In the Lab

On Cancer: Tumor DNA in Bloodstream May Yield New Cancer Insights

By Jim Stallard, MA, Writer/Editor  |  Monday, August 11, 2014
Experimental pathologist Jorge Reis-Filho says “liquid biopsies” could provide more comprehensive ways to assess tumors. Experimental pathologist Jorge Reis-Filho says “liquid biopsies” could provide more comprehensive ways to assess tumors.

For many people with cancer, the needle biopsy — an invasive procedure that isolates tumor tissue for analysis — is an uncomfortable part of diagnosis and treatment. Recent research reveals that the information obtained from this biopsy might be less accurate than previously thought.

As researchers identify genetic changes that trigger cancer and promote its growth, there is a growing awareness that effective treatments could be undermined by tumor heterogeneity — the variation among cancer cells both within the primary tumor and within distinct tumors formed by a cancer’s spread, or metastasis, to distant sites. For example, genetic mutations present in one biopsy sample might be absent from another biopsy taken from a different part of the tumor. This type of discrepancy complicates clinical decisions and confounds research efforts.

Now scientists are looking for less invasive and more comprehensive ways to examine tumors. One approach might be to rely on “liquid biopsies,” which analyze tumor cells or tumor cell DNA that has entered the bloodstream simply by drawing a blood sample from the patient.

Memorial Sloan Kettering experimental pathologist Jorge Reis-Filho, who studies the genetic alterations that drive the malignant behavior of breast cancer cells, discussed with us the potential of analyzing circulating tumor DNA (ctDNA) or circulating tumor cells (CTCs) in a blood sample and explained ways that ctDNA might provide insights for treatments that go beyond current conventional clinical tests.

What are the differences between CTCs and ctDNA and what can they reveal about a tumor?

CTCs have been recognized over the last decade as clinically useful cancer biomarkers in certain cases, but researchers have recently begun to focus on ctDNA as a potentially superior source of genetic information.

One limitation of CTCs is that they require elaborate methods for detection and retrieval. First, they need to be isolated from other blood material, which is difficult because they are very rare. Then extracting the DNA to study the genetic makeup of the cells is itself a cumbersome process. In addition, it is unclear what causes cells to break off from a tumor. We don’t know whether CTCs represent the entire makeup of cancer cells in the tumor or only a subpopulation, so it has yet to be defined how many CTCs need to be analyzed with molecular methods in each patient.

ctDNA, on the other hand, is much easier to isolate. When cells die, including cancer cells, some of the DNA they shed ends up in the [blood] plasma. This DNA in the circulation can be extracted from plasma and used for the characterization of the genetic makeup of tumor cells. In patients with advanced disease, tumor DNA can be collected even from a single vial of blood. As this is minimally invasive, multiple vials can be collected and the DNA extracted can provide important insights into the biology of the tumor.

More important, ctDNA may actually provide a more global — and therefore more accurate — picture of the cancer in the body. We have shown that at least in some patients, sequencing the ctDNA can capture all genetic alterations found in cancer cells in different parts of the body, because the cells dying and releasing DNA would come from all parts of the primary tumor and metastases.

What has made the study of ctDNA such a focus of interest recently?

The big change has been the introduction of next-generation sequencing, a new technology that allows us to sequence entire genomes of tumor cells to detect mutations with great accuracy. For example, we have a test here called MSK-IMPACT, which can look for hundreds of cancer-related mutations at once with very high sensitivity. We have used this test to detect every clinically relevant mutation present in ctDNA from cells all over the body. Just a few years ago, this type of analysis would have pertained to the realm of science fiction.

Being able to detect the entire range of mutations could be of great help in planning therapy. If ctDNA sequencing can detect all mutations present in the different cancer cells from a patient, we have a better chance to select the optimal drug or combination of drugs for that particular patient.

How could analysis of ctDNA be used to monitor disease progression and treatment?

As with CTCs, there have been studies showing that disease progression can be tracked by monitoring ctDNA levels, particularly by how concentrated the ctDNA is in the blood. As cancer progresses, the levels go up. In addition, in the last two years, research here at MSK and elsewhere has shown that next-generation sequencing of ctDNA from the plasma of breast cancer patients could identify genetic alterations that arise at different stages of the disease and potentially tell us whether a tumor will recur.

When it comes to treatment, we are beginning to see that genomic analysis of ctDNA over time makes it possible to track how cancer cells evolve in response to therapy. We recently published a study demonstrating this approach in a single patient with advanced breast cancer that had spread to the liver and bone. Genetic analysis from the primary tumor had revealed a mutation in a gene called AKT, so the patient was enrolled in a clinical trial testing a targeted therapy. We also sequenced DNA from the liver metastasis and from plasma samples collected both before the patient started the trial and at multiple points after receiving the drug.

We saw that the mutations in the cancer cells changed as the patient first responded to the drug and then relapsed. The changes in genetic composition mirrored the response to the drug as indicated by PET scans of the tumor and metastases. Significantly, however, we detected increases in the amount of ctDNA in plasma that preceded the detection of disease progression by imaging analysis or biochemical tests.

Although this analysis was done in a single patient, it does establish an important proof of principle that we might use this technique to track new mutations as they arise in response to targeted therapy. Analyzing ctDNA could give us a fuller picture of disease progression and drug resistance.

What applications do you see for this technology in the near future?

The use of ctDNA from plasma is still in its infancy. We can’t lose sight of all that we don’t yet know. With circulating tumor cells, a lot of initial hopes have not yet been fully realized; however, the experience accrued with the analysis of CTCs may prove instrumental for the use of ctDNA in clinical practice. The lessons we have learned over the last decade about the development of molecular tests for treatment decision-making, coupled with technological developments, are cause for optimism that this type of liquid biopsy will lead to noninvasive — yet highly sensitive — ways of detecting and monitoring cancer in the body.


When a cancerous tumor is treated and the ells die, they also release the Specific cell ctDNA and one might see arise in the ctDNA. Would there then be an expected bump in the ctDNA signifying death rather than and increase in the number of cancer cells? Would this technology also work for lymphomas and noy just solid tumors?

Raul, thank you for your comment. We passed on your questions to Dr. Reis-Filho, and he responds:

1. When the cancer cells die, there is an increase in the amount of tumor DNA in plasma; however, the DNA from a cancer cell dying will remain for a limited time in the circulation. For example, after chemotherapy, there is a short-lived increase in tumor DNA in plasma.

2. In a way akin to solid tumors, tissue sampling can be problematic for some types of lymphomas. In these cases, the analysis of ctDNA in plasma may offer a minimally invasive approach to assess the molecular characteristics of those cancers.

In past four years i have TURBT, breast cancer,lung cancer,malignant melanoma. My question is:'How it can be with one person?'

Dear Lerner Larisa, We are sorry to hear that you've been through so much. There are many reasons that people are diagnosed with more than one cancer, including for genetic reasons. If you would like to make an appointment with one of our genetic counselors for a consultation, please call 646-888-4050. To learn more about our support resources, please go to Thank you for reaching out to us.

It is nice to know that someone had discover new methods to detect and cure cancer. If it is really work for the best of the patient, then why not use it. I know that's going to be beneficial.

How would this technology work for monitoring the progression of the cancer - as opposed to dead cells? Is it applicable only in vivo or can it be used in vitro? What specific attributes of the disease can be monitored?

James, thank you for your comment.

As mentioned in the story, progression of the cancer could be tracked by monitoring ctDNA levels, particularly by how concentrated the ctDNA is in the blood. As cancer progresses, levels of the ctDNA go up. In addition, in the last two years, research here at MSK and elsewhere has shown that next-generation sequencing of ctDNA from the plasma of breast cancer patients could identify genetic alterations that arise at different stages of the disease and potentially tell us whether a tumor will recur.

Regarding the use of ctDNA to monitor specific attributes of the disease, the technique has the potential to track new mutations as they arise in response to therapy. Once adequately validated, ctDNA analysis will likely help us determine which populations of cancer cells are resistant to therapy and their genetic characteristics.

When we talk about capturing enough of ctDNA in plasma, accepted that this will indicate a pharmacogenomic understanding. But would it not be late for giving the medication or too late for patient to start responding to the treatment as a result of such a time when detection is made.

Venkatesh, thank you for your comment. We consulted with Dr. Reis-Filho, who responds: "The advantages that ctDNA analysis provides include the ability to identify recurrences before they are clinically or radiologically apparent, and to define the genetic changes present in the cancer cells. This information may ultimately be helpful for the therapy to be optimally tailored."

How do we know if the DNA in the circulation comes from the cancer cells or other tissues? If there are some mutations in other tissues but not the cancer cells, it would make the treatment plan more complicated. For example, we detected both KRAS mutation and EGFR mutation in patient but actually he only got KRAS mutation in the lung. Should the doctor give the patient EGFR-TKI treatment?

Thomas, we sent your questions to Dr. Reis-Filho, who responded:

Although mutations that were not inherited may also be present in noncancerous cells, the number of noncancerous cells harboring a given mutation is often low. Given that ctDNA correlates with the amount of cells harboring a mutation, these alterations present only in noncancerous would not be readily detected.

ctDNA is a promising tool for monitoring disease burden and to characterize the mutations present in a cancer, however at the moment, it is a research tool. Critical decisions about the therapy a patient will receive should not be based on ctDNA analysis at present; additional validation of the clinical utility of this approach is still required.

Thank you for your comment.

How important will mutation amplification such as ICE COLD-PCR and allele - specific PCR in analysis of cfDNA?

Peter, we sent your question to Dr. Reis-Filho, who responded:

These techniques have higher sensitivity than traditional PCR, however they require prior knowledge of the mutations a cancer has; hence, their use is limited to characterize the constellation of mutations in a cancer based on ctDNA analysis. They may be of help for disease monitoring on the basis of ctDNA, if the tumor has already been sequenced or if a tumor type harbors a given mutation in the vast majority of cases. Techniques with similar or even higher sensitivity and that allow for the characterization of a large number of mutations affecting many genes in a single experiment are currently being tested for the analysis of ctDNA.

Thank you for your comment.

The author states that the use of the MSK-IMPACT™, test which can look for hundreds of cancer-related mutations at once has been used to detect every clinically relevant mutation present in ctDNA from cells all over the body."The author later states that,"mutations in the cancer cells changed as the patient first responded to the drug and then relapsed," we know that each treatment regime causes further mutations in the ctDNA, and in the patient as a whole, it seems to be important for MSK, to measure mutations before, during treatment, and post treatment and relevant to measure mutation appearing after each treatment. So we can evaluate which therapies work best, based on disease state mutations and during treatment mutations, so that we can adapt therapies to best treat the patient.

I don't want to discourage anyone or undermine these line of investigations, but I think these all studies are more towards finding "a cure" for a more or less of a lost cause i.e disease is already metastatic or reaching metastasis. I agree either of methodology has its pros and cons; but in the end it gives us another "read out of progressed disease". My personal philosophy is based on "precaution is better than cure" (may be being a non-clinical researcher...). Thus the focus should be more on developing some kind of basic tests like a high levels of glucose in blood is a strong indicator that one "might" develop diabetes in future. So why not to try and find a similar "signature" by doing a blood profile from every healthy individual of a high risk group (smoker or red meat eater or sexually active individuals) as a mandatory once an year deal (may be start collecting and analyzing blood only after reaching puberty...) and based on that one may be able to find when exactly that small "blip" came on radar, which eventually lead to development of a cancer. And thus may help us far better in designing the "precaution strategies" and even help to find "the cure".

For the past four years I have been having problem with my thyroid , it kept growing nodules , I did the needle biopsy in 2012 and 2013 . The results showed no cancer . My throat became more enlarged ,and painful , with burning on the right side of my neck . I was constantly choking . July 21 , 2014 I had total thyroid ectomy . Pathology report showed cancer , /parathyroid cancer , I have to have the radioactive iodine treatment soon , I am so scared .

Zona, we're sorry to hear you're having these troubles. If you would like to make an appointment to see an MSK doctor, you can call 800-525-2225 or go to for more information. Thank you for your comment.

This is really a great discovery. when it will be commercially available?
This is helpful for the patients who live in resource poor area. I am doing my oncology practice at Jodhpur India for last 30 years. Patient do not want to go for Biopsy/ FNAC as tumor touch by prcedure spread it. This is a great myth.
please keep us updated about development in this field
All the best

Dr. Sanchetee, thank you for your comment. This technology is still in the investigative stage and must be validated with further research and testing, so it is difficult to tell how soon a clinical test will be commercially available.

Dear Dr. Reis-Filho,

Thank you for setting up this blog, it is very informative. I am a product development manager for Horizon Diagnostics and we are currently developing a range of ctDNA/cfDNA reference standards. I would really love to talk to you to get an insight of how our reference standards may be of use in your research, possible collaborations etc. When is the best time to get in touch with you?

Jessie, thank you for your comment, We will pass your message on to Dr. Reis-Filho.

This is wonderful to press for use of and education of the public with such technology for all patients. A key point of your description of ctDNA test utility in combatting one's tumor evolution for a single patient is tumor banking for these analyses prior to treatment in a trial or even for "standard" treatment. This would be analagous to culturing an infection at the outset.

Add a Comment

We welcome your questions and comments. Because this is a public forum, please do not include contact information or other personal details. Also, keep in mind that while we can provide general information and resources, we cannot offer personal medical advice. To make an appointment with one of our experts, contact our Physician Referral Service at 800-525-2225 or online.
Your e-mail address is kept private and will not be shown publicly.

More information about formatting options