Author: Janis C. Kelly

Bits of tumor cell somatic DNA shed into the circulation or released when cells die can now be detected and counted, thanks to advances in gene sequencing. This circulating tumor DNA (ctDNA) is derived from somatic mutations that occur in the tumor during an individual’s life, unlike hereditary mutations that are present in every cell in the body, so ctDNA is a specific cancer biomarker that can be detected, measured, and tracked.

Monitoring ctDNA is expected to provide clinicians with faster, cheaper, less invasive ways to assess cancer patients’ clinical status and response to therapy. ctDNA assay for multiple genes via next-generation sequencing (NGS) might become a “liquid biopsy” alternative to invasive tissue biopsy, experts told Medscape Medical News.

However, they also cautioned that rigorous testing of this concept is needed before the test can be used in practice, saying: “for now, we would counsel clinicians not to jump the gun on this.

Faster, Cheaper, More Accurate Tumor Tests
Paul B. Chapman, MD, a medical oncologist with the Melanoma and Sarcoma Service at Memorial Sloan Kettering Cancer Center in New York City and Chair of the Medical Advisory Panel at the Melanoma Research Alliance in Washington, DC, said that ctDNA assay is less invasive than biopsy, requires no radiation exposure, is relatively inexpensive, uses fresh DNA not exposed to preservatives, and allows near real-time monitoring of response to treatment.

“The beauty of ctDNA monitoring is the speed,” Dr Chapman said. “If you are looking for a change in a tumor, based on CT scan, you are talking about not only killing billions of tumor cells but also waiting for the resulting cell debris to be cleared by the body before the change shows up on imaging. That can take weeks. But cell death happens in minutes to hours, so you would expect the change in ctDNA to be a quick effect, and it is.”

Chetan Bettegowda, MD, Luis A. Diaz Jr, MD, and colleagues at Johns Hopkins Medical Institutions in Baltimore, Maryland, working in collaboration with 16 other institutions worldwide, recently reported that ctDNA levels differentiate early vs advanced tumors in a wide variety of cancers. In 640 patients, the researchers found that ctDNA was detectable in over 75% of advanced pancreatic, ovarian, colorectal, bladder, gastroesophageal, breast, melanoma, hepatocellular, and head and neck cancers but was found in less than 50% of primary brain, renal, prostate, and thyroid cancers. In 206 additional patients with metastatic colorectal cancers, ctDNA for KRAS gene mutations had 87.2% sensitivity and 99.2% specificity. The researchers also found that 96% of patients who relapsed after epidermal growth factor receptor (EGFR) blockade had ctDNA indicating one or more mutations in genes in the mitogen-activated protein kinase pathway.

Counting the methyl groups on ctDNA fragments might be even more precise. An assay for hypermethylation of one or more breast cancer genes was 95% accurate in detecting identifying patients with metastatic breast cancer.

However, Ben Ho Park, MD, PhD, from John Hopkins Sidney Kimmel Comprehensive Cancer Center and the department of chemical and biomolecular engineering at Johns Hopkins University in Baltimore, urged some caution in interpreting these data.

Dr Park told Medscape Oncology that key unanswered questions include: “Can we use ctDNA for identifying patients that have mutations that are susceptible to targeted therapies? Can we use ctDNA to follow response to therapies in metastatic disease that actually makes a positive difference for patient care? And can we use ctDNA as tumor biomarkers for assessing residual disease for early stage cancers, and to determine if we can tell patients that they are cured after surgery and do not need additional adjuvant therapies [including chemotherapy]?”

Advances in Gene Sequencing Led to Clinically Useful ctDNA Tests
A major factor in the surging interest in ctDNA has been the rapid drop in cost for analyzing genetic information. In a development that puts Moore’s Law to shame, costs for whole genome sequencing dropped from $1 billion for the first complete genome sequenced by the Human Genome Project to $350,000 per genome in 2008 and more recently to about $1,000 for a genome sequenced with Illumina’s commercially available HiSeq X Ten sequencing system.

Cancer researchers have been challenged by the problems of how to discriminate ctDNA from normal cell-free DNA, to detect extremely low levels of ctDNA, and to count accurately the number of mutant ctDNA fragments in a blood sample. The sensitivity of polymerase chain reaction (PCR)-based digital approaches improved with the addition of NGS, in which DNA is fragmented into small segments that can be quickly sequenced in millions of parallel reactions known as “reads” and then reassembled so that the set of reads shows the entire DNA sequence.

The half-life of ctDNA is about two hours, and changes in ctDNA levels can be apparent days to weeks before changes in imaging or in protein biomarkers. Because ctDNA is specific for the individual patient’s tumor, it is likely to avoid some of the false-positive problems associated with other cancer biomarkers.

Detecting Cancer and Monitoring Cancer Stages Without Biopsy
Most ctDNA strands contain between 180 and 200 base pairs, similar to the 180 base-pair multiples characteristic of apoptosis, and are thought to result mainly from passive release into the blood of ctDNA after cell death.

The presence of ctDNA after resection (but before adjuvant chemotherapy) indicates residual disease. Absence of ctDNA might identify a patient subgroup at low risk for recurrence who could be spared the risk, expense, and discomfort of adjuvant therapy. In 2008, a team of Johns Hopkins researchers reported that mutation-specific probes for 18 subjects undergoing multimodality therapy for colorectal cancer and monitored for two to five years showed that “ctDNA measurements could be used to reliably monitor the tumor dynamics in subjects with cancer who were undergoing surgery or chemotherapy.”

The authors suggested that ctDNA levels reflect the total systemic tumor burden because they decreased after complete resection and increased as new radiologically-apparent lesions developed. The researchers also pointed out that micrometastatic lesions (smaller than a few millimeters) contribute to tumor burden and to ctDNA levels although they are not detectable by imaging.

“For a melanoma patient who is free of disease after surgery but at risk for recurrence, ctDNA could be a nice way to follow without having to do frequent CT scans,” Dr Chapman said. It is also expected to be useful in situations in which tissue biopsy is undesirable or cannot be done.

However, an important unanswered question is how often these tests should be done. “We need to know how meaningful small changes in the ctDNA level are — sensitivity, specificity, and lead-time bias,” he said.

Monitoring Tumor Burden, Response to Therapy, and Resistance
“Another key advantage is that ctDNA could overcome the issue of tumor heterogeneity,” commented Dr Park. “Different sites of disease often have different mutational profiles. Since blood, and therefore ctDNA, acts as a ‘reservoir’ for all sites of disease, ctDNA is representative for all sites of metastases.”

A research team from Dr Chapman’s lab led by Parisa Momtaz, MD, reported at this year’s American Society of Clinical Oncology (ASCO) Annual Meeting that in melanoma patients with BRAF-v600E mutations, ctDNA correlated well with tumor burden measured by radiographic imaging.

“The ASCO cohort of 11 patients was to convince ourselves that we could get the assay to work (proof-of-principle) and that it correlated with what was clinically going on,” Dr Chapman said. “We have now studied ctDNA in about 60 patients. We are focusing on patients treated with immunotherapy because radiographic evaluation of these responses are somewhat equivocal. We hope that ctDNA will add more clarity and tell us whether the immune system is attacking the tumor or not.”

Also at ASCO, Nicholas C. Turner, MD, consultant medical oncologist at the Institute of Cancer Research, London, United Kingdom, and colleagues reported that in primary breast cancer tumor-specific ctDNA levels can predict early relapse.

ctDNA might also provide early warning that the patient has developed treatment-resistant disease. Sarah B. Goldberg, MD, MPH, assistant professor of medicine at Yale University School of Medicine, New Haven, Connecticut, and colleagues reported that ctDNA could be used to detect both sensitizing and resistance EGFR mutations in patients with advanced lung cancer treated with EGFR tyrosine kinase inhibitors. The researchers suggested that using NGS to detect sensitizing and resistance mutations in plasma ctDNA might allow earlier identification of resistance in patients treated with targeted therapies.

Similarly, a team of researchers from nine cancer centers in Italy found that KRAS mutations in ctDNA could be detected in over 35% of patients with non–small-cell lung cancer who became resistant to tyrosine kinase inhibitors.

In discussing this research, Luis A. Diaz Jr, MD, from the Ludwig Center for Cancer Genetics and Therapeutics at Johns Hopkins University School of Medicine in Baltimore, and Alberto Bardelli, PhD, from the Laboratory of Molecular Genetics, Institute for Cancer Research and Treatment, University of Torino, Italy, wrote, “This understanding of the mechanisms of acquired resistance to targeted agents at the molecular level can be used to plan combinatorial treatments with drugs that will suppress the expansion of the clones that are responsible for most of the current failures of medical treatment. This knowledge could result in the early adoption of alternate therapies before clinical resistance is detected.”

Dr Chapman said, “I have this fantasy that you might use this to screen chemotherapy drugs in a patient. You could give them one dose of drug, then measure their ctDNA response. If there is no tumor death, the ctDNA levels wouldn’t change.”

ctDNA: From Bench to Bedside
Currently, monitoring genetic changes in a tumor requires multiple biopsies. “In the future, it might just be a matter of drawing a tube of blood,” Dr Chapman said.

However, Dr Park warned that there has yet been little to no validation of ctDNA testing and that published studies show considerable variability due partly to lack of quality control and uniform standards.

“How the plasma is prepared makes a huge difference, which isn’t always appreciated. How the ctDNA is analyzed is also quite variable among studies, with some technologies being better suited for specific applications. So for now, we would counsel clinicians not to jump the gun on this. We have to be extremely thoughtful and careful when dealing with ctDNA and its applications,” Dr Park commented. He quoted another researcher (Dan Hayes, MD, from the University of Michigan), who says it best: “A bad test can be just as bad as a bad drug.”

“Therefore I believe we need to apply the same rigorous standards of testing drugs to the development of ctDNA as a liquid biopsy,” Dr Park said.

Dr Chapman disclosed anticipated research funding from Trovagene. Dr Park reported no relevant financial interests.

Ann Transl Med. Published online January 2014. Abstract; Sci Transl Med. Published online February 2014. Abstract