited tumor susceptibility.

Although clinical validation has been emerging, liquid biopsies remain not universally accepted or used. Multiple studies have looked into concordance of results with tumor tissue, and the results have continued to be promising. It is always more difficult when there is absence of a mutation as to whether to stop there or to proceed with a tissue biopsy. It also is becoming clear that liquid biopsy can reveal important aspects of tumor heterogeneity among metastases, information that cannot be revealed without multiple tissue biopsies from multiple lesions.

The study by Razavi and colleagues used a 508-gene circulating tumor DNA assay at depth of more than 60,000 times on blood samples to compare with tissue collected within 6 weeks of the blood collection. The analysis which included both circulating tumor DNA and white blood cell DNA from the same blood samples included 161 evaluable patients with metastatic breast, lung or prostate cancers.

Researchers reported high concordance for various detected abnormalities, ranging from 58% for copy number variations and 67% for gene fusions, to 75% for single nucleotide variants or indels. This concordance is consistent other studies. The researchers claim the high intensity and large number of genes in the panel is unique compared with commercially available platforms but, of course, this is a rapidly evolving and adapting field. Guardant360 (Guardant Health) tests 73 genes from a blood draw, Foundation Medicine reports more than 60 genes analyzed through its FoundationACT assay, and Personal Genome Diagnostics reports 64 genes analyzed in its PlasmaSELECT 64 panel. It remains unclear whether use of 508 genes is superior to 60 or 70+ genes.

The researchers comment on and are interested in further development of blood-based high-intensity sequencing for early cancer detection. That remains a far-away goal, with the need to test normal individuals as well as those at risk with anticipated issues regarding yield from the testing. This likely will encounter similar issues as other tumor markers, such as PSA, carcinoembryonic antigen or CA125 that are not routinely used in the general population for early cancer detection. Issues such as cost and yield will be important. However, in selected populations at high risk whether they are patients with strong family histories or those who already had cancer but went into remission this technology is very promising. Consistent with this prediction, an article by Charles Swanton, MD, PhD, Christopher Abbosh, MD, and colleagues that made the cover of the journal Nature on May 25 suggested that liquid biopsies could be used for early detection of relapse after surgery for nonsmall cell lung cancer.

Reference:

Abbosh C, et al. Nature. 2017;doi:10.1038/nature22364.

Charu Aggarwal, MD, MPH

This is an important study because it not only looks at circulating tumor DNA, but it also looks at concordance of mutations in tissue. From 161 eligible patients, 124 actually had concordance in tissue. Researchers observed similar mutations in both blood and tissue.

We run into issues where we are not able to get gene sequencing on tissue either because it is unavailable, out of sight/inaccessible or because a patient does not want to undergo another biopsy. As more and more targeted therapies come to the forefront, it is important to characterize the gene sequencing and genomic profile of the patients.

Researchers looked at circulating tumor DNA on a broad platform — a 500-gene panel — and their tissue sequencing also involved a broad panel of 410 genes. This involves a lot of information, and the researchers should be applauded for this type of study and their ability to be able to get matched tissue and blood on a great majority of patients.

This has implications not just for targeted therapy, but also for immunotherapy because there is correlation between tumor mutation burden and response to immunotherapy. If you are able to demonstrate that you do not always need tissue — and instead you can use blood — it is a great advancement because blood is so much easier to retrieve.