Liquid biopsies provide extensive information, but technology carries ‘potential for overuse’

Researchers have known for decades that tumors shed cells and DNA into the bloodstream.

Recent technological advances finally have enabled researchers to capture these fragments through the so-called “liquid biopsy.” The simple blood draw yields a wealth of data that can help guide treatment decisions.

“Liquid biopsies consisting of either circulating tumor cell-associated biomarkers or circulating tumor DNA both can give extensive information about the biology of the disease,” Massimo Cristofanilli, MD, deputy director of Translational Research at the Kimmel Cancer Center and director of the Breast Cancer Center at Jefferson University Hospitals, told HemOnc Today. “Particularly in advanced disease, they can help monitor response to therapy by assessing changes in genomic abnormalities or even potentially help in selecting molecular treatments. This is a real-time assessment of the disease biology that can be repeated in a longitudinal way.”

Although liquid biopsies primarily have been utilized in metastatic disease across solid tumor types, new findings have shown promise for their use in early-stage disease. Some researchers suggest liquid biopsies eventually could be used to screen asymptomatic individuals at high risk for cancer.

Despite the optimism, concern remains that researchers’ abilities to perfect this approach might outpace the clinical community’s ability to apply the resulting knowledge to practice.

“Several of these technologies are trying to compete with each other without really understanding that the real question should be how to properly use the information obtained with these tests for patient management,” Cristofanilli said. “Researchers have the opportunity, very early on in the technology development of liquid biopsy methods, to focus on the fundamental questions of clinical utility that need to be addressed in order to include them in the standard management of patients. Academia has to be part of this development, and not left behind.”

HemOnc Today spoke with several researchers about the advantages of liquid biopsies, how they can complement standard tissue biopsies and the controversy that surrounds their potential use as a tool to screen healthy individuals for early signs of a variety of malignancies.

Blood vs. tissue

Klaus Pantel, MD, PhD, and Catherine Alix-Panabières, PhD, coined the phrase “liquid biopsies” in 2010 to describe blood draws that evaluate circulating tumor cells (CTCs). Since then, the term has evolved to include circulating plasma tumor DNA.

Multiple studies have demonstrated enumeration of CTCs can predict prognosis, while genotyping CTCs and circulating plasma DNA can identify targetable mutations and detect recurrence.

Nearly 50 companies are creating technologies to assay CTCs, and 10 are in development for circulating DNA.

Liquid biopsies are cheaper and less invasive than traditional tissue-based biopsies, which — particularly for patients with metastatic disease — can be complex and often are associated with complications.

“The traditional biopsy requires a certain amount of tissue that is not always feasible to collect,” Cristofanilli said. “Although indicated for the majority of patients with advanced disease, in some patients the liquid biopsy can be the only way to interrogate the tumor because there is no possibility to perform a traditional biopsy, such as in patients with bone metastases.”

Researchers can repeat liquid biopsies serially — a practice that increases the risk for complications in traditional biopsies — and receive real-time results.

“If you can get a rapid result on a blood test instead of spending weeks collecting a tumor specimen and processing it, it is potentially very clinically practical,” Geoffrey R. Oxnard, MD, a thoracic oncologist at Dana-Farber Cancer Institute, said in an interview. “Currently, for an oncologist to get a tumor genotype, they have to dig up a specimen that may have been collected a month ago, request the tissue and have it sent to another lab to be processed, analyzed and reported back. The blood test can be performed immediately in the lab, obviating some of the logistical challenges of finding archived tumor and getting it processed appropriately.”

Traditional biopsies for metastatic disease not only cause complications and pose logistical challenges, but the results may not capture the heterogeneity of the disease.

“In metastatic disease, we already know there is great tumor heterogeneity — not only between patients and types of tumors, but even within the patient,” Ben Ho Park, MD, PhD, associate professor of oncology at Johns Hopkins Medicine, told HemOnc Today. “We’re trying to prove right now that blood-based biopsies can reliably detect a spectrum of mutations that accurately reflect at least one site of metastatic disease, whereas a traditional biopsy only informs us about a single site.”

Despite the advantages of liquid biopsies, researchers question whether they can replace tissue-based biopsies.

One concern is that liquid biopsies may never reach 100% sensitivity.

“Some tumors shed DNA while others don’t. If your tumor doesn’t shed DNA, we are not going to be able to analyze the tumor’s biology through a liquid biopsy,” Oxnard said. “I see liquid biopsies as a convenient step on the way to conventional biopsy, perhaps eliminating a subset of biopsies. But in the end, conventional biopsies are going to be essential.”

Park suggested liquid- and tissue-based biopsies may complement each other as researchers identify different mutations in metastases through the conventional approach and then track them continually through the blood-based approach.

Yet, Cristofanilli said blood-based diagnostics might replace tissue-based diagnostics in the near future.

“The blood can provide a variety of different information that the tissue — in a static, one-time collection — cannot,” Cristofanilli said. “This is not science fiction, it’s reality. In my opinion, the blood-based testing will replace the tissue-based approach. It remains to be seen how long it’s going to take and how we’re going to get to that point.”

Utility of CTCs

Cristofanilli and colleagues were the first to demonstrate in a prospective trial that CTCs offer prognostic value in patients with metastatic breast cancer.

The study showed patients with ≥5 CTCs per 7.5 mL of whole blood at baseline had shorter median PFS (2.7 months vs. 7 months; P<.001) and OS (10.1 months vs. >18 months; P<.001) than patients with baseline CTC counts <5 cells per 7.5 mL of whole blood. Failure to clear CTCs after treatment was associated with worse outcomes.

Subsequent studies demonstrated similar results in patients with colorectal and prostate cancers, yet many clinicians remained unconvinced of the approach’s clinical value, Daniel F. Hayes, MD, professor of internal medicine and clinical director of the Breast Oncology Program at the University of Michigan Comprehensive Cancer Center, said in an interview.

“Because of this skepticism, we tried to conduct a study we hoped would actually convince others that enumerating CTCs had clinical utility,” Hayes said. “In breast cancer, there are at least 10 or 15 chemotherapies that work now, and we’re trying to pick them empirically in a given patient. Our thought process was that, if CTCs indicated you picked the wrong therapy, maybe a different one would work instead.”

This analysis — published in June in Journal of Clinical Oncology — included 595 patients with metastatic breast cancer. Researchers calculated patients’ CTCs at baseline and after 21 days of therapy. Patients who had elevated counts at baseline but demonstrated lower counts at first follow-up remained on therapy, whereas those with persistently high counts switched therapies at that time.

Hayes and colleagues observed similar median OS among patients who continued on the same therapy and those who switched treatments (10.7 months vs. 12.5 months; P=.98).

“On the one hand, we were disappointed [because] patients did not benefit from switching therapies,” Hayes said. “On the other hand, this study is quite important because it tells us that having these cells in your blood is a lot different from protein markers. These cells are biologically important, and not clearing them after one cycle of chemotherapy resulted in an awful prognosis.”

Overall, researchers reported significantly shorter median OS (13 months) among patients who had increased CTCs after first-line chemotherapy than patients who did not have increased CTCs at baseline (35 months) and those whose CTC counts improved after therapy (23 months; P<.001). Among patients who continued to have high CTC counts after initial chemotherapy, nearly 50% had died by 12 months, and about 75% had died by 18 months. These results were consistent across disease subtypes.

“This tells us that we can use CTCs to identify patients who should receive investigational therapy instead of one chemotherapy after the other,” Hayes said.

CTC research is moving beyond enumeration to explore more in-depth questions about the disease through genotyping.

“We can use CTCs to learn about the ongoing genetic changes that are occurring, such as if a patient is developing resistance to therapy,” Terence Friedlander, MD, a genitourinary cancer specialist at the University of California San Francisco Medical Center, told HemOnc Today. “Radiographically, we may not know about the development of treatment resistance for a long time, because these treatment-resistant cells will not be radiographically detectable until there are thousands or even millions of resistant cells.”

A study by Antonarakis and colleagues, presented at ASCO, sought to detect the androgen receptor splice variant-7 (AR-V7) — which confers resistance to hormonal therapy — in CTCs of 62 men with metastatic castration-resistant prostate cancer. Half of the men were treated with enzalutamide (Xtandi, Astellas/Medivation), and the other half received abiraterone (Zytiga, Janssen).

Men who were AR-V7–positive demonstrated significantly shorter median PFS than AR-V7–negative patients, regardless of treatment (2.1 months vs. 6.1 months for enzalutamide; 2.3 months vs. not reached for abiraterone; P<.001 for both).

No patients with AR-V7–positive CTCs demonstrated a PSA response, whereas 52.6% of AR-V7–negative patients treated with enzalutamide and 68% AR-V7–negative patients treated with abiraterone demonstrated a PSA response.

“These results are quite impressive and statistically very powerful. This is a very big step,” Friedlander said. “It moves CTCs away from a simple tool to assess prognosis to a much more powerful tool to predict which therapies we should use to benefit the patient.”

Circulating tumor DNA

The ability to identify and track mutations in the blood through circulating plasma tumor DNA, also known as circulating cell-free tumor DNA, is another area of promise for liquid biopsies.

Dawson and colleagues first provided proof-of-concept for circulating tumor DNA in a study published in 2013 in The New England Journal of Medicine. They assayed plasma samples from 30 women with metastatic breast cancer for cancer antigen 15-3 (CA 15-3), CTCs and circulating tumor DNA. All women had detectable somatic genomic alterations identified from sequencing tissue samples.

Researchers detected circulating tumor DNA in 29 women, CTCs in 26 women and CA 15-3 in 21 women. Circulating tumor DNA provided the earliest indication of treatment response in 10 of 19 women.

“Reaching for DNA is very practical. It is the stable molecule of life and it is relatively easy to handle, collect and process,” Oxnard said. “The most important biomarkers in solid tumor oncology are DNA biomarkers, and we can get at these same biomarkers using floating DNA instead of actually doing a tissue biopsy.”

The most rapid assimilation of liquid biopsies into clinical practice likely will occur in solid tumor types for which biomarkers currently exist, Oxnard said.

The majority of research in this area has focused on metastatic disease.

“In metastatic disease, sometimes the amount of tumor DNA is almost the same as the amount of total DNA because there is so much cell death and turnover of cancer cells, even if the patient is not undergoing therapy,” Park said. “That’s just the nature of cancer.”

A study by Bettegowda and colleagues, published in February in Science Translational Medicine, found 82% of patients with metastatic solid tumors outside of the brain had detectable circulating plasma tumor DNA. When stratified by tumor type, more than 75% of patients with metastatic ovarian, colorectal, bladder, gastroesophageal, pancreatic, head and neck, breast and hepatocellular cancers harbored detectable DNA, whereas less than 50% of patients with medulloblastomas or metastatic cancers of the kidney, prostate or thyroid demonstrated detectable DNA.

Tracking circulating DNA also can be used to detect resistance to therapies and assess evolving cancer biology.

“Liquid biopsies can be used predominantly in situations where we’re exploring how tumors have changed over time, where the primary tumor is no longer representative of what we’re treating,” E. Scott Kopetz, MD, PhD, FACP, associate professor in the department of gastrointestinal medical oncology at The University of Texas MD Anderson Cancer Center, said in an interview. “This is a general problem that we face in many tumor types. By the time we get to a later-line therapy and use a targeted agent, sometimes the target that we were trying to address is no longer there.”

Kopetz and colleagues utilized the Guardant360 (Guardant Health) assay to monitor for recurrence in 71 patients with metastatic colorectal cancer treated with an EGFR inhibitor.

Results of the study, presented at ASCO, showed 30% of patients developed KRAS mutations, 14% developed new EGFR mutations and additional 14% developed an increased copy number of MET.

“These results demonstrated that there are different ways cancer cells can develop resistance to EGFR inhibitors,” Kopetz said. “This is relevant because, when we understand how and why the cancer develops resistance, we can inhibit these particular new mutations.”

A study at MD Anderson, currently enrolling patients, is designed to do just that — assign patients to new therapies based on an understanding obtained through the plasma analysis of why patients are progressing. Yet, researchers disagree about whether this approach can be applied to early-stage disease.

“In a very acquiescent or early-stage, slow-burden cancer, I’m uncertain as to whether a liquid biopsy is going to be adequate,” Oxnard said. “We may need a new generation of this technology to push past where we are now to reliably and effectively assess low-burden cancer states.”

However, using liquid biopsies in early-stage disease has the potential to reduce overtreatment, Park said.

Park and colleagues evaluated data from 29 patients with early-stage breast cancer using the Droplet Digital PCR (Bio-Rad) platform. They evaluated pre- and post-surgery plasma samples for PIK3CA mutations and compared them with tumor sequencing analyses.

The results — published in Clinical Cancer Research — indicated 14 of the 15 PIK3CA mutations detected through tumor sequencing also were detectable in the pre-surgery plasma samples, equating to a 93.3% sensitivity and 100% specificity. Among 10 patients evaluated after surgery, five still had detectable mutations.

“The fact that we were still detecting mutations postoperatively really spoke to us,” Park said. “These women still had additional therapies to undergo, but now we can use this technology to follow microscopic residual disease to determine whether a woman is still at risk or if she has a better prognosis.”

In the future, this technology may be able to identify patients for whom further treatment is unnecessary.

“It’s going to take time to develop and prove that we can safely say patients who have nondetectable plasma tumor DNA after the primary therapy are OK, allowing us to reduce the risk for overtreatment,” Park said. “But that is the hope, and I think that is going to have an enormous impact on many solid malignancies.”

Screening for cancer

The possibility that technology may evolve to where liquid biopsies routinely detect cancer in asymptomatic individuals has triggered tremendous optimism.

For example, if a patient is known to be at high risk for pancreatic cancer, a malignancy in which RAS is frequently mutated, a liquid biopsy could be ordered to look for that mutation.

However, the use of liquid biopsies for this purpose is cause for significant concerns, including whether the knowledge yielded by CTCs and circulating DNA may raise more questions than researchers are able to answer.

“Using this in the general population will open up Pandora’s box,” Park said. “All of a sudden, we’ll detect things before they become clinically or radiographically apparent. What would you do with that information? This technology has potentials for overuse, and we have to be very careful and cognizant of what we’re trying to address as a field.”

If a marker is detected in the blood, researchers will not know the tumor site unless it is a very specific marker for that tissue, such as PSA for prostate cancer.

“As the technology changes, we’ll have a chance to do broad genomic screening in the blood,” Hayes said. “But it’s Star Wars stuff right now. It’s not reality.”

Further, screening tests require high specificity to reduce the risk for false-positives, and liquid biopsies currently do not achieve that level of specificity.

Friedlander referred to the 2004 study by Cristofanilli and colleagues, where the technology used has been shown to infrequently detect one to two CTCs in healthy control patients.

“Based on that, you might think that a patient with a detectable CTC therefore has cancer, and in fact those patients did not have cancer,” Friedlander said. “There needs to be more refinement of the technology in order to accurately screen patients.”

Despite these challenges, the potential exists that liquid biopsies may one day be used as a cancer screening tool.

“Early detection certainly is the dream, and that is where we eventually want to go with this technology,” Oxnard said. “It’s going to be a few years before we have an assay that has the appropriate degree of sensitivity and reliability and a low false-positive rate that will allow us to screen for cancer. I think there’s promise for this, but it’s going to be much more challenging than the initial goal of understanding cancer’s biology.” – by Alexandra Todak

References:

Antonarakis ES. Abstract #5001. Presented at: ASCO Annual Meeting; May 30-June 3, 2014; Chicago.

Beaver JA. Clin Cancer Res. 2014;doi:10.1158/1078-0432.

Bettegowda C. Sci Transl Med. 2014;doi:10.1126/scitranslmed.3007094.

Cristofanilli M. New Engl J Med. 2004:351:781-791.

Dawson SJ. New Engl J Med. 2013;368:1199-1209.

Klaus P. Trends Mol Med. 2010;16:398-406.

Morelli MP. Abstract #11117. Presented at: ASCO Annual Meeting; May 30-June 3, 2014; Chicago.

Oxnard GR. Clin Cancer Res. 2014;doi:10.1158/1078-0432.

Smerage JB. J Clin Oncol. 2014;doi:10.1002/JCO.2014.56.2561.

For more information:

Massimo Cristofanilli, MD, can be reached at massimo.cristofanilli@jefferson.edu.

Terence Friedlander, MD, can be reached at terence.friedlander@ucsf.edu.

Daniel F. Hayes, MD, can be reached at hayesdf@umich.edu.

E. Scott Kopetz, MD, PhD, FACP, can be reached at skopetz@mdanderson.org.

Geoffrey R. Oxnard, MD, can be reached at geoffrey_oxnard@dfci.harvard.edu.

Ben Ho Park, MD, PhD, can be reached at bpark2@jhmi.edu.

Disclosure: Hayes reports research funding from and royalties on patents licensed to Janssen Diagnostics. Park reports consultant roles with GlaxoSmithKline and Novartis, as well as scientific advisory board roles with Horizon Discovery and Loxo Oncology. He also is entitled to a share of the royalties received by The Johns Hopkins University on patents licensed to Horizon Discovery. Cristofanilli, Friedlander, Kopetz and Oxnard report no relevant financial disclosures.

 

 

Is circulating plasma tumor DNA the ideal target for liquid biopsies?

To the extent that somatic mutations are drivers of the disease and are molecularly actionable, circulating plasma tumor DNA offers an attractive approach to genotype noninvasively.

The diagnostic potential of circulating plasma tumor DNA (ctDNA) was conceived in the 1970s, when evidence of elevated levels of ctDNA in cancer patients was first reported by Leon and colleagues. For 40 years, ctDNA has not been clinically implemented — in part due to unrealistic expectations of total ctDNA concentrations as a specific marker of solid tumor, and in part due to lack of appropriate technology. Fast forward 40 years. We now live in a time in which genotype-directed therapy is the standard of care for advanced lung cancer, melanoma and colorectal cancer. Specific somatic mutations in EGFR, ALK, ROS1 and BRAF are predictors of the efficacy of small-molecule kinase inhibitors and are commonly used in either clinical practice or clinical trials. As these tumor-unique mutations are hallmarks that distinguish normal tissue from cancer cells, testing for these genotypes could allow for a uniquely specific approach for diagnosis and monitoring of cancer.

Technologies have emerged that allow quantitative genotyping of ctDNA with single-molecule sensitivity. We have recently witnessed several high-profile publications demonstrating the potential power of ctDNA genotyping. Diehl and colleagues examined the mechanism of resistance to EGFR inhibition by monoclonal antibodies in colorectal cancer. The investigators performed genotyping on ctDNA specimens from 24 patients, demonstrating that clinical resistance arose via point mutations in the corresponding KRAS that were detectable in ctDNA up to 10 months prior to radiographic progression. Similarly, Oxnard and colleagues demonstrated the presence of the EGFR activating mutation prior to EGFR therapy, its reduction or disappearance during therapy and re-emergence along with the drug-resistant EGFR T790M mutation during treatment in a ctDNA with high accuracy.

Although still early, the available data points to ctDNA as the molecule of choice when one studies validated oncogenic drivers — such as EGFR and BRAF mutations — given its inherent specificity, ease of collection, an analytical workflow that is easily transferable into the clinical setting from academic centers to community hospitals, sample stability and turnaround time from sample collection to results reporting. The time is ripe to transform ctDNA genotyping from a research tool into a clinical biomarker.

 

References:

Leon SA. Cancer Res. 1977;37:646-50.
Diehl F. Nat Med. 2008;14:985-90.
Oxnard GR. Clin Cancer Res. 2014;20.6:1698-1705.

For more information:

Cloud P. Paweletz, PhD, is head of the Translational Research Laboratory of the Belfer Institute for Applied Cancer Science at Dana-Farber Cancer Institute. He can be reached at Belfer Institute for Applied Cancer Science, 4 Blackfan St., Boston, MA 02115; email: cloudp_paweletz@dfci.harvard.edu. Disclosure: Paweletz reports no relevant financial disclosures. 

Circulating tumor cells provide more opportunities for molecular analysis.

Although analysis of circulating tumor cells (CTCs) and circulating plasma tumor DNA (ctDNA) often are viewed as competing emerging technologies, they are in fact highly complementary, and both are poised to revolutionize how we sample and monitor cancer. In essence, the appeal of these technologies stems from our increasing understanding about how tumor cells evolve and adjust in response to increasingly effective therapies: Cancer cells need to be monitored in real time to optimally direct targeted therapies and monitor their effectiveness.

Compared with free DNA in the circulation, molecular analysis of whole cancer cells presents powerful capabilities. Enumeration of CTCs is well correlated with clinical prognosis in patients with metastatic cancers of the breast, prostate and colon, and within individual patients, the trend in CTC numbers following therapy is indicative of treatment response or disease progression. Molecular characterization of CTCs includes analyses of protein and RNA markers in addition to DNA, thus enabling the application of diverse biomarkers — ranging from Ki67 (eg, proliferative index) to signaling outputs (eg, androgen signaling) and epigenetic changes (eg, epithelial-to-mesenchymal transition). Probably the most exciting emerging application of CTCs is their expansion in vitro, enabling detailed tumor cell genotyping and personalized drug sensitivity testing.

The primary area of overlap between analyses of CTCs and ctDNA is tumor genotyping, both the screening for driving mutations that may be therapeutically targeted at the time of diagnosis and the ongoing monitoring for the acquisition of drug resistance mutations that may be an indication for a switch to an alternative treatment. Both ctDNA and CTCs have low and variable purity even after optimal preparation, which complicates sequencing analysis. The advantage of ctDNA is that, although impure and contaminated with circulating free DNA derived from normal cells, the total amount of DNA available for analysis is high enough that it can be easily subjected to next-generation sequencing analysis.

Alternatively, advantages of CTC analyses include the fact that they allow sequencing of intact RNA (including tumor-associated translocations, which are not as easily measured by DNA sequencing) and that once optimized, sequencing of single CTCs will allow unparalleled analysis of tumor cell heterogeneity, something that can only be estimated by sequencing bulk DNA molecule preparations. In addition, although testing for known mutations (eg, BRAF mutations in melanoma) is the most straightforward application of ctDNA-based sequencing, absence of such a mutation is harder to interpret because it may either indicate a cancer with wild-type BRAF or insufficient amounts of tumor-derived DNA within the circulation. In contrast, initial isolation of CTCs followed by whole-genome amplification and next-generation sequencing will allow for optimal interpretation of a cancer cell’s mutational profile.

Ultimately, molecular analyses of CTCs and ctDNA have tremendous potential to provide detailed real-time information about the properties of cancer cells through noninvasive, easily repeated blood sampling. Specific tumor genotyping is likely to be the sweet spot for ctDNA measurements, whereas more comprehensive multiparameter studies will require CTC studies. The bottleneck for CTC analyses is the technology for isolation of rare cells from blood specimens. Although currently available commercial technologies are limited in sensitivity and analytic capabilities, the field is exploding with new and more powerful tools. These are likely to be optimized and broadly available within a few years, at which time the full breadth of molecular single-cell analytics will be brought to bear and will likely change the way in which we monitor cancer.

 

Daniel Haber, MD, PhD, is director of Massachusetts General Hospital Cancer Center, professor of oncology at Harvard Medical School and a researcher at Howard Hughes Medical Institute. He can be reached at dhaber@mgh.harvard.edu. Disclosure: Haber reports research support from the Johnson and Johnson Center for CTC Technologies at Massachusetts General Hospital, as well as scientific advisory board roles with Life Technologies and Cell Signaling Technologies.