Solid tumor biopsies are expensive and invasive, making them less than ideal for patients who are older or very young. Furthermore, with certain forms of cancer such as non-small cell lung cancer, solid tumor biopsies are not obtainable for a significant
number of patients or do not have sufficient DNA to perform NGS profiling.
Cell-free DNA (cfDNA – can also be referred to as circulating tumor DNA and circulating nucleic acid), derived from blood or some other biological liquid, represents a more accessible material from which DNA can be obtained for NGS testing. 10 cc blood
draws are typically standard. cfDNA yields vary widely between patients, cancer type, and cancer stage but are nearly always low, averaging 20-30 ng/10 cc blood draw. The cfDNA is characterized by a uniform fragmentation size of 167 bps indicating
its being derived from nucleosome-protected DNA resulting from in vivo cleavage occurring as cells undergo apoptosis and/or necrosis. Therefore, cfDNA yield is related to the type of cancer. Cancers associated with a high degree of apoptosis
will have the larger yields.
Collection of blood for cfDNA studies requires special procedures. cfDNA is usually purified from the plasma fraction which is devoid of white blood cells. This is done to prevent gDNA contamination resulting from WBC lysis. Genomic DNA contamination
is detected as high molecular weight (HMW) DNA. This sort of contamination will dilute out the tumor cfDNA, preventing detection of rare variants. We have developed a reverse SPRI cleanup assay that preferentially removes HMW DNA from cfDNA contaminated
with gDNA and have used it on both new samples and archival plasma samples. Results have been very good, although there will be residual remnants of WBC gDNA and some degree of tumor cfDNA loss during the purification.
CCGD has made large strides to improving cfDNA library construction. We have found that overnight adapter ligation improves the conversion percentage of sequenceable material. Increased depth of unique coverage has been facilitated by our introduction of IDT
dual-matched adapters that contain a unique molecular identifier. UMI usage increases unique depth of coverage, while the dual-matched adapters allow detection of barcode hopping, significantly increasing confidence in detected rare variants.
Our current cfDNA input recommendations are to use 20 ng or more per library prep, although we have been able to obtain data with less than 2 ng. However, the level of unique depth of sequencing is directly dependent on the amount of cfDNA used. Assuming
one genome is equivalent to 6.6 ng of DNA then 2 ng would permit a theoretical maximum depth of ~300X; 20 ng would grant a tenfold increase to 3,000X. Keep in mind those calculations assume 100 percent conversion of the cfDNA to sequenceable material
and 100 percent hybrid capture efficiency, which is not feasible.
Another consideration is that the low tumor content associated with most cfDNA samples dictates how the NGS assay is performed. To sequence to => 5,000X requires:
- Sufficient material (at least 30 ng)
- Low sample plexing which increases sequencing cost (<=20 samples/HiSeq 2500 run)
- Small hybrid capture bait set – target region is 200 Kb or less; baits with large off-target capture cannot be used
Finally, NGS with hybrid capture is best suited for profiling a patient sample since it only requires knowledge of which gene(s) to target. ddPCR requires a priori knowledge of the precise mutation but will always give better sensitivity than
NGS/hybrid capture at the expense of target multiplexing. Amplicon-Seq straddles the two for sensitivity and a priori knowledge of what is being investigated.