Cell-Free DNA: A Diagnostic Revolution in the Works?

When cells die or undergo stress, nucleases digest cellular DNA and release fragments of approximately 170 base pairs in length known as cell-free DNA (cfDNA). cfDNA is found in many tissues, but in blood (or plasma) it is known as circulating cfDNA. Although cfDNA is found in healthy individuals its levels increase in certain diseases, including cancers and many inflammatory conditions. Hence the idea that cfDNA signatures might inform on a patient’s health or disease status as a type of liquid biopsy. Much work remains before liquid biopsies enter the mainstream of medical diagnostics and prognostics, but the research field is booming and practical applications are looming, as liquid biopsies have the potential to replace many invasive surgical biopsies with a simple blood draw.

Actionable genes

Success will depend on how well genomic information from cfDNA samples correlates to disease phenotypes. In June, scientists from Dana-Farber Cancer Institute published results of a study using the Resolution ctDx-Lung™ assay, from Resolution Biosciences, for detecting actionable gene fusion mutations in non-small cell lung cancer.

The Resolution assay detected 13 out of 16 (81.3%) fusions (allele frequency range 0.17–62.8%), compared with just seven mutations (43.8%) for a competing assay kit. According to Resolution CEO Mark Li, the shorter probes used by the Resolution assay “maintain high specificity while allowing better detection of DNA fragments containing fusions.”

Separately and more recently, the company’s Resolution HRD™ assay was granted Breakthrough Device Designation by the U.S. Food and Drug Administration (FDA), for use in the diagnosis of advanced prostate cancer. HRD (homologous recombination deficiency) mutations, which are suggestive of poor prognosis, exist in up to 10% of men with localized prostate cancer.

Li explains the difficulties in detecting oncogenic fusions generally, and specifically from plasma samples. “The two main challenges are the short fragment length of cfDNA, and the need to detect a wide range of breakpoints and partner genes.”

The task, therefore, becomes detecting gene fusions without a priori knowledge of the gene partner and breakpoint. “With a fusion, we often know the target oncogene (ALK, RET, ROS1, FGFR3, etc.), but not always the fusion partner,” Li adds. “While many canonical partners exist, such as EML4 with ALK, each patient may have a unique breakpoint in the introns. Because of this, most other methods, including PCR-based detection, require pre-programmming for specific partner and breakpoint pairs prior to detection.”

Resolution’s liquid biopsy assay detects fusions without prior knowledge of the fusion partner or specific breakpoint, thus enabling the detection of more clinically actionable driver mutations, according to the company.

Sequence length for extracted cfDNA is the other issue. cfDNA molecules are typically about 160–170 base pairs in length, whereas the industry standard capture probe size is 120 base pairs. “The long capture baits almost guarantee a lower fusion capture rate, which lowers sensitivity and the overall detection rate,” Li says. Resolution’s platform uses probes of just 40 base pairs to provide a higher capture rate while maintaining specificity and reducing bias. “This capability has two benefits: We can easily capture any area of the genome equally, and we do not waste resources on sequencing reads that are irrelevant. Our on-target rate is incredibly high at 99%.”

Resolution liquid biopsy assays are based on the company’s patented cfDNA NGS analysis platform, which includes proprietary targeted capture NGS chemistry tightly coupled with cloud-based bioinformatics.

Pre-analysis considerations

cfDNA studies involve a complex series of operations. Given the low frequency of many species of interest, their quite large concentration dynamic ranges, and the presence of interfering species, sample preparation becomes the dominant, if not the rate-limiting step of many workflows.

“Even more challenging than actual sample preparation methods is what happens during the pre-analytical handling, between blood collection and the actual assay,” says Dr. Georg Wieczorek, associate principal scientist for MDx sample technologies at QIAGEN. The object is to collect plasma that reflects the patient’s actual status and stabilize it so it continues to do so to the point of analysis. QIAGEN has selected the PAXgene Blood ccfDNA tube from PreAnalytiX’s, a joint subsidiary of QIAGEN and BD, for use with several QIAGEN cfDNA analytical kits (e.g. QIAsymphony PAXgene Blood ccfDNA Kit). The PAXgene tube uses a non-crosslinking-based anticoagulant technology to preserve blood samples.

The next consideration involves sample volumes required to complete a study. Blood is 40% to 50% plasma, so a 10 mL collection tube will contain between 4–5 mL of plasma. That may not seem like much blood to healthy blood donors, but it can be for very sick patients subjected to multiple blood draws. “In our experience a 10 mL draw is fine for most patients but 20 mL is pushing it, and volumes higher than that might be a challenge,” says Dr. Thorsten Voss, associate director, PreAnalytiX at QIAGEN. Voss suggests working backwards from the very end of the process, from readout to assay to patient, to estimate required blood collection volumes.

With molecular techniques now everywhere in biological research and medical diagnostics, the International Standards Organization (ISO) published a guidance in September 2019, ISO 20186-3:2019, Molecular in vitro diagnostic examinations—Specifications for pre-examination processes for venous whole blood—Part 3: Isolated circulating cell free DNA from plasma. The document specifically addresses how investigators should handle samples for ccfDNA analysis during the critical time between blood draw and assay.

As noted in the guidance regarding modern methods for analyzing nucleic acids, proteins, and metabolites in human tissues and body fluids, “...the profiles of these molecules can change drastically during the pre-examination process, including the specimen collection, transport, storage, and processing.” Specifically, ccfDNA profiles may change through the release of genomic DNA, ccfDNA degradation, or the appearance or disappearance of certain species. “Therefore,” the guidance continues, “special measures need to be taken to secure good quality specimens for ccfDNA examination.”

Assuming samples are collected by the book, the next challenge is isolating as much ccfDNA as possible from the sample. “Although ccfDNA concentrations are elevated in cancer patients, the absolute concentrations are quite low, and even when not the target regions may themselves be rare,” says Voss. “But really it’s the final readout, be it digital PCR, qPCR, or NGS, that determines how much plasma you need in your prep.”

Already the subject of intense investigation, cfDNA has the potential to be a game-changer in diagnostic medicine. Preliminary studies show sensitivity and specificity numbers typical for in vitro cancer tests. The trove of data available through a single cfDNA test holds independent scientific value as well. “I believe we are only scratching the surface of uses for cfDNA genotyping,” says Resolution’s Mark Li. He cites a recent study from Memorial Sloan Kettering Cancer Center demonstrating the use of cfDNA to guide cancer therapy in non-small-cell lung cancer, a paper from Vanderbilt University monitoring resistance to the cancer drug ensartinib, and an article from AstraZeneca showing that cfDNA can be used to confirm dosing schedules.