Numerous prospective clinical studies use droplet digital PCR (ddPCR)-based liquid biopsy to quantify driver and resistance mutations associated with tumors such as lung cancer, melanoma, and breast cancer. These studies have reported positive predictive values of up to 100 percent, turnaround times of two to three days, and a significant increase in sensitivity compared to other methods.
David Polsky, MD, PhD, and George Karlin-Neumann, PhD, joined MLO to discuss how ddPCR technology is shaping the cutting edge of melanoma care—and the issues that must be addressed before it can be widely used for clinical decision-making in oncology.
MLO: What is the appeal of liquid biopsy to clinical labs?
David Polsky (DP): The term “liquid biopsy” became popular when we began looking at cell-free DNA (cfDNA) in the plasma of patients with solid tumors. The concept of a biopsy, of course, involves getting a sample that tells you what kind of tumor the patient has. Ordinarily, we need to get a piece of the solid tumor to do this; with a liquid biopsy, the molecular diagnosis is derived from bodily fluids. That information can give us insight into which treatments a patient’s tumor may be susceptible to.
The other aspect of liquid biopsy is monitoring disease. For example, after surgery, liquid biopsy testing could identify patients who have the highest risk of relapse. This could alert a physician to consider adjuvant therapy.
George Karlin-Neumann (GKN): In clinical use, liquid biopsy allows more rapid determination of the genetic alterations driving a patient’s cancer than tissue biopsy does. Thus, it may allow for more rapid decision-making on a course of therapy. This can help to reduce a patient’s anxiety about how the disease will be treated while waiting for test results.
How does ddPCR detect blood-based biomarkers in melanoma patients?
GKN: ddPCR works by finely dividing plasma cfDNA samples (or other sources of DNA) into tiny droplets, using end‐point PCR to amplify targets of interest in each droplet. The number of positively and negatively fluorescent droplets produced in a reaction is used to calculate the concentration of the target or targets in that sample. Using primers and Taqman probes designed against DNA sequences of interest, ddPCR enables absolute and highly sensitive nucleic acid quantification that is less susceptible to PCR inhibition than qPCR and less costly and complicated than NGS workflows.
DP: We helped develop new digital PCR tests for two mutations in a promoter region of the TERT gene. These complement a handful of commonly-known hotspot mutations present in one-half to two-thirds of melanoma patients. Together with BRAF and NRAS mutations, the digital PCR assays could be useful for at least 70 percent of metastatic melanoma patients.
What are the limitations of ddPCR technology?
DP: We only test for one gene at a time, so we need to know what mutation the patient has in his or her tumor to choose the appropriate test to run. In terms of disease monitoring, it’s not an ideal method for taking a broader look at the patient’s tumor to identify potential resistance mutations early in treatment.
For example, it’s been described that some patients with a BRAF-mutant tumor, when they fail their treatment, can develop NRAS-mutant tumors, which you don’t want to treat with a BRAF inhibitor. If you know that early, you might be able to switch treatment faster.
GKN: Its major limitations are the need for defined target sequences or regions against which the assays are designed and the smaller number of markers that can be interrogated in a single reaction. On the other hand, multiplexing does currently allow for half a dozen or so markers to be assayed per reaction well, making ddPCR technology well suited to arriving at specific, actionable clinical decisions in a variety of cancers. Where a larger number of candidate mutations need to be considered to make a treatment decision, a broader profiling technology such as NGS is better suited.
How might ddPCR-based molecular diagnostics improve melanoma treatment?
DP: We published a paper in 2016 looking at disease monitoring in metastatic melanoma patients. Clinicians use lactate dehydrogenase (LDH), which is a serum-based biomarker, for detecting disease progression or response to treatment. But it’s known to have a low sensitivity and specificity for monitoring disease. So it’s not thought to be very useful, although it is monitored.
We found that ddPCR assays for either BRAF or NRAS mutations were more sensitive than LDH at detecting the presence of metastatic disease as well as detecting disease progression. We think that with further validation studies we may be able to use these for disease monitoring, and eventually for diagnosis as well.
GKN: Tissue biopsy is the standard for diagnosis, but it is not amenable to repeated sampling to evaluate a patient’s response to surgery or systemic therapy. ddPCR testing of cell-free DNA in blood or other fluids is especially well suited for diagnosing and monitoring treatment response in melanoma because only a small number of recurrent mutations are responsible for driving disease in the majority of these patients’ tumors.
What are the next steps for validating molecular diagnostic tests for melanoma?
DP: You not only need to analytically validate liquid biopsy tests for laboratory use, including for sensitivity, specificity, and reproducibility; you also need clinical validation that asks, “Okay, I can give you a number, but what does that mean?”
That work has to be done in conjunction with clinical trials and show that the information you get from the tests has value in helping oncologists manage their patients. We showed in our paper that it was more sensitive than LDH, but you have to show that again, and show it with another set of patients, and other types of uses, all in the clinical setting.
GKN: Also, the significance of that number might change depending on how you’re using the test. As a diagnostic, your primary question is probably whether a mutation is present, or potentially present above or below a certain threshold, to decide on a treatment. For monitoring, the question becomes more quantitative: how much is the level of cell-free tumor DNA changing, and how closely does that concord with how the patient is responding?
In addition to analytical and clinical validation, there’s an overall, pre-analytical workflow to consider: How do I draw the blood sample? What kind of tube do I use? How soon do I need to isolate plasma from the sample? How best should the cell-free DNA be purified? These are things that we need to test for the sake of reliable results because, ultimately, the test will be used with patients.
What new tools will be available to clinical labs in the near future?
DP: In the next year or so, I think probably we’re going to see clinical labs start to adopt ddPCR. The tests are commercially available, so it’s just a matter of laboratories doing all the analytical validation work and getting permission from their state agencies to run the tests. But showing the value of the test in managing patients is more in the five-year timespan. I’m hoping we can get to where managing the patient based on these blood tests will be able to extend survival.
GKN: I agree. There is an understandable prejudice—it makes intuitive sense—that if we can get these results and act earlier, when the tumor is smaller or the patient is perhaps in better health, the patient is going to do better. But showing real clinical utility means testing these assumptions and measuring tumor progression, patient survival, and quality of life based on using these tests for measuring plasma tumor DNA levels versus acting without this information.
On the technological side, we also expect more evolved instrumentation that enables larger numbers of targets to be assayed per ddPCR reaction (perhaps half a dozen to a dozen, and that integrates steps from sample processing through detection and analysis to report generation. This will be built around a better-informed view of sample types and preparations that offer the highest biological signal for the most informative biomarkers.
David Polsky, MD, PhD, is a dermatologist at NYU Langone Medical Center and a professor of dermatology and pathology at the New York University School of Medicine.
George Karlin-Neumann, PhD, serves as Director for Scientific Affairs for Bio-Rad’s Digital Biology Group.
