Circulating tumor cell test shows clinical mettle

May 2006
Feature Story

Anne Paxton

If you think of magnetic nanoparticles zeroing in to capture and count, out of billions of cells in a 7.5-mL blood sample from a cancer patient, the exact number of circulating tumor cells—whether that number is zero or 700 or in between—you have an insight into why Arthur Clarke wrote that “any sufficiently advanced technology is indistinguishable from magic.”

The assay that accomplishes this feat is the CellSearch Circulating Tumor Cell Kit, sold by Veridex, and it was no conjuring trick to develop it. Officials with manufacturer Immunicon Corp., Huntingdon Valley, Pa., say the process took 10 years and $120 million.

Because circulating tumor cells, or CTCs, in peripheral blood are few and difficult to isolate, Immunicon’s technology was a major technical advance. But something that is just now coming into focus may constitute the real magic of the circulating tumor cell test. It is the way the assay’s results might be used to lengthen the lives of patients with metastatic breast cancer.

Since the findings of a clinical trial on CTCs were published in the New England Journal of Medicine in 2004 (Cristofanilli M, et al. N Engl J Med. 2004;351:781–791), and the Food and Drug Administration granted Veridex pre-marketing approval for CellSearch the same year, the strong link between CTCs and breast cancer patients’ overall survival rates is becoming more widely accepted. “CTCs are the strongest independent prognostic factor for survival of metastatic breast cancer patients ever found,” asserts Immunicon president and CEO Byron Hewett.

The research to date has repeatedly confirmed that CTCs are a potent prognostic tool. Among the most dramatic examples: Overall survival for patients being treated for metastatic breast cancer is about 11 months if they have five or more CTCs and 23 months if they have fewer than five CTCs—so they essentially have an additional year of survival if they do not have CTCs, says Howard Robin, MD, medical director of laboratory services at Sharp Memorial Hospital, San Diego.

New analyses of the data and new clinical trials are indicating how oncologists can apply that information to prescribe the best treatments for breast cancer patients.

“There has been a growing interest in this test over the last two years,” says Massimo Cristofanilli, MD, associate professor in the Department of Breast Medical Oncology, University of Texas MD Anderson Cancer Center, Houston, which was the lead institution in the 2004 New England Journal clinical study.

Quest Diagnostics has been offering the test for more than a year and a half, and as of Dec. 31, 2005, Immunicon had shipped 47 CellTracks Analyzer II or CellSpotter Analyzers, and 42 CellTracks AutoPrep Systems. Immunicon has a strategic alliance with Johnson & Johnson, which through its subsidiary Veridex markets, sells, and distributes CellTracks technology as part of the Veridex CellSearch system.

Among clinical researchers, there isn’t unanimity about using the CTC test to decide when to switch therapy for women with metastatic breast cancer. At least some have questioned this application, and many oncologists say they are not ready to abandon their traditional treatment approaches.

But increasing numbers of oncologists are requesting the test—and patients are too, Dr. Cristofanilli says. “When I talked about it in the past, not many people were aware of it, but now they’re trying to find which will be the best use for it.”

MD Anderson Cancer Center acquired the Veridex CellSearch system for the clinical laboratory and has since instituted the CTC test as a routine chargeable test procedure, says Herbert A. Fritsche, PhD, professor of laboratory medicine and chief of the clinical chemistry section at MD Anderson. “Our volume right now for clinical purposes is somewhere between 10 and 15 samples per week,” he says.

“At this stage the data seem to be upholding the prognostic value of the test. But we’ve spent the last year running patient data, and it will take a couple more years of followup before you can really address whether patients are living longer.”

Circulating tumor cells can be detected in several ways, explains Daniel F. Hayes, MD, clinical director of the Breast Oncology Program at the University of Michigan Comprehensive Cancer Center, another participant in the New England Journal clinical trial. Most of the methods are hospital-developed home-brew tests that do not cross state lines and do not require FDA approval, he notes.

“One way is by separating the CTCs from white cells based on their physical characteristics. They’re bigger and heavier, so you can do density gradient separation or filtration. The second way is by performing RT-PCR for transcription factors that don’t belong in the blood; in this case you’re looking for epithelial tissue that shouldn’t be there.”

“The third way is to use immuno-magnetic separation, as the CellSearch does. The difference between CellSearch and all the others is that it is a highly automated, highly reproducible assay.”

All of the CellSearch 510(k)s were cleared by the FDA in 2004, and the first commercial shipment of the system was in October 2004, Hewett says. More applications are pending; Veridex just received another 510(k) clearance from the FDA for use of the Linux operating system for the CellTracks Analyzer II. “We pursued that to use it as a sort of next-generation software platform to allow us to introduce a number of features that will enhance ease of use and give customers more flexibility.”

But Hewett says one of the most important steps forward since the 2004 trial was the FDA’s 510(k) clearance in October 2005 of additional data proposed for the CellSearch package insert.

“The claim we have is quite broad,” he explains. “It says patients with five or more CTCs have shorter progression-free survival and shorter overall survival. But that information alone does not tell the physician how to use the test. We did further analysis of the data and noticed that patients with fewer than five CTCs at all points had much longer survival than those with persistent circulating tumor cells. And that’s been added to the package insert.”

Also in the insert is a “revealing” chart, he says, titled “A reduction in CTC count below 5 after the initiation of therapy predicts longer overall survival.” It breaks the patients into four groups: green for those with fewer than five CTCs at all time points, blue for those with elevated counts at the start but lower counts later, red for those with elevated counts that persist, and amber for patients that start low and go to a higher cell count. (See chart, this page.)

“On the green line, we had a median survival of 22.6 months, while the red line had a median survival of only 4.1 months. This is a stark difference, but what’s most significant is those patients that responded to therapy, the blue line, had a survival not statistically different from those that started with fewer than five CTCs: 19.9 months. Both the blue and green lines had almost a fivefold survival advantage over those with high CTC counts.”

This tells the physician, Hewett says, that elevated cells, as a strong indicator of survival, “identify patients on treatments that may not be working and the physician should consider introducing other therapies.”

“With CTCs, at the end of the first cycle you can see if a therapy is having the desired effect, while with imaging studies you have to wait considerably longer, possibly months, to evaluate what’s going on. The other major difference between CTCs and imaging studies is that imaging reflects tumor burden, while CTCs are an indicator of how active and aggressive the disease is. That’s information the physician never had before,” he says.

“We found this to really resonate with oncologists. They can use this test to predict an outcome for the patient literally after each cycle of therapy.”

Tumor burden is less informative, he says. “Many oncologists tell us they have some patients with a relatively large tumor mass who can live a long time, while others have a small mass but all of a sudden there are metastatic sites all over their body. People know the key is the metastatic process, not the tumor.”

CTCs provide a “real-time” biopsy of a patient’s cancer, Hewett says. “After removal of the primary tumor the only potential sources of tumor tissue are metastatic sites, and biopsying these areas involves significant risk. But CTCs give you a glimpse into what’s going on with that specific patient’s cancer today.”

As it continues to explore applications of CTCs to cancer treatment, Immunicon is researching the development of tumor phenotyping and genotyping reagents for HER2/ neu, EGFR, Bcl-2, MUC 1, and others, Hewett says. “The idea is to look at the protein and gene expression on the CTCs. We are designing trials now to show that CTCs may be used to help select therapy for patients because they provide a real-time picture of what’s going on with the disease.” (Veridex last fall launched the HER2, EGFR, and MUC 1 CellSearch Tumor Phenotyping Reagents for research use only.)

Discussion about possible correlations between circulating tumor cell levels and survival time of metastatic breast cancer patients began in 1999, Dr. Cristofanilli says. “It was the first time we had the possibility with a blood test to be able to have a more rational discussion of treatment with the patient, and to develop some therapies that would be more effective” than the previous standard approaches.

And since the initial study released in 2004, “we’ve been expanding these observations and confirming the prognostic aspect of the test, which strikes me as a very significant factor in patients with metastatic disease.”

Dr. Fritsche says circulating tumor cells are seen in about 50 percent to 60 percent of patients at the time they have documented development of metastatic progression. The CTC test is qualitative in nature, he stresses, and has very good reliability. “We’re not talking about using a test in a serial fashion where a 50 percent change in the marker might indicate a tumor is worse or better. You can’t look at a change from five to 50 cells as having any significance. We are basically looking at whether there are more or less than five cells present,” Dr. Fritsche says. With their test method, they’re able to determine consistently with precision whether cells are absent or present defined by the five-cell cutoff value.

In Dr. Fritsche’s experience, the test is also reliable in terms of biological variability. “Do patients with fewer than five cells come back with more than five cells? No, we do not see that.” In the control subjects they have tested on multiple occasions, the results remain negative. “But in patients undergoing therapy, changes from negative to positive suggest progression of disease, and changes from positive to negative suggest a response to treatment.” In those CTC-positive subjects who have greater than five cells, Dr. Fritsche says, biological variation may contribute to the variability in cell numbers observed in subsequent measurements.

The assay started as a “mom and pop” test, very hands-on, “with several cycles of manual washing required,” Dr. Hayes says. “But it was improved over time and eventually changed into what it is now a ‘black-box’ assay that basically just requires sticking blood in the machine, and in a couple of hours you receive a readout.”

Immunicon realized that circulating tumor cells are fragile and do not survive well outside the body. Therefore, the test developers combined a cell fixative with an anticoagulant in a vacuum draw tube called CellSave. The preservative keeps the cells intact for a minimum of 72 hours so the sample can be shipped around the world as needed. CellTracks’ AutoPrep instrument takes the tube, reduces it to a 320-µL aliquot, and places it into a cartridge, which is inserted into the CellTracks Analyzer II.

Images are then presented on a computer monitor where they can be reviewed by a technologist or pathologist, producing a reproducible assay that is “really from A to Z pretty self-reliant,” Dr. Hayes says.

The key study published in the New England Journal in August 2004 was unusual in being a prospectively written protocol, he notes. “It’s one thing I am probably the most proud of. I’ve written several papers through the years regarding the design and evaluation of tumor markers, and I’ve made the point that most of the studies are ‘studies of convenience’—they’re put together without a lot of the controls and clinical trials that we perform for a new drug. But that wasn’t the case with this test. It was a prospective, multi-institutional trial, funded by Immunicon.”

The expense of such trials means they seldom materialize. But one of the reasons so few tumor markers have been widely accepted, he points out, is because the data are so unreliable. “There’s a new tumor marker a week published in various journals, and they all end up with P-values of .03, but they are not directed toward understanding how to use them in the clinic. A lot of the things that get published never actually make it into routine clinical use because no one can figure out how to use them.”

In the past, clinicians at MD Anderson would have used radiographs to determine if a particular therapy for metastatic breast cancer was working, Dr. Fritsche says. “But we would almost have to require completion of the treatment before there was radiographic evidence that the tumor was not responding. If you’ve given three or four courses of a toxic chemotherapy, the patients are likely not going to do as well as they could have if given the most appropriate therapy to begin with.”

Since several active chemotherapy agents are available to treat patients with metastatic breast cancer, the challenge is picking the therapy most likely to work with the fewest side effects. “It’s not always true that you have to wait,” Dr. Hayes says. “For example, if the patient has horrible pain and then comes in pain-free, you did the right thing. Or if there’s nothing palpable on the chest wall, then the patient has lumps and bumps,” a change might be in order. “But those are the extremes. More commonly, using conventional radiographs, it takes two or three months to figure out if a therapy is working or not.”

There are clinical and financial price tags attached to using the wrong therapy, he notes. First, it means something else that might work is not being used. Added to that are the toxicity and possible side effects of the therapy, and finally the cost. “Chemotherapy is not cheap,” he says, suggesting that the roughly $600 cost of a circulating tumor cell test may easily be justified if it saves a patient $5,000 in chemotherapy costs.

Many findings of the New England Journal study have generated provocative questions that researchers are now pursuing, Dr. Hayes says. For example, the full set of data on median time to progression of disease suggests that patients who still have five circulating tumor cells after three to five weeks of therapy have a 50 percent chance of showing classic signs of progression in the next month, although 10 or 15 percent of those patients did not progress rapidly.

“So these data suggest that most of the patients who continue to have elevated cells after a single cycle of chemotherapy are probably on inactive treatment. The question is whether these patients benefit from changing therapy right then, or whether they are patients for whom nothing works. And, of course, that small group of patients with elevated CTCs at this first followup time period who don’t progress rapidly probably would not benefit from an early change in therapy.”

Through the Southwest Oncology Group and the Breast Cancer Intergroup of North America, a cooperative of research institutions, Dr. Hayes and his colleague, Jeffrey Smerage, MD, PhD, are spearheading a trial funded by the National Cancer Institute and scheduled to start later this year that will address some of these issues. Unlike the New England Journal trial, which looked at a more heterogeneous group of patients with metastatic disease, the new study will look only at patients who will be starting a first-line chemotherapy for their metastatic breast cancer.

“They will be starting first-line chemotherapy and receiving whatever chemotherapy the doctor thinks is appropriate,” he says. “CTCs will be analyzed at baseline and at first followup, and if they have elevated cells at first followup they will be randomly assigned to remaining on the first-line chemotherapy chosen by their oncologist—which is the standard of care without CTC data—or to changing to a different therapeutic regimen—again, left to their physician’s discretion. The idea is to see if switching chemotherapy early makes sense.”

When the necessary reviews are completed, Dr. Hayes says, “we hope we can actually start putting people on the study this summer.” The study has generated excitement because it will test a new way to use a tumor marker. Markers like CA 15-3 and CEA, he says, have been considered inaccurate indicators at first followup, because 25 percent of patients treated with new chemotherapy will have the level of these “soluble” protein markers spike—become falsely elevated—before it goes down, probably because the therapy is killing cells that release the antigen. “We think that unlike circulating markers like CA 15-3, CTCs don’t go through this spike. That’s the basis of the next randomized trial.”

Several other developments have taken place since the New England Journal study. The investigators in the initial Immunicon trial are now working on sub-analyses of that study to compare the accuracy of CTC levels versus assessment of classic clinical response to predict overall survival. Thomas Budd, MD, of the Cleveland Clinic is leading this analysis. Says Dr. Hayes: “With patients with non-measurable disease—one-third to one-half of metastatic breast cancer patients—it’s notoriously difficult to decide if the patient is progressing or not. Many patients with metastatic breast cancer have it principally in their bones, and it’s harder to a get a handle on the cancer.” Lung metastases, for example, can be measured in two dimensions, but radiographic studies of bone often do not show the cancer. Rather, they show cancer-related epiphenomena happening around the cancer, like the appearance of lysis or sclerosis associated with osteoclastic and blastic activity.

An interesting finding so far is that CTCs were poor predictors for clinician-determined progression. “But they were actually better predictors for survival than was clinician-determined progression. In other words, clinicians’ ability to determine progression is very bad when compared to the gold standard, whether the patient is alive or dead,” Dr. Hayes says. CTCs, however, were more likely to predict survival accurately.

Most patients with metastatic breast cancer have subtle symptoms, he points out. “Ninety percent of them have their cancer internally, so there’s nothing to feel or examine. And they’re usually older women with arthritis or other conditions that can mimic cancer symptoms.”

He is using the assay now to help decide if he should change therapy after several months of a therapy in patients for whom other parameters, such as history, physical exam, other markers, and radiographs, are uncertain. “I am, of course, very excited about our prospective randomized trial to determine if CTC levels at the very first followup time period after a patient starts chemotherapy will direct a change that improves a patient’s outcome.” He says he personally would not use CTC levels now at baseline to make decisions, though Dr. Cristofanilli is designing studies to determine whether the test might be used to direct “more or less ‘aggressive’ therapy right up front.”

Other uses for the CTC test are taking shape. “We’re talking right now about metastatic breast cancer,” Dr. Fritsche says, “but there has been a shift to using this test in two other opportunities. One as an adjuvant treatment, after a primary treatment of either surgery alone or surgery with radiation, followed by an anti-hormonal treatment like tamoxifen or chemotherapy. In the adjuvant setting, you’d be treating patients with early-stage disease to postpone or reduce their chance for recurrence and metastasis.”

Neoadjuvant use would be another. “That’s chemotherapy before you have surgery, the primary treatment. The reason you do that is to shrink the tumor to make surgery more effective,” Dr. Fritsche says. “So we have protocols in place where we’re trying to identify who needs more than the current level of treatment at adjuvant and neo-adjuvant settings.”

“We know in both breast and colon cancer that 20 percent to 30 percent of cases supposedly cured by surgery do develop metastatic disease. Most people believe that metastasis occurs through bloodborne CTCs, and that’s why the CTC test is so important in early-stage cases.”

Such adjuvant therapy might be particularly helpful in colon cancer, Dr. Fritsche notes. “Colon cancer most often metastasizes to the liver, so if you have evidence there is some disease you can maybe find earlier, you can do a hepatic resection.”

In early prostate cancer there do not seem to be many patients who have CTCs. “With metastatic disease there does seem to be a large number with CTCs,” he says. “However, successful chemotherapy for prostate cancer is really only at the early stages now, so we don’t have the wide variety of treatment options that we have in metastatic breast cancer. So while there is interest, there’s probably much less benefit.”

The early data submitted to the FDA implied that CTCs have a valuable prognostic use in managing patients with late-stage disease. “Our current data is substantiating that,” Dr. Fritsche says. In his view, now might be the time for oncologists to start getting their own level of confidence and experience with this test.

One problem with testing for CTCs in the past was the lack of a reliable, robust assay technology. Though there is a continuing need for better cancer markers, Dr. Fritsche says, “what we have now is a completely automated process for making the observation. So for the first time we have a way of generating results across the U.S. and providing a level of service that physicians could be comfortable in using. A lot of new technologies out there don’t promise you that.”

Dr. Cristofanilli hopes that other clinical trials will help clarify the appropriate uses of CTC testing in deciding on therapies. “Many times a patient will have a minimal response to a therapy and we don’t know exactly what to do. But eventually the interpretations of CTC test results may lead us to make changes in treatment. One day we’ll get more confident in this test to let it be really useful in moving us slowly away from expensive therapies.”

Dr. Robin at Sharp Memorial Hospital recently received training on the CellSearch system and says one has been installed in his laboratory for research and clinical services. “Oncologists on the East Coast have been using the test for about a year and believe it has been very helpful in identifying when drugs are not working,” he reports.

Potentially, the CTC test could supplant imaging studies that are used to gauge whether a therapy is working, he says. “Rather than measuring the size of tumor at baseline, then at one month, two months, and three months, which means significant exposure to x-rays as well as the costs of the studies, this is a simple blood test that may demonstrate effectiveness of therapy earlier than the imaging studies.”

The CellSearch system enriches and stains CTCs and then presents the images to a pathologist, who then interprets them as being circulating tumor cells or mononuclear cells or unidentified cells,” Dr. Robin says.

“It’s a really intriguing process and different from any other test performed in our laboratory,” he adds. “The iron-coated antibodies react with epithelial cells, the specimen is placed in a magnetic field that enriches for tumor cells now covered with antibodies, magnets hold the CTCs to the side of the container, and the pathologist enumerates the cellular images on the computer screen.”

“The pathologist now plays an integral part in helping the oncologist choose the right therapy for his patient, and the right patient gets the right drug at the right time, which hopefully will reduce the use of expensive chemotherapeutic agents as well as reduce the significant risk they present.”

But many other factors will decide whether the CTC test gains traction in the medical oncology community, Dr. Robin says. “Medical oncologists are concerned about what to do with the results if it appears the patient is not responding to therapy. Should they change the medication based on the laboratory test? They will have to get over their initial skepticism and determine if the CTC test helps them identify optimal therapy for their patients.”

One oncologist he has spoken to said the test was helpful in a patient who developed an adverse reaction to a particular chemotherapeutic drug but didn’t want to stop the drug if it was effective for his patient’s tumor. When a followup CTC assay showed it was working, he kept her on the drug. On the flip side, if a woman had five or greater CTCs before therapy, and they did not decline after therapy, “Why subject her to chemotherapy? Why not change therapy or stop the drug?”

That will be a subject of ongoing debate as use of the test expands, Dr. Robin predicts.

Related Links

• image (PDF, 6,755 K)