Liquid Biopsy in Oncology: Have We Just Scratched the Surface?

In addition to circulating tumor DNA, circulating tumor cells and circulating extracellular vesicles (EVs) should be integrated into liquid biopsy assays to improve their utility in oncology, according to Sanja Dacic, MD, PhD, who added that validation with large cohorts and implementation of machine learning will be critical in this effort.

“As you can imagine, most of these things are in the research stage although some of these things are making their way into clinical practice,” Dacic, a professor of pathology, vice chair and director of Anatomic Pathology, and medical director of the Tumor Profiling Laboratory at Yale School of Medicine in New Haven, Connecticut, said in a presentation delivered during the 25th Annual International Lung Cancer Congress.¹

Understanding the Significance of Early Research With EVs

EVs have gained substantial attention over the past few years, according to Dacic, who explained that the marker, which is released by all living cells including cancer cells, has an important role in intercellular communication. She added that a large number of vesicles are released into the blood’s circulation, reaching over 20,000 vesicles in a 48-hour span. Although EVs are a stable structure, housing DNA, RNA, proteins, and lipids, investigators have found the entity difficult to isolate.

“Most of the studies that are published are looking into subsets, but there are hopes that we are going to analyze individual [components] of EVs, particularly exosomes because they carry lots of information. Because of their important role in intercellular communication, there are certain areas where these will be analyzed, particularly in the area of resistance to certain types of therapies,” Dacic said.

For example, a study published in the Journal of Cellular and Molecular Medicine illustrated the potential involvement of EVs in mediating resistance to EGFR TKIs. In the study, investigators evaluated whether exosomes shed by EGFR T790M–mutant resistant non–small cell lung cancer (NSCLC) cells could transfer drug resistance to cells sensitive to gefitinib (Iressa). To do this, exosomes were isolated from the T790M–mutant NSCLC cell line H1975 and sensitive cell line PC9. Exosome-derived miRNA expression profiles from both cell lines were subject to small RNA sequencing and confirmed by qRT-PCR. Investigators discovered that exosomes that were shed by H1975 could transfer gefitinib resistance to PC9 both in vitro and in vivo through activation of PI3K/AKT pathway signaling. Further analysis confirmed that miR-3648 and miR-522-3p were the 2 most differentially expressed miRNAs.²

Similarly, it has been shown that intercellular transfer of exosomal wild-type EGFR triggers osimertinib (Tagrisso) resistance, Dacic said. This study, published in Molecular Cancer,demonstrated EGFR-mutated sensitive cells and EGFR wild-type resistant cells promoted an osimertinib-resistant phenotype in EGFR-mutated cancer cells, whereas the removal of exosomes from conditioned medium or blockade of exosomal EGFR by neutralizing antibody alleviated this phenotype. Additionally, osimertinib promoted the release of exosomes through upregulation in RAB17, an effect that was negated by RAB17 knockdown. Moreover, exosomes were shown to be internalized by EGFR-mutated cancer cells via Clathrin-dependent endocytosis, after which the internalized exosomal wild-type EGFR protein activated downstream PI3K/AKT and MAPK signaling pathways, leading to osimertinib resistance.³

“What’s more interesting is when these assays were compared with the existing FDA-approved assays for T790M, they showed a higher sensitivity than existing assays. If they’re coupled with existing assays, their sensitivity and specificity goes even higher, and they can provide information even earlier than what we use as traditional assays in practice,” Dacic said.

Going Beyond Surface-Level PD-L1: The Biomarker of Most Resistance

“I know that not everyone is happy with the PD-L1 biomarker, including pathologists, but it’s here. We currently use the PD-L1 membranous expression, but there are many other forms of PD-L1 that can influence the development of resistance or give us some answers about resistance including soluble PD-L1, exosomal PD-L1, and nuclear cytoplasmic PD-L1; but none of these are currently used in clinical practice,” Dacic said.

To better understand the predictive capacity of EV PD-L1 dynamics, investigators conducted a study that compared dynamic changes in EV PD-L1 in plasma samples before and during treatment with checkpoint inhibitors in patients with NSCLC. Data were tested in independent training (n = 33) and validation cohorts (n = 39). The median follow-up in the training and validation cohort was 12.4 months (range, 2.5-33.1) and 13.1 months (range, 3.5-56.5), respectively.

Results showed that an increase in EV PD-L1 occurred only in non-responders (training, P = .017; validation, P = .050) and was an independent biomarker for shorter progression-free survival and overall survival. Moreover, PD-L1 expression performed via tissue biopsy was not predictive of durable response or survival.⁴

“This study very nicely shows that we are living in this era wherein we’re moving beyond looking into a single marker. We have to look into multiple markers and try to build these machine learning algorithms that are going to give us an answer about who is going to benefit from certain therapies,” Dacic said.

Plasma DNA Methylation: Probing the Possibilities

“When you think about DNA methylation, we are basically looking into nearly 30 million CpG sites across the human genome. And luckily, nowadays, we have established cancer-specific methylation patterns that could be obtained from the plasma and [help us] investigate how we’re going to use this information. One area where [investigators] tried to show how we can use this information or this type of assay is [in the form of] a targeted plasma DNA methylation assay,” Dacic said.

As part of the larger Circulating Cell-free Genome Atlas study (CCGA; NCT02889978), investigators conducted a prespecified sub study that included 4077 individuals in an independent validation set (cancer: n = 2823; non-cancer: n = 1254).

Results indicated that the use of the multi-cancer early detection test using cell-free DNA sequencing in combination with machine learning produced an overall accuracy rate for detecting the cancer signal of origin of 88.7% (95% CI, 87.0%-90.2%). The specificity for cancer signal detection was 99.5% (95% CI, 99.0%-99.8%), and the overall sensitivity for cancer signal detection was 51.5% (95% CI, 49.6%-53.3%).⁵

Concluding Thoughts: Going for Gold

“We really have to start thinking beyond the blood and looking into the cerebrospinal fluid (CSF) and the pleural fluid in patients with lung cancer because [that information] can give us the same information as we are waiting for in the blood. The assays are not there [however], and everything has to be standardized in terms of the sample collection and in terms of indications where this could be done. But particularly for brain metastases, it’s been shown that information from the CSF is more accurate and correlates more with the nature of the disease than a tissue biopsy. Does it mean that we are not going to do brain biopsies? Probably yes, in the future, but we are not there yet.”

“[With regard to] the pleural fluid, [we receive] an enormous…sample, and most of the pathology laboratories are making the cell blocks and trying to do oncogenic molecular testing on cell blocks, [which is] very time consuming. But the pleural effusion is an excellent sample, usually rich in tumor cells, and the sensitivity and specificity is high and overperforms plasma,⁶ but we are not there yet,” Dacic said.

References

1. Dacic S. Novel liquid biopsy data beyond ctDNA. Presented at: 24th Annual International Lung Cancer Congress; July 25-27, 2024; Huntington Beach, CA.
2. Liu X, Jiang T, Li X, et al. Exosomes transmit T790M mutation-induced resistance in EGFR-mutant NSCLC by activating PI3K/AKT signaling pathway. J Cell Mol Med. 2020;24(2):1529-1540. doi:10.1111/jcmm.14838
3. Wu S, Luo M, To KKW, et al. Intercellular transfer of exosomal wild type EGFR triggers osimertinib resistance in non-small cell lung cancer. Mol Cancer. 2021;20(1):17. doi:10.1186/s12943-021-01307-9
4. De Miguel-Perez D, Russo A, Arrieta O, et al. Extracellular vesicle PD-L1 dynamics predict durable response to immune-checkpoint inhibitors and survival in patients with non-small cell lung cancer. J Exp Clin Cancer Res. 2022;41(1):186. doi:10.1186/s13046-022-02379-1
5. Klein EA, Richards D, Cohn A, et al. Clinical validation of a targeted methylation-based multi-cancer early detection test using an independent validation set. Ann Oncol. 2021;32(9):1167-1177. doi:10.1016/j.annonc.2021.05.806
6. Bao Y, Zhang D, Guo H, et al. Beyond blood: advancing the frontiers of liquid biopsy in oncology and personalized medicine. 2024;115(4):1060-1072. doi:10.1111/cas.160097