Oncologists must be ‘very cautious’ as liquid biopsy use expands
So-called “liquid biopsies” — noninvasive tests used to evaluate blood for cancer cells or DNA — have been refined extensively over the past decade, leading to a growing number of applications in cancer treatment.
These uses range from monitoring cancer growth to detecting genetic mutations, identifying signs of relapse and predicting sensitivity to immunotherapy.
Although these biopsies are becoming popular because of their noninvasive nature and quick result turnarounds, they are not yet ready for use in early-stage settings. Further, although specificity has improved, the risk for false-positive or -negative results remains, especially as the list of biomarkers expands with better understanding of the underlying biology of various cancers.
Still, technological upgrades and ongoing clinical trials have researchers optimistic that liquid biopsies will move into earlier settings in some diseases, while continuing to be a reliable complement to tissue biopsies in metastatic or recurrent settings.
“Liquid biopsies in solid tumors are where many significant advancements are coming,” Joshua M. Lang, MD, associate professor in the department of medicine at University of Wisconsin and director of UW Circulating Biomarker Core, told HemOnc Today. “As a medical oncologist, there is a certain jealousy of hematologists who can access tumor cells just from a blood draw, enabling closer disease monitoring and intervention for early signs of treatment-resistant disease. We are trying to model what they have successfully done in leukemias and other hematologic malignancies. The challenge is that these circulating events are much rarer in solid cancers.”
HemOnc Today spoke with researchers about the advantages of liquid biopsies, technological improvements that continue to expand their applicability, their utility in monitoring patients for relapse and determining best treatments, and whether the challenges to using liquid biopsies in earlier disease settings can be overcome.
Advantages across malignancies
Advantages of liquid biopsies include their feasibility, low cost, rapidly improving accuracy and real-time monitoring.
Feasibility has been demonstrated in clinical trials across several different malignancies. The liquid biopsy method is well-liked by clinicians and patients because blood draws are routine and much less invasive than traditional tumor tissue biopsies.
Accuracy also has been tested across several clinical trials, with positive results. However, the potential for false-positive results leading to overtreatment remains a concern.
In some instances, depending on the cancer type and the genetics, the tests can help guide clinicians in choosing new therapies for patients who may have developed resistance to their current treatments.
“These allow us to test patients with cancer faster, so there is no delay in setting up and conducting the biopsy. They are also less risky than a tissue biopsy because they are noninvasive,” Sameek Roychowdhury, MD, PhD, associate professor and member of the translational therapeutics program at The Ohio State University Comprehensive Cancer Center, told HemOnc Today. “Because blood draws are easy, you can do them repeatedly every week or every 2 weeks if necessary. This lets us monitor the patient’s cancer to see if it is changing or becoming resistant to treatment.”
Liquid biopsies also provide an opportunity for analysis of heterogeneity across different metastatic sites that can’t be performed with tissue biopsies, Lang said.
“The ability to sample tumor-derived content — such as cells, DNA or exosomes — that have been shed into circulation from different metastatic or primary sites of cancer is potentially very important,” Lang said. “When cancers become resistant to an initially successful therapy, tumor heterogeneity across these different metastatic sites may hold the key to understanding resistance.”
However, different malignancies are more sensitive to liquid biopsy than others.
“There are reports of clinical efficacy and integration in lung cancer, colorectal cancer and breast cancer,” Charu Aggarwal, MD, MPH, Leslye M. Heisler assistant professor for lung cancer excellence at Perelman School of Medicine of the University of Pennsylvania and a HemOnc Today Editorial Board Member, told HemOnc Today. “All three of these malignancies have large clinical trials that have demonstrated that liquid biopsies can help improve outcomes by delivery of targeted therapy or, in the case of colorectal cancer, a clinical trial examining who should get certain adjuvant therapy.”
For example, results of a phase 3 trial presented at last year’s European Society for Medical Oncology Congress showed circulating tumor DNA (ctDNA) detected through liquid biopsy appeared to be an independent prognostic marker for identifying patients with colorectal cancer who are likely to relapse after surgery.
Taieb and colleagues analyzed plasma samples of 805 patients from the IDEA-FRANCE trial who underwent liquid biopsy before adjuvant chemotherapy for stage III colorectal cancer. Of the patients sampled, 109 (13.5%) had ctDNA in their blood.
Results showed 2-year DFS rates of 64% for ctDNA-positive patients compared with 82% for ctDNA-negative patients (HR = 1.75; 95% CI, 1.25-2.45).
A multivariate analysis that accounted for age, sex, microsatellite instability, perforation, T stage, N stage and treatment group confirmed ctDNA as an independent prognostic marker (HR = 1.85; 95% CI, 1.31-2.61).
Lung cancer also has been the focus of liquid biopsy research due to the availability of targeted treatments for the disease.
Results of the NILE study presented at last year’s American Association for Cancer Research Annual Meeting showed liquid biopsies accurately identified guideline-recommended biomarkers in metastatic non-small cell lung cancer at a rate similar to that of standard tissue tests, but with a shorter turnaround time.
Leighl and colleagues sought to show the noninferiority of comprehensive cell-free DNA (cfDNA) analysis vs. standard tissue genotyping in identifying biomarkers among 282 patients (81.9% white; 54.3% women) with previously untreated, nonsquamous cell metastatic NSCLC.
Standard tissue biopsy identified a guideline-recommended biomarker in 60 patients, whereas cfDNA detected a biomarker in 77 patients (21.3% vs. 27.3%; P < .0001). Moreover, cfDNA increased the total number of patients with a recognized guideline-recommended biomarker by 48%, from 60 patients to 89; this included patients whose tissue biopsies produced results of negative (n = 7), not assessed (n = 16) or insufficient quantity (n = 6).
The cfDNA test demonstrated a positive predictive value of 100% for four biomarkers — EGFR, ALK, ROS1 and BRAF — that are targets of FDA-approved therapies.
Test turnaround time — defined as time from the test order to test results — was 9 days with cfDNA vs. 15 days with tissue genotyping (P < .0001).
“Other cancers will follow appropriately as more drug development and marker identification occur,” Edward S. Kim, MD, chair of the department of solid tumor oncology at Levine Cancer Center at Atrium Health and a HemOnc Today Editorial Board Member, said in an interview. “We could see it soon in breast cancer and thyroid cancer, but other tumor types do not have as many of the targets or corresponding therapies that make this feasible.”
Liquid biopsies also have demonstrated utility in prostate cancer because of the disease’s very high burden of circulating tumor cells.
A study presented in 2017 at Genitourinary Cancers Symposium showed liquid biopsy identified genetic changes in ctDNA in 94% of 514 patients with advanced prostate cancer, including TP53 (36%); the androgen receptor gene (22%); APC (10%); NF1 (9%); EGFR, CTNNB1 and ARID1A (6% each); and BRCA1, BRCA2 and PIK3CA (5% each).
Further, a higher number of genetic changes, including changes in the androgen receptor gene, were associated with shorter time to failure (HR = 1.42) and poorer survival (HR = 2.51).
“The reason for the high burden of circulating tumor cells is not known at this point,” Lang said. “Maybe because it’s the presence of these cells in bone metastases, but we do find these rare cells in prostate cancer.”
These studies highlight the advances and accompanying advantages that liquid biopsies have brought to both the understanding of cancer biology and treatment, Kim said.
“They also show the need for further research into the utility of these tests with increasing knowledge of different disease sites,” he said.
Liquid biopsies, once a far-from-perfect technology, have benefited from focused research to improve their accuracy and applicability across cancer types.
“The technology has changed significantly over the past decade,” Aggarwal said. “Liquid biopsies have gone from single-gene assays or detection of circulating tumor cells to being able to detect a broad range of genes on the ctDNA. Newer approaches are being developed to look at other circulating material, such as exosomes or other molecules. This includes other approaches to detect circulating tumor cells.”
Technology also has made it convenient for clinicians to process the blood on site or ship it to a central laboratory where the results are gathered. The simplicity of the process has allowed physicians to focus on the results and how best to interpret them.
“We didn’t realize the magnitude of the specificity issues at first, but now that we are aware of them, the technology is rapidly improving to address them,” Lang said. “We are the furthest along in cell-free DNA, which involves sending the collected sample to a central lab for analysis. As we collect samples more broadly for analysis and send them out to be analyzed, we have to figure out the specific clinical contexts where liquid biopsies are most useful.”
Some of the most notable advances over the past 5 years have come through next-generation sequencing.
A study published in 2017 in Cancer Research demonstrated that next-generation sequencing of ctDNA revealed unique somatic profiles with targetable alterations among patients with carcinoma of unknown primary.
Kato and colleagues used targeted clinical-grade, next-generation sequencing of circulating tumor DNA to evaluate 54 to 70 genes in 442 patients.
The most common variants included TP53-associated genes (37.8%) and those involved in the MAPK pathway (31.2%), PI3K signaling (18.1%) and cell-cycle machinery (10.4%).
Of the 290 patients with characterized alterations, 255 (87.9%) showed distinct genomic profiles, and 289 (99.7%) had potentially targetable alterations.
In a study by Aggarwal and colleagues published in 2018 in JAMA Oncology, the addition of plasma next-generation sequencing to solid tissue biopsy improved detection of therapeutically targetable mutations in patients with stage IV NSCLC.
The study included 323 patients (60.1% women; median age, 65 years) who underwent plasma next-generation sequencing using a 73-gene commercial platform (Guardant360, Guardant Health). The researchers assessed 166 patients at initial diagnosis and 157 upon disease progression.
Thirty-five percent of the patients had targetable mutations in EGFR, ALK, MET, BRCA1, ROS1, RET, ERBB2 or BRAF.
Among those who underwent both plasma and tissue next-generation sequencing or were unable to have tissue sequencing (n = 229), therapeutically targetable mutations were found in 47 patients (20.5%) with tissue biopsy alone and 82 patients (35.8%) with the addition of liquid biopsy. Liquid biopsy detected targetable mutations in one-third of patients (n = 33) who opted for that test alone.
Further, 36 of 42 patients (85.7%) who received therapy as indicated by their plasma next-generation sequencing results achieved complete or partial response or stable disease.
Researchers caution, however, that the technology has a long way to go and still carries the risk of providing false results or no results compared with traditional tissue biopsy.
For instance, research by Torga and colleagues published in JAMA Oncology showed two commercial liquid biopsy platforms that evaluated tumor samples from the same patient population yielded significantly different results.
Researchers sent identical samples from 40 men with prostate cancer for analysis by the Guardant360 platform — which assessed at least part of the coding sequences of 73 genes — and the PlasmaSELECT (Personal Genome Diagnostics) platform, which sequenced segments of 64 genes.
Results showed 25 patients (62.5%) had at least one genetic mutation reported within the overlapping sequences covered by both platforms.
Only three patients (7.5%) had identical results for all mutations from both platforms. Both platforms’ results matched for at least one reported mutation for six patients (15%). In 40% of cases, no mutations that could have been covered by both panels were detected by both platforms.
“We don’t fully know when liquid biopsies work well and how to measure their quality,” Roychowdhury said. “If you perform liquid biopsy and you don’t find something, we are left wondering whether it worked or not. One reason nothing was detected could be because the tumor is not shedding DNA into the blood, because the amount of tumor present is not enough to be detected. It could also be that the type of mutation, such as fusion, is not easy to detect in the small DNA fragments.
“If you consider the automobile, the first ones were very different than the ones we have today,” he added. “There is still active refinement in the way we perform liquid biopsy testing and the technology and computer science behind it. There is great promise that will require well-controlled retrospective and prospective studies to understand the limits and applications.”
Monitoring tumor evolution
A major use of liquid biopsies has been to track patients who might relapse or become refractory to treatment. Assessing ctDNA that is shed into circulation for biomarkers and independent prognostic markers is a much more practical means of serial monitoring of patients’ cancer treatment and progression than use of standard tissue biopsy.
However, cancers that don’t have a high burden of disease or that don’t have many sites of metastatic disease are less likely to shed — a major limitation of liquid biopsies cited by experts who warn against completely replacing tissue biopsies.
“There is evidence that patients with EGFR-mutant NSCLC and residual ctDNA in the bloodstream after starting targeted therapy are more likely to experience progression of disease earlier,” Aggarwal said. “Those patients are enrolled onto studies with combination therapy and are possibly scanned earlier. It’s now about getting them onto future randomized clinical trials.”
For example, the genetic makeup of a glioblastoma tumor can change over time. Currently, the only ways for clinicians to learn about these changes are to conduct repeat brain tumor sampling through a craniotomy with tumor resection or through stereotactic brain biopsy. Both methods are highly invasive.
This challenge is what prompted Stephen Bagley, MD, MSCE, assistant professor of medicine at Hospital of University of Pennsylvania, and colleagues to conduct a prospective pilot study to assess the utility of plasma cfDNA among 42 adults (median age, 65 years; range, 20-81) with newly diagnosed glioblastoma.
The researchers calculated plasma cfDNA before initial tumor resection and regularly throughout chemoradiotherapy. They also drew blood from 42 age-matched healthy adults (median age, 62 years; range, 22-82) who served as controls.
Results, published last year in Clinical Cancer Research, showed patients with glioblastoma had higher plasma cfDNA concentration than healthy adults (mean, 13.4 ng/mL vs. 6.7 ng/mL; P < .001).
The concentration of plasma cfDNA appeared significantly associated with total radiographic tumor burden on first post-radiation MRI scan (P = .003). Concentration tended to increase before or during radiographic tumor progression.
The 14 patients with preoperative plasma cfDNA concentration above 13.4 ng/mL demonstrated shorter median PFS than the 28 patients with lower concentration (4.9 months vs. 9.5 months; P = .038).
Researchers detected one or more somatic mutations in plasma cfDNA among 55% of patients, which appeared associated with shorter median PFS (6 months vs. 8.7 months) and OS (5.5 months vs. 9.2 months), although the differences did not reach statistical significance. These mutations were different than those detected by tissue biopsy.
“We need larger studies to confirm our results, but our initial results raise the question of whether liquid biopsy is, indeed, complementary to tissue biopsy in glioblastoma,” Bagley told HemOnc Today when the study was published. “If we can confirm in larger studies that liquid biopsy provides meaningful genetic information about the tumor, this may lead to us being able to track the glioblastoma tumor genetics over time with fewer invasive brain procedures.”
Despite the promise of using liquid biopsies to monitor for disease evolution, this use is challenged by the constant discovery of new prognostic markers. Thus, the scope of liquid biopsies will have to adapt to these new biomarkers in rapid fashion.
“Liquid biopsy platforms are not as broad as tissue platforms, and that creates a problem as the number of markers continues to expand,” Kim said. “There will come a point where the technology must expand so that it can measure more challenging markers.”
Experts with whom HemOnc Today spoke agreed the ultimate use of liquid biopsy would be to screen patients for cancer prior to diagnosis. Early cancer detection in the general population could substantially improve outcomes in the long run.
However, experts acknowledge such an approach is years away from being implemented.
“Screening liquid biopsies are in no way ready for prime time universally, but such an approach is the holy grail of liquid biopsies,” Roychowdhury said. “It’s a very exciting concept, but we are still far from that reality.”
Rather, the first focus of such an approach should be to more effectively screen individuals at high risk for disease, Kim said.
“It’s absolutely a clinical need because one of the first things my patients ask me when I see them is if their son or daughter is at risk for cancer,” he said.
“Specifically, researchers are looking at high-risk populations — such as those with a BRCA gene mutation — to monitor whether they have developed cancer,” he said. “This raises the probability that we could find something at an early stage and start treatment right away.”
For now, liquid biopsies are common in community practices and academic centers as a complement to tissue biopsies.
“This is all about using them in the right setting and at the right time,” Kim said. “If doctors are using these as the only biopsy, that’s not good practice. Tissue biopsies are still needed upfront at diagnosis. But we know that some patients have diseases that are challenging for biopsies because it is tough to get enough tissue samples. This is where liquid biopsies can help.”
A good comparator is PSA screening for prostate cancer, Kim added. Although it is a very sensitive test, it is not specific for cancer or for distinguishing between more aggressive and indolent cancers.
“Some of the new liquid biopsy platforms show greater molecular specificity, but we have to be very cautious because the risk is much higher in terms of the harm that can be done to patients,” Lang said. “Our understanding of the biology of solid tumors and what we find in blood continues to evolve. We are just at the tip of the iceberg of what can be found with liquid biopsies, and I think the next 10 years are going to be very exciting in this field.” – by John DeRosier
Click here to read the , “Will liquid biopsies be used routinely to screen for cancer?”
Aggarwal C, et al. JAMA Oncol. 2018;doi:10.1001/jamaoncol.2018.4305.
Bagley SJ, et al. Clin Cancer Res. 2019;doi:10.1158/1078-0432.CCR-19-2533.
Kato S, et al. Cancer Res. 2017;doi:10.1158/0008-5472.CAN-17-0628.
Leighl NB, et al. Clin Cancer Res. 2019;doi:10.1158/1078-0432.CCR-19-0624.
Sonpavde G, et al. Abstract 149. Presented at: Genitourinary Cancers Symposium; Feb. 16-18, 2017; Orlando.
Taieb J, et al. Abstract LBA30_PR. Presented at: European Society for Medical Oncology Congress; Sept. 27-Oct. 1, 2019; Barcelona, Spain.
Torga G, et al. JAMA Oncol. 2018;doi:10.1001/jamaoncol.2017.4027.
For more information:
Charu Aggarwal, MD, MPH, can be reached at email@example.com.
Edward S. Kim, MD, can be reached at firstname.lastname@example.org.
Joshua M. Lang, MD, can be reached at email@example.com.
Sameek Roychowdhury, MD, PhD, can be reached at firstname.lastname@example.org.
Disclosures: Aggarwal reports advisory board roles with AstraZeneca, Bristol-Myers Squibb, Celgene, Merck and Roche, and research funding to her institution from AstraZeneca, Incyte, MacroGenics and Merck. Kim reports consultant roles with AstraZeneca, Boehringer Ingelheim, Merck and Takeda; research funding from AstraZeneca, Boehringer Ingelheim, Ignyta and Merck; and a data and safety monitoring board role with Celltrion. Lang and Roychowdhury report no relevant financial disclosures.