Ushering in Liquid Biopsy for the Microsatellite Status: Advantages and Caveats

In this issue of Clinical Cancer Research, Georgiadis and colleagues (1) and Willis and colleagues (2) have reported liquid biopsy–based assessments of the microsatellite status in patients with cancer. The reports are on the feasibility and performance of the assays while validating results in tumor tissues and in some patients receiving immune checkpoint blockade (ICB). The results of both studies suggest that liquid biopsy–based microsatellite status is a viable and sensitive option with high translational potential. The availability of these tests in the clinic will likely be palpable and can circumvent a few unmanageable circumstances where tumor tissue is simply unavailable/sparse.

Microsatellite instability (MSI), albeit uncommon, is highly associated with durable tumor response and long overall survival in >50% of patients. MSI is often observed in localized uterine/endometrial carcinoma (28%), colon (17%), and gastric (22%) adenocarcinoma, but also observed at low frequency (≤3%) in other cancers (3, 4). The MSI rates reduce considerably in metastatic solid tumors. MSI is a hallmark of Lynch syndrome (in ∼30%) but may also harbor actionable NTRK fusions/mutations. Pembrolizumab (in all cancer types) and nivolumab (in colorectal cancers) have been approved by the U.S. FDA for patients with MSI tumor, thus, necessitating MSI assessment in all patients with an advanced solid tumor. Although, multiple tissue-based tests are available but, in some patients MSI testing is not feasible due to inadequate tumor tissue. Therefore, liquid biopsy–based MSI assessment is attractive. Liquid biopsy has several other advantages (Fig. 1A): sequential assessment for response/resistance monitoring which is problematic in tissue-based assays because of sampling issues and inherent intratumoral heterogeneity (ITH). In general, liquid biopsy (DNA, RNA, mutated proteins, post-translational modifications, fusions, exosomes, miRNA, etc.), when it achieves a higher level of technologic sophistication, holds tremendous promise of early cancer detection, individualized therapy, disease monitoring, resistance characterization, and immune profiling (cytokines, TCR/BCR, etc.) among others.

MSI is a genomic alteration that involves mainly the microsatellites, short tandem nucleotide repeats (mono-, bi-, tri-, tetra-), abundantly dispersed throughout the genome, characterized by gains and/or losses of nucleotides within the repetitive sequences of microsatellite tracts. Many approaches are available for MSI assessment. IHC, probably, the least reliable and tissue consuming, was developed to assess integrity of DNA mismatch repair proteins. However, DNA-based methods [PCR and next-generation sequencing (NGS)] may be far more reliable. Early testing of MSI was performed on only a few canonical microsatellite loci, via PCR followed by capillary electrophoresis analysis. Over time, the NGS technologies and computational algorithms (5) have allowed unbiased, genome-wide screening of the molecular fingerprints of MSI (3), greatly increased the sensitivity of MSI detection. However, sensitive detection of MSI in liquid biopsies via circulating-tumor DNA (ctDNA) or cell-free DNA (cfDNA) still appears to be in early development due to: (i) great technical and bioinformatics challenges (Fig. 1B) for efficient molecular capture, sequencing, mapping, variant calling, error correction at microsatellite loci, and the highly repetitive genomic context and (ii) low tumor fraction in circulation and a high level of technical noise due to polymerase slippage. To address these challenges, Georgiadis and colleagues (1) developed a hybrid-capture–based 98-kb pan-cancer gene panel which covers the targeted microsatellite loci. They applied a multifactorial error correction approach for accurate calling of insertions and deletions (indels) in cfDNA fragments and a digital peak finding algorithm for quantification of MSI-H alleles. Willis and colleagues (2) utilized digital sequencing error correction method to define true indels at microsatellite loci. Both studies applied molecular barcoding (uniquely labels each DNA template) to help eliminate false positives and accurately quantify indels among duplicated fragments. These results by Willis and colleagues show high overall accuracy and is a significant technology advance.

Key factors that could influence accurate detection of MSI and lead to discrepancy between different studies include: (i) the number and type of microsatellite markers used in a study. For example, Georgiadis and colleagues (1) utilized a hybrid-capture–based 98-kb gene panel for both MSI detection and tumor mutation burden (TMB) measurement, while Willis and colleagues (2) applied Guardant360 and selected 90 informative microsatellite loci. Different microsatellite loci and gene panels have been proposed for sensitive detection of MSI and proper selection of microsatellite targets and microsatellite panels are the key. Notably, both the quantity and the quality of the microsatellite loci matters. Willis and colleagues (2) demonstrated that some microsatellite loci, including a couple of widely selected microsatellite loci in traditional MSI tests, performed poorly during capture, with either low/no coverage or a noisy background. Therefore, performance assessment is highly recommended for a capture design, which helps evaluate whether the microsatellite loci can be effectively captured, sequenced, and mapped and are truly informative for MSI testing. In addition, tumor-type specificity has been reported for some well-known targets of MSI, which again draws attention to the panel design. (ii) Tumor cell fraction. Willis and colleagues (2) showed that the efficiency of MSI detection declines remarkably for samples with a high level of normal cell contamination. For a subset of patients whose liquid biopsies had a low tumor cell fraction (<0.2%), MSI assessment will be technically unevaluable and the analysis will not be informative. (iii) The heterogeneous and dynamic nature of cancer. Intertumoral heterogeneity and ITH represent a fundamental property of cancer and also a key factor contributing to discrepancy between patients/cohorts. In addition, tumor cells evolve in response to environmental and therapy stress and clonal evolution may alter the genomic architecture. Nevertheless, compared with tissues specimen, liquid biopsy is expected to better capture the ITH as it is not limited by region or size of the biopsy and sequential sampling is more ideal.

Willis and colleagues (2) presented the clinical outcomes in a small cohort of patients with cfDNA MSI-H gastric cancer treated with pembrolizumab or nivolumab and demonstrated a similar response rate (an objective response rate of 63%) as previously reported for MSI-H patients defined by tissue testing. Georgiadis and colleagues (1) analyzed cfDNA sequentially during pembrolizumab therapy in patients with metastatic cancers and validated that direct detection of MSI in baseline cfDNA predicts progression free and overall survival to ICB. In addition, they reported an inverse correlation between the overall and progression-free survival with the residual MSI allele levels at last dose, indicating that cfDNA residual MSI allele may serve as an earlier molecular predictor of response than radiographic imaging. However, it is noteworthy that not all MSI-H patients benefit from ICB, the study from Willis and colleagues (2) showed that approximately 40% of MSI-H gastric cancer cases did not respond to such therapy. Similarly, <50% of patients with endometrial cancer with MSI-H tumor respond to ICB, suggesting other factors than MSI/TMB determine response thus these mechanisms are yet to be elucidated. On the other hand, some microsatellite stable tumors are highly sensitive to ICB, such as a unique subset of colon/endometrial cancers characterized by a hypermutated phenotype associated with POLE/POLD1 hotspot mutations. In such cases, a TMB measurement as performed by Georgiadis and colleagues (1) would be extremely useful.

Finally, the major impact of liquid biopsy in general and also for MSI assessment would be higher in localized cancer (higher rate of MSI and value in early detection), however, in this setting, ctDNA fraction is expected to be low thus reducing its potential. In this circumstance, tissue-based assessments supersede liquid biopsy till more advancements are made.

No potential conflicts of interest were disclosed.

This work was supported by the start-up research funds kindly awarded to L. Wang by U.T. MD Anderson Cancer Center. This work was also supported in part by the U.S. Department of Defense (DOD) grant-CA160445 to J.A. Ajani.