Hope and Hype: Questions Surround Liquid Biopsies

By Kathy Boltz, PhD

Questions surrounding liquid biopsies are plentiful. Such as, what is the correlation between blood and FFPE biopsy tissue? What about location and stage of disease? Along with, what should be tested, how comprehensively to test, and how to keep the price feasible?

To address some of the uncertainty surrounding liquid biopsies in oncology, we interviewed both Vincent Miller, MD, Chief Medical Officer of Foundation Medicine and David Spetzler, PhD, MBA, Chief Scientific Officer of Caris Life Sciences for this article.

Comprehensive genomic profiling through liquid biopsy

Liquid biopsies are noninvasive tests, usually of blood, that detect DNA, cells, or pieces of cells that are shed from tumors, including primary and metastatic sites. The appeal is obvious. Use a blood draw, rather than a biopsy of a solid tumor, to conduct comprehensive genomic analyses on patients whose solid tumor samples will not allow for that; monitor disease and its response to treatment through serial blood draws; and ideally, detect cancer very early through the blood testing.

According to Vincent Miller, MD, Chief Medical Officer of Foundation Medicine, genomic profiling is becoming accepted as warranted and helpful for patients with advanced stage disease. However, the amount of tissue obtained under an initial diagnostic biopsy may not be enough to allow genomic profiling for some patients. Patients with insufficient tissue may not be able to undergo a new biopsy because of their condition, the treating physician’s preference, or issues such as time and logistics. So, Dr. Miller feels that the clearest indication of when to use a liquid biopsy is in patients where no genomic profiling would be feasible from a traditional tissue-based biopsy.

Liquid biopsies have the potential to allow patients whose tissue biopsy is nondiagnostic or who do not have an area of tissue amenable to biopsy access to clinical trials. Foundation Medicine will soon launch Foundation Act, a genomic profiling circulation tumor DNA (ctDNA) assay that is designed to allow for the richest profiling possible from ctDNA.

This type of biopsy assay may benefit both patients seeking treatment in clinical trials and pharmaceutical companies seeking enrollment in clinical trials. As more genomic data become available, biopharma will be able to further analyze trial outcomes and possibly pinpoint biomarkers.

Dr. Miller sees two different types of liquid biopsies: the ctDNA genomic profiling assay and the monitoring type assay. “The ctDNA would be at a higher price point and it’s much more detailed and rich on the genomic side,” he said, while the monitoring assay “would be done at a lower price point and have narrower genomic content.”

Along with Foundation Act, Genomic Health is also launching a liquid biopsy assay, called Oncotype SEQ™, in the first half of 2016. Oncotype SEQ is a blood-based mutation panel that uses next-generation sequencing to identify actionable genomic alterations for the treatment of patients with late-stage lung, breast, colon, melanoma, ovarian, or gastrointestinal cancer. Genomic Health has plans to deliver several more liquid biopsy tests through its Oncotype IQ™ Genomic Intelligence Platform.

Monitoring biomarkers in NSCLC

Non-small cell lung cancer (NSCLC), which is typically detected at a late stage, has been an active area in the development of liquid biopsies. Difficulties obtaining adequate tissue samples are not uncommon in NSCLC. As third-generation EGFR tyrosine kinase inhibitors (TKIs) targeting the EGFR mutation T790M are becoming available, noninvasive approaches to detecting T790M are increasingly important to guide treatment over the course of the disease.

To examine the use of liquid biopsies for detecting the EGFR T790M mutation, a multi-institutional Stand-Up-To-Cancer collaboration compared tumor biopsies with analysis of circulating tumor cells (CTCs) and ctDNA from blood samples in 40 patients with NSCLC.1 The CTC- and ctDNA-based genotyping were unsuccessful in 20%-30% of the cases, but combining the analyses together enabled genotyping in all patients who had a blood sample available. The combined analyses identified the T790M mutation in 35% of patients (14/40) whose concurrent biopsy was negative or indeterminate.

At ASCO 2015, an innovative study in NSCLC was presented on monitoring urinary circulating tumor DNA for the early acquisition of T790M and understanding ctDNA kinetics in patients on second-line anti-EGFR treatment.2 When urine T790M was compared with tissue-positive samples from 14 patients, the sensitivity was 100%. Further, urine liquid biopsy detected the EGFR T790M mutation as much as 3 months before radiological progression occurred based on data from 24 patients, and ctDNA spikes that possibly correlated with tumor lysis were observed among the 13 patients monitored who were treated with second-line, anti-EGFR TKIs.

A different study in patients with NSCLC compared circulating free DNA (cfDNA) in baseline blood draws with biopsy tissue with known mutations.3 The study analyzed samples from 97 patients whose tissue samples indicated 1 of 2 EGFR mutations was present: either the exon 19 deletion or the L858R mutation in exon 21. These mutations were successfully assessed in corresponding cfDNA from baseline blood in 76 of the 97 patients (78%). 

Honing in on ctDNA

These small studies suggest the power of liquid biopsies, but the data to fully support their use across a range of cancers still needs to be generated.

“The proportion of time that you actually find evidence of colon, breast, or brain cancer in ctDNA is largely unknown,” said Dr. Miller. “It’s almost certainly stage-dependent, with earlier stage cancer shedding less DNA than later stage cancer, and it remains to be elucidated in a tumor-specific manner.”

Dr. Miller stated that a fundamental difference between tissue and liquid biopsies is that much more substantial sequencing is required for liquid biopsies to know if an alteration is real or not, so that problems such as false positives are avoided.

Dr. Miller further explained that Foundation Medicine has undertaken a large study of their ctDNA assay paired with FoundationOne® testing (NCT02620527 on ClinicalTrials.gov). The study, with estimated enrollment of 2000 patients, will include patients with very tight temporal linkage between tissue and blood and also patients with a variable but bigger difference between tissue and blood. For the latter group, the question will be how much the genome is diverging, while samples collected from blood and biopsy at the same timepoint are expected to be quite similar.

The Foundation Medicine study also seeks to answer key questions, such as what proportion of the time does stage IV head and neck cancer shed ctDNA vs thyroid cancer vs colon cancer vs lung cancer? If ctDNA is found, what is the concordance between what shows up on the ctDNA assay and on the tissue assay? What is missed by using the ctDNA assay? For example, Dr. Miller explained that it is harder to detect amplifications, such as MET and HER2, with a ctDNA assay than with a tissue-based assay. If the ctDNA test does not find anything, is that result final, should it be believed, or should the tissue test remain the gold standard and the first-choice option?

“We take a lot of time to do it right, with real rigorous analytical and clinical validation studies, both for FoundationOne and now for Foundation Act,” said Dr. Miller. “The onus is on us to provide the data so discerning clinicians can decide when it’s really prime time."

“If we get the data across ten tissue types that confirm liquid biopsy and can’t really supplant tissue-based testing, then it’s less of a universal, binary switch from tissue to blood across the board—I think it’s naive to assume that’s the way things are going to move.”

Another path to a wealth of liquid biopsy data 

Currently, robust concordance and validation studies comparing ctDNA with tumor tissue have not yet been done. At Caris Life Sciences, David Spetzler, Chief Scientific Officer said, “Theoretically, liquid biopsies could help address tumor heterogeneity, but that has yet to be proven. The false positive rate is a challenge.” Dr. Spetzler explained that Caris found a false positive rate of 40% a few years ago with ctDNA, and so they are focusing on protein content from exosomes.

“Exosomes are secreted by all cells as part of normal function, so they represent a significant source of biological information that includes early perturbations in pathways leading to cancer,” said Dr. Spetzler. “They serve as a biological substrate that allows a systems biology approach to understanding the changes in the system associated with the emergence of cancer. These signals are found earlier than ctDNA and impart more information.”

A validation study of the exosome approach for breast cancer screening for dense breasts was recently presented at the San Antonio Breast Cancer Symposium.4 The ADAPT Biotargeting System™ measures millions of protein complexes simultaneously in a nondestructive manner, and Dr. Spetzler said it can address the complexity of interacting proteins, and thus it enables a systems biology approach to characterizing disease.

Meeting the demands of regulators and payers

Developing data to support the use of liquid biopsies is key to moving them to market.

“Treatment decision making is becoming more binary,” Dr. Miller said, meaning choosing a treatment depends on if a marker is present, and if it is, a veritable homerun or a strikeout could result if the marker is absent. “The mandate should be higher on assays being used to make those decisions. We embrace that philosophically, and we’re well-positioned to address regulatory oversight that might come forward.”

Similarly, Dr. Spetzler emphasized the importance of high quality testing. He said, “We believe more regulations are appropriate to ensure quality and utility standards. We expect studies that test impact on patient survival to be required.”

“Payers demand demonstration of clinical utility,” said Dr. Spetzler. He stated that many tests use code stacking, but he expects that to change, with studies testing impact on patient survival becoming required.

“I believe more studies and clinical utility data are needed to further validate the most relevant approaches for a true liquid biopsy,” said Dr. Spetzler. “That said, this is such an exciting time for molecular science and the role it plays in helping cancer patients fight their disease."

Liquid biopsies offer a lot of promise, and research to develop them will move personalized treatment forward. Questions still remain. What is best to test? How comprehensive should this testing be? Can liquid biopsies get us to a “panomic” approach that encompasses the genome, the proteome and other “-omes”? If so, how can this technique be priced to be accessible to patients?


1. Sundaresan TK et al. Detection of T790M, the acquired resistance EGFR mutation, by tumor biopsy versus noninvasive blood-based analysis. Clinical Cancer Research 2015; doi: 10.1158/1078-0432.CCR-15-1031.

2. Husain H et al. Kinetic monitoring of EGFR exon 19 del, L858R, and T790M in urinary circulating tumor DNA predicts radiographic progression and response in patients with metastatic lung adenocarcinoma. American Society for Clinical Oncology 2015 Annual Meeting. Abstract 8081.

3. Karachaliou N et al. Association of EGFR L858R mutation in circulating free DNA with survival in EURTAC trial. JAMA Oncology 2015; doi:10.1001/jamaoncol.2014.257.

4. Domenyuk V et al. Adaptive dynamic artificial poly-ligand targeting (ADAPT): aptamer-based profiling of liquid biopsies to improve the accuracy of breast cancer diagnoses in women with dense breast tissue. San Antonio Breast Cancer Symposium, December 8-12, 2015. Poster P2-01-08.

© Caribou Publishing. All rights reserved. Reproduction in whole or in part is prohibited.