https://orcid.org/0000-0002-5169-7085
It is intriguing to highlight the work by Gouda et al on developing mutation-agnostic liquid biopsy.1 It emphasizes the need to realize that there are different kinds of liquid biopsies (Fig. 1). The landscape of available liquid biopsies is changing. While looking at data on any type of liquid biopsy, it is important to consider not only the specific tumor’s biology and shed, but also what the assay is trying to measure. In an article published in this issue of The Oncologist, the authors report that for colorectal cancers, liquid biopsies measuring mutations only from circulating tumor DNA (ctDNA) can potentially miss a subset of these cancers; thus, the authors considered a liquid biopsy based on methylation or epigenomic signatures.1 This is what the team has termed “mutation-agnostic” detection of colorectal cancer.
Types of liquid biopsies detecting cancers through different lenses. In some patients, cancer is detected on circulating tumor DNA (ctDNA) only, in some on methylation or epigenomics signatures only, and in some on both platforms.
This is not the first liquid biopsy using this approach of using liquid biopsy-based methylation-specific signatures. Numerable assays that are in the early detection or minimal residual disease (MRD) space already utilize a similar approach. Some are methylation-based only, eg, the multi-cancer early detection (MCED) test in the early detection space for a multitude of cancers, including colorectal cancer.2,3 Others are a composite of ctDNA and methylation-based assays in the MRD space specifically for patients with colorectal cancer (Fig. 1).4 In the article looking at plasma-only MRD assay for patients with colorectal cancer, the sensitivity of the assay was noted to go up by 25%-36% when the epigenomic signature lens was added to a panel-based genomic detection of colorectal cancer.4 Specifically, in their cohort of patients with colorectal cancer, “…47% of samples were positive by both epigenomic and genomic calls, while 28% and 25%, respectively, were positive by either genomic or epigenomic calls alone”.4
This is a good problem that we can now look at the same cancer, in this case colorectal cancer, through different lenses (Fig. 1). This can increase the proportion of patients who can benefit from such an approach as the assays continue to be developed incorporating genomic, epigenomic, and other lenses (eg, proteomics and exosomes).
Like any other assay or approach, there are always limitations to consider when interpreting the results, in this case methylation-specific signatures. The specificity of this assay as reported by the authors was 92%, as a patient with pancreatic cancer was also classified as colorectal cancer.1 This was thought to be due to potentially overlapping methylation patterns between gastrointestinal cancers. We had reported on a similar story of an MRD being false positive in a patient with colorectal cancer who was subsequently noted to have more than 1 primary cancer (hepatocellular cancer and a head and neck cancer).5 This is not an uncommon scenario. With the aging patient population, it is not uncommon to have more than 1 primary cancer. This is also not surprising. We have long known, from mutational testing in practice for solid and liquid tumors, that mutations (eg, KRAS, BRAF, IDH, and HER2) can be seen across tumor types. It should not be surprising that we will find instances of possible overlap of methylation patterns as well.5-7 In the early detection work by Klein et al, cancers that shared oncogenic pathways (eg, human papilloma virus (HPV)-driven cancers, such as anal or cervical cancer), the site of origin was hard to predict.3
Similarly, we learned that for tumor-uninformed mutation- or genomic-based liquid biopsies, we must remember and filter for possible clonal hematopoiesis of indeterminant potential.8 For methylation and epigenomic-based liquid biopsies, it would be of value to have positive cancer controls (individuals with other cancers), and perhaps even non-cancer positive controls (eg, individuals with other chronic or active diseases) that potentially may lead to a methylation signal being detected. This would be important as our understanding of the methylation/epigenomic field continues to evolve.
The future might very well be in assays that focus on multiple approaches and targets. A recent liquid biopsy for hepatocellular cancer, for example, incorporated not only methylation markers but also the glycoprotein alpha fetoprotein (AFP) and even sex.9 In summary, it is a good problem that we have a multitude of assays looking at the cancer through different lenses.10 As liquid biopsies continue to be incorporated in clinical trials and practice, it is important to understand what the liquid biopsy platform is measuring, the data on both positive and negative controls, the shedding, the biology, the unique patient situation, and the timing of testing for genomic and epigenomic signatures that can affect the sensitivity and specificity of these assays.
Acknowledgment
The author thanks DrawImpacts for their work on the figure that accompanies this commentary.
Conflict of Interest
Pashtoon Murtaza Kasi reported consultancy/advisory board relationship with Elicio, Illumina, Natera, Foundation Medicine, Daichi Sankyo, Tempus, Bayer, MSD Oncology/Merck, Delcath Systems, QED, Taiho Oncology (self/institution), Saga Diagnostics, Servier, Seagen, Lilly, and Ipsen (to institution), and research funding/trial support from Boston Scientific (to institution), Tersera (to institution), and Advanced Accelerator Applications, as well as travel grant for IIT from AstraZeneca.
