Microsatellite Instability Detection in Cancer: A Multiplex qPCR Approach that Obviates the Need for Matching Normal Samples

Abstract

Background

Microsatellite instability (MSI) indicates DNA mismatch repair deficiency in certain types of cancer, such as colorectal cancer. The current gold standard technique, PCR–capillary electrophoresis (CE), requires matching normal samples and specialized instrumentation. We developed VarTrace, a rapid and low-cost quantitative PCR (qPCR) assay, to evaluate MSI using solely the tumor sample DNA, obviating the requirement for matching normal samples.

Methods

One hundred and one formalin-fixed paraffin-embedded (FFPE) tumor samples were tested using VarTrace and compared with the Promega OncoMate assay utilizing PCR-CE. Tumor percentage limit of detection was evaluated on contrived samples derived from clinical high MSI (MSI-H) samples. Analytical sensitivity, specificity, limit of detection, and input requirements were assessed using synthetic commercial reference standards.

Results

VarTrace successfully analyzed all 101 clinical FFPE samples, demonstrating 100% sensitivity and 98% specificity compared to OncoMate. It detected MSI-H with 97% accuracy down to 10% tumor. Analytical studies using synthetic samples showed a limit of detection of 5% variant allele frequency and a limit of input of 0.5 ng.

Conclusions

This study validates VarTrace as a swift, accurate, and economical assay for MSI detection in samples with low tumor percentages without the need for matching normal DNA. VarTrace's capacity for highly sensitive MSI analysis holds potential for enhancing the efficiency of clinical work flows and broadening the availability of this test.

Introduction

Microsatellite instability (MSI) is characterized by length changes in repetitive DNA sequences called microsatellites and is often caused by deficiencies in DNA mismatch repair (dMMR) (1, 2). MSI is an important biomarker in several cancers (3, 4), most notably colorectal cancer (1, 5) where approximately 15% of patient tumors are classified as having high MSI (MSI-H) (6). MSI can indicate dMMR tumors and is also an important screening factor for Lynch syndrome (6–8), an inherited disorder caused by germline mutations in mismatch repair (MMR) genes that confers a high lifetime cancer risk. Patients with MSI-H tumors have a favorable prognosis and derive great benefit from immunotherapy compared to aggressive chemotherapy (9–17).

Current standard methods for detecting MSI and dMMR in colorectal cancer include immunohistochemistry (IHC), PCR amplification of microsatellite markers followed by capillary electrophoresis (PCR-CE) (18–20), or next-generation sequencing (NGS) (21–29) as recommended by the National Comprehensive Cancer Network. IHC can reflect MSI status through identifying a loss of MMR protein expression but may miss some MSI tumors with intact protein expression (30, 31). NGS and PCR-CE directly classify tumors as MSI-H or microsatellite stable (MSS) but often require comparison to matched normal tissue (18–27). While NGS can accurately assess MSI and MMR gene status from minute inputs (26), it requires specialized instrumentation. PCR-CE is considered the gold standard for MSI detection due to its high efficiency and sensitivity in detecting repeat markers. The Bethesda Guidelines recommend a panel of 5 well-characterized homopolymer markers for MSI testing by PCR-CE: BAT-25, BAT-26, NR-21, NR-24, and MONO-27 (32). However, neither IHC nor PCR-CE alone ensures accurate diagnosis (18, 30, 31, 33, 34). Using both methods simultaneously increases sample requirements and screening costs substantially. Moreover, the multistep PCR-CE process still demands advanced instrumentation, with only marginal cost savings over NGS. Alternative methods such as droplet digital PCR (33) and targeted NGS panels (26–29) have also emerged for accurate MSI screening, but require further validation.

To reduce costs and sample input while preserving the sensitivity for MSI detection, here, we introduce a new quantitative PCR (qPCR) assay, NuProbe VarTrace, which utilizes blocker displacement amplification (BDA) technology (35) to preferentially amplify MSI alleles over wild-type sequences. Our assay requires only 0.5 ng DNA input and can directly detect MSI in DNA samples extracted from just 10% tumor content without pre-amplification. Combined with automated data analysis, MSI status can be determined within 8 h of sample receipt, which is faster than existing PCR-CE protocols that require at least 10 h or an overnight process. This qPCR approach could potentially be applied to both tumor tissue and cell-free DNA pending clinical validation. By enabling rapid, robust MSI detection from minute DNA quantities, this method could greatly expand test accessibility in laboratories with qPCR instruments and facilitate Lynch syndrome screening and precision medicine in colorectal cancer management.

Materials and Methods

Samples

Formalin-fixed paraffin-embedded (FFPE) samples were obtained from several sources. Twelve samples were purchased from the BioChain Institute, and included 7 confirmed to have MSI status by IHC and 5 confirmed by the Promega OncoMate MSI Dx assay. Forty-one adenocarcinoma samples were purchased from OriGene, with MSI status undetermined. Additionally, 48 samples with MSI status confirmed by PCR were purchased from Discovery Life Sciences. All clinical sample information was collected from vendors and is summarized in Table 1 and online Supplemental Tables 1–3. This study was conducted in accordance with the Declaration of Helsinki. The research protocol was approved by the Institutional Review Board at NuProbe USA Inc (IRB number: NP257).

Table 1.

Sample information.

		Totala	MSI-Hb	MSI-L/MSSb	NAb
VARIABLES		101	45	50	6
SEX	Female	46	20	23	3
	Male	41	12	26	3
	Unknown	14	13	1	0
AGE	≥50	59	19	37	3
	<50	12	3	8	1
	Unknown	30	23	5	2
STAGE	I	12	9	3	0
	II	25	18	7	0
	III	31	13	14	4
	IV	29	2	25	2
	Unknown	4	3	1	0
TUMOR TISSUE TYPE	Primary	73	44	25	4
	Metastatic	28	1	25	2
TUMOR LOCATION	Colon	61	40	18	3
	Rectum	5	1	3	1
	Other (cecum, small intestine, stomach, liver)	32	1	29	2
	Unknown	3	3	0	0
TUMOR ORIGIN	Colon	87	41	42	4
	Rectum	7	1	4	2
	Other (cecum, small intestine, stomach)	4	0	4	0
	Unknown	3	3	0	0

^aFFPE samples are all commercially purchased from vendors, with a focus on samples with known MSI status from BioChain and Discovery Life Sciences, and samples from colorectal or other related cancer from OriGene.

^bFFPE sample information is categorized by MSI status detected using Promega OncoMate MSI Analysis, as this is the current gold standard. Promega reports indicate ‘NA’ when either the marker or the sample fails in the test.

Seraseq^® MSI Reference Panel Mix AF5% (Seracare) was used as commercial reference standard for sensitivity, specificity, and limit of input (LoI) study. In addition, VarTrace positive and negative controls were also included for the sensitivity, specificity and LoI analyses. The VarTrace negative control is background human genomic DNA GM25485, which is known to be microsatellite stable (MS)S. The VarTrace positive control is a mixture of GM25485 and 5 synthetic MSI-H DNA fragments, and a variant for each marker was quantified as approximately 5% variant allele frequency (VAF) using qPCR individually. DNA fragments used in the positive control are plasmid-synthesized and purified in GenScript. Additional wild-type (WT) and variants templates used in limit of blank (LoB) and limit of detection (LoD) study, respectively, are synthetic gBlock fragments ordered from IDT; sequences are listed in online Supplemental Table 4.

NuProbeVarTrace MSI Assay

NuProbe developed the VarTrace qPCR assay to detect MSI using a previously published BDA technology without matching normal samples. BDA can manipulate reaction thermodynamics by introducing a rationally designed set of blocker oligonucleotides that perfectly bind to WT targets and share the same binding region with the primer (Fig. 1A). Since the blocker oligonucleotides perfectly match the target, they bind more strongly to the WT targets than the primer, thereby inhibiting the primer from binding and extending. In contrast, a mismatch bulge will be created when the blocker binds to the variant template. This bulge weakens the blocker's binding stability, enabling binding of the primer and preferential amplification of the variant. Consequently, the BDA technology enables the detection of low-frequency variants by preferentially amplifying these variants over WT templates, which is then reflected in the fluorescence signals observed in the qPCR setting.

Fig. 1.

Detecting microsatellite instable variants using qPCR with BDA. (A), Single-plex design diagram. Blocker sets (long solid block) overlapping the forward primer binding site are designed to perfectly match common WT sequences: WT (0, ±1, ±2, ±3). This allows the blockers to outcompete forward primer binding to WT sequences. WT (0) represents the reference sequence with the highest population frequency, and WT (+1) represents sequences with a 1 nucleotide (nt) addition in the mononucleotide repeat region relative to WT (0). MSI variants (short solid block) create a mismatch that enables forward primer binding and PCR amplification. (B), qPCR detection of NR-21 variants in synthetic MSI-H samples with different VAFs. WT (0) and no-template controls are shown in solid gray and dashed black, respectively. All variants and WT curves are tested and shown in triplicate; No template control (NTC) is a single test. (C), Overview of the VarTrace assay composition and coverage. The VarTrace MSI Kit comprises 2 tubes for 5 markers in total. Tube 1 is designed to target NR-21, NR-24, and BAT-25, while Tube 2 targets BAT-26 and MONO-27. Both tubes also contain primers and probes specific for GAPDH as an internal reference control, without the use of a blocker for GAPDH. Frequencies of genotypes covered by WTs (−3 to +3) were determined from population sequence data reported in a previous publication (18). Additionally, the microsatellites for WT (0) are listed for all 5 markers, with poly-A lengths of 21, 23, 26, 25, and 27 nt, respectively. Color figure available at https://academic.oup.com/clinchem.

BDA Design Principle for MSI Markers

As MSI WT alleles vary in homopolymer lengths across populations (18) and polymerase slippage causes stutter artifacts, multiple blockers per marker are designed to ensure comprehensive suppression. A set of 7 blockers corresponding to WT 0, ±1, ±2, and ±3 bp (to span BDA coverage) is simultaneously included in the qPCR reaction (Fig. 1A). WTs (0) are selected to maximize coverage of natural variation in WT microsatellite length, enabling MSI detection without requiring a matched normal sample (Fig. 1C).

Assay Composition

The VarTrace MSI Kit comprises 2 tubes of 4× oligo mixes that can detect 5 MSI markers selected from the Bethesda guidelines: NR-21, NR-24, BAT-25, BAT-26, and MONO-27 (Fig. 1C), a positive control, and a negative control. Primers targeting GAPDH are included as a reference control in both tubes of oligo mixes to evaluate and quantitate sample quality and input amount. The background DNA used in both the positive and negative controls is cell line GM25485 (BioChain), which is known to be MSS. TaqPath ProAmp Multiplex Master Mix (Applied Biosystems) is used to conduct all VarTrace experiment. Sequence information is listed in Supplemental Table 4.

Assay Implementation

All VarTrace experiments were conducted in strict accordance with the provided user manual (online Supplemental Methods). Positive (5% VAF, 5 ng input per tube) and negative controls (5 ng input per tube) are included as in-plate controls for each run. qPCR readouts are performed using the Applied Biosystems 3500 Fast Dx, as detailed in Table 2.

Table 2.

Comparison of operating characteristics and specification between Promega OncoMate and NuProbe VarTrace.

	Promega OncoMate	NuProbe VarTrace
Normal sample required	Yes	No
Instrument setting	ABI 7500 Fast (qPCR) and ABI 3500 Dx (Sanger)a	ABI 7500 Fast (qPCR)
Sample throughputb	12	46
Time to result	2-day result	Same-day result
Tumor percentage of FFPE sample	≥30%	≥10%
Input requirement	1 ng	0.5–10 ng
Success rate	94%	100%

^aABI 7500 Fast (qPCR): Applied Biosystems™ 7500 Fast Dx Real-Time PCR; ABI 3500 Dx (Sanger): Applied Biosystems 3500 Dx Genetic Analyzer.

^bSample throughput is calculated based on instrument specification. ABI 3500 Dx (Sanger) throughput is 24 capillaries per run, including matched samples. ABI 7500 Fast (qPCR) throughput is 96 wells per run, including 2 tubes per sample as well as 2 positive controls.

Data Analysis

The QuantStudio 3/5 Real-Time PCR Design & Analysis 2 software is used to extract quantification cycle (Cq) values under default settings. Upon successful Cq extraction, MSI status can be analyzed using the following step-by-step analysis pipeline (Supplemental Methods and online Supplemental Fig. 1):

For each sample:

$$\text{Sample Successful, if}C{q}_{\mathit{G}\mathit{A}\mathit{P}\mathit{D}\mathit{H}}^{\mathrm{Sample}}\le 37;\text{else Failed}.$$

For each marker:

$$\mathrm{\Delta }C{q}_{\mathrm{marker}}=C{q}_{\mathrm{marker}}-C{q}_{\mathit{G}\mathit{A}\mathit{P}\mathit{D}\mathit{H}}$$

$$\text{Sample Marker Positive},\mathrm{if}\mathrm{\Delta }C{q}_{\mathrm{marker}}^{\mathrm{Sample}}-\mathrm{\Delta }C{q}_{\mathrm{marker}}^{\mathrm{PosCtrl}}\le 1$$

Sample call:

• MSI-H, if the number of positive markers > 1

• MSI-L, if the number of positive markers = 1

• MSS, if the number of the positive markers = 0

Both low MSI (MSI-L) and MSS are considered as negative in the final report, and only MSI-H is considered as positive.

MSI Testing using the PCR-CE Method

Promega OncoMate MSI Dx is an assay cleared by the FDA to determine MSI status, and considered to be the gold standard. All FFPE samples underwent MSI testing using the Promega OncoMate MSI Dx Analysis assay per the manufacturer's protocol. Testing was performed by a third-party CLIA-certified lab, which carried out the DNA extraction, quantitation, and Promega assay analysis on the samples. Leftover DNA was returned to NuProbe and stored at 4°C for 2 years prior to additional testing.

DNA Extraction and quantification

For the OriGene and BioChain samples, DNA was extracted from one slide per sample without tumor preselection using the GeneRead FFPE Kit (Qiagen) per manufacturer's protocol with a slight modification (details in Supplemental Methods).

Considering only one slide per sample was obtained from Discovery Life Sciences, DNA extraction was performed in the third-party CLIA lab for Promega OncoMate testing. Leftover DNA was returned and stored at 4°C. Prior to the VarTrace assay test, a portion of the DNA samples from Discovery Life Sciences was rehydrated, purified, and concentrated for a second time to remove any magnetic bead residue.

DNA concentration was quantified prior to VarTrace testing using an Invitrogen Qubit 4 fluorometer (ThermoFisher Scientific) and the Qubit™ dsDNA Quantification Assay, High Sensitivity (ThermoFisher Scientific).

Results

Blocker displacement amplification enables qPCR Detection of MSI At low VAFs

The feasibility of single-plex qPCR with BDA in detecting MSI was first demonstrated by amplifying NR-21 instable variants spiked into MSS background DNA at VAFs down to 0.1% (Fig. 1B). There was a notable Cq separation of over 4 cycles between the WT and 0.1% VAF samples, while no template controls showed no amplification. All 5 markers were tested individually with samples with a range of VAFs, and calibration curves for all 5 markers from 0.1% to 50% VAF showed high coefficients of determination (R²) ranging from 0.93 to 0.99 (online Supplemental Fig. 2).

Detection of MSI In FFPE Tumor samples using theVarTrace assay

A total of 101 FFPE tumor samples obtained from commercial sources were tested using the VarTrace assay, of which 60 had known MSI status from the vendor. Successful DNA extraction, quantitation, and qPCR testing were achieved in the first round for all 53 samples extracted in-house. The remaining 48 samples from Discovery Life Sciences were extracted by a third-party CLIA lab. Leftover DNA was sent back and stored at 4°C for 2 years. All 48 samples were successfully analyzed. In addition, the 101 samples were all successfully analyzed on the first pass with no invalid results or need to retest.

The 101 samples encompassed a wide range of tumor percentages from 10% to 90% (median 40%). The VarTrace assay detected MSI-H in 47 samples, with tumor percentages ranging from 15% to 80% (median 45%) (Fig. 2A). Most MSI-H samples (40/47) were positive for ≥4 markers, while the remaining 7 samples with <50% tumor content were positive for only 2 to 3 markers (Fig. 2C). The MONO-27 marker showed no amplification in MSI-negative samples, while BAT-25 showed positive signals in all MSI-positive samples, indicating potential biases in sample selection. Employing samples with a broader diversity in marker deletion lengths and variant frequencies can potentially enhance the assay's sensitivity and specificity through precise adjustment of the threshold for reporting positive markers.

Fig. 2.

MSI results of 101 FFPE tumor samples. Details on sample information are provided in the Materials and Methods and Table 1. (A), Tumor percentage distribution across all samples and the MSI-H subset. Background shading indicates samples below the 30% tumor percentage, which is below the Promega OncoMate specification for sample input sample. (B), Representative qPCR curves from an MSI-H sample (C504247) showing positive detection of all 5 markers. Tube 1 can detect 3 markers and the remaining 2 markers are detected in Tube 2. GAPDH is present in both tubes and used as a reference control. (C), Summary heatmap of MSI results from VarTrace for the 5 markers and overall MSI status. Samples are grouped by positive (MSI-H) and negative (MSI-L/MSS) status and sorted by decreasing tumor percentage within each group. White indicates positive detection or MSI-H status; black indicates no detection or MSI-L/MSS status. Sample C504247 in (B) is outlined with a box. (D), Concordance of MSI status between the 95 samples successfully tested by both VarTrace and OncoMate methods. Six samples failed because of loss of marker signal in the OncoMate analysis. Diagnostic sensitivity is 100% (95% CI, 92%–100%), and diagnostic specificity is 98% (95% CI, 89%–100%). (E), Timeline comparison of the work flow steps between the VarTrace and OncoMate assays. See Table 2 for additional assay specification comparisons. Color figure available at https://academic.oup.com/clinchem.

Comparison of theVarTrace assay toPromegaOncoMate MSI Analysis

The MSI status of the 101 FFPE samples was also tested using the Promega OncoMate assay at a third-party CLIA lab (online Supplemental Table 5 and online Supplemental Figs. 3–10). However, 6 samples showed invalid results due to marker readout failure, yielding a sample test success rate of 94%. Of the remaining 95 samples, the VarTrace and OncoMate assays were concordant with the exception of one sample that scored as MSI-H by VarTrace but MSI-L by OncoMate. The one discrepant sample was confirmed as MSI-H by the vendor report, suggesting OncoMate likely missed low-frequency MSI signals due to limitations in its pre-amplification efficiency and analysis algorithm (Supplemental Fig. 11). Considering both MSI-L and MSS as negative and MSI-H as positive, the VarTrace assay demonstrated 99% concordance, 100% (95% CI, 92%–100%) sensitivity and 98% (95% CI, 89%–100%) specificity compared to OncoMate (Fig. 2D).

Receiver operating characteristic (ROC) analysis summarizing the 5 markers (online Supplemental Results and online Supplemental Figs. 12–14) was performed, yielding the area under the ROC curve (AUC) values of markers ranging from 0.91 to 0.98. Under the default algorithm settings, VarTrace demonstrated >95% specificity for 3 markers (NR-21, BAT-26, and MONO-27) and >95% sensitivity and concordance for the 3 other markers (BAT-25, BAT-26, and MONO-27).

Comparison ofVarTrace andOncoMate assays to vendor validation

Of the 101 samples, 60 had MSI status validated by the commercial vendor, with 44 determined to be MSI-H. Compared to the vendor results, VarTrace concordantly classified 59/60 samples, with one sample scored as MSS by VarTrace but MSI-H by the vendor (online Supplemental Tables 5 and 6). This yielded a sensitivity of 98% (95% CI, 88%–100%) and specificity of 100% (95% CI, 79%–100%) for VarTrace.

Similarly, the OncoMate assay had 2 samples that scored negative, which the vendor claimed to be MSI-H. After excluding 3 invalid samples from the validation dataset, the OncoMate assay demonstrated 95% sensitivity (95% CI, 84%–99%) and 100% specificity (95% CI, 77%–100%) vs vendor validation (Supplemental Table 6). Therefore, the VarTrace assay showed slightly better concordance with vendor results compared to OncoMate.

Determination of tumor percentageLoD

Serial dilutions using WT DNA to make contrived samples of different tumor percentages were performed on 41 MSI-H samples (Fig. 3A and online Supplemental Figs. 15–35), with each sample tested once at each dilution due to the limited sample amount. The LoD of one sample is defined as the smallest tumor percentage with a positive result, and all tests with higher tumor percentages also reported as positive. When evaluating the LoD of each individual sample, a tumor percentage LoD below 5% was achieved in 35/41 samples, with only one sample showing LoD above 10% (Fig. 3B and online Supplemental Table 7). This sample, originally carrying a tumor percentage of 30% and testing positive for 3 markers, may suggest it is an early-stage mutated MSI-H sample with a limited number of detectable markers. The process of serial dilution using WT DNA could result in the loss of detection for these positive markers, potentially resulting in a higher LoD for this specific sample. All of the samples that tested positive for 5 markers when undiluted DNA was tested had a tumor percentage LoD ≤5%. Similarly, all of the samples that tested positive for 4 markers at original testing showed tumor percentage LoD ≤10%.

Fig. 3.

Tumor percentage (TP) LoD. (A), Contrived sample dilution demonstration. WT genomic DNA from cell line GM25485 was used as the MSS background ( tube on the top left) for all contrived sample dilutions. DNA extracted from an FFPE slide (C504247, 40% tumor percentage) was quantified by GAPDH in the original VarTrace test (Fig. 2B) and used as the starting MSI-H material ( tube on the top right). The WT genomic DNA and FFPE DNA were mixed to create contrived samples at the desired tumor percentage (tube at bottom). Serial dilutions were made to generate LoD datasets for this sample, with each dilution tested once. LoD was defined as the lowest tumor percentage at which all samples with ≥ LoD tumor percentage scored as MSI-H. (B), TP LoD was determined by diluting 41 MSI-H samples with WT sample. Dot colors indicate the number of positive markers detected from the undiluted sample. (C), Positivity of the VarTrace assay across samples grouped into tumor percentage brackets, with the number of samples per bracket showed in the bar graph. The values in brackets/parentheses on the x-axis represent the range of tumor percentage. For example, “[1,5)” represents a tumor percentage of ≥1% but <5%. Color figure available at https://academic.oup.com/clinchem.

A total of 246 tests across varying tumor percentages were included in the analysis to evaluate the accuracy of the VarTrace assay at different tumor content levels, of which 41 tests using the original, undiluted sample were included. All tests were considered independent experiments, regardless of differences in starting material used to create the contrived diluted specimens. The positivity rate was calculated at different tumor percentage ranges to assess the performance of the VarTrace assay as tumor content decreased. The VarTrace assay demonstrated 100% positivity for samples with ≥15% tumor, 97% positivity from 10% to 15%, and 92% positivity from 5% to 10% tumor percentage (Fig. 3C). Additionally, 5 tests were scored MSI-H at tumor percentages below 1% (online Supplemental Table 8). These results indicate reliable MSI detection by the VarTrace assay even at low tumor content.

Analytical performance ofVarTrace assay using synthetic samples

The analytical sensitivity and specificity of the VarTrace assay was evaluated using 2 types of synthetic samples: Seraseq MSI Reference Panel Mix AF5% and internally generated VarTrace positive controls (5% VAF) and negative controls. The VarTrace assay demonstrated 100% sensitivity and 100% specificity in detecting all 5 markers when tested with both 2 ng input (1 ng per reaction tube) of the Seraseq reference samples and 10 ng of the internal controls (Table 3, online Supplemental Table 9 and online Supplemental Figs. 36–37). No false–positive results occurred for any markers in the negative control samples in either analytical study.

Table 3.

Analytical performances of VarTrace using synthetic samples.

Analytical validation
Samples	Seraseq MSI Reference	VarTrace internal control
Input	2 ng	10 ng
Specificity	100% (CI: 83% to 100%)	100% (CI: 83% to 100%)
Sensitivity	100% (CI: 83% to 100%)	100% (CI: 83% to 100%)

Limit of blank: 10 ng input
Samples	WT (+3) 100%VAF	WT (0)	WT (−3) 100%VAF
Concordance	12/12 (100%)	12/12 (100%)	12/12 (100%)

Limit of detection: 10 ng input
Concordance at samples	Synthetic reference material (−5)	Synthetic reference material (−7)
50%VAF	6/6 (100%)	6/6 (100%)
10%VAF	6/6 (100%)	6/6 (100%)
5%VAF	4/6 (67%)	6/6 (100%)
1%VAF	0/6 (0%)	4/6 (67%)

Limit of input
Concordance at input	Synthetic reference material (−7) 5% VAF	Seraseq MSI Reference 5%VAF
1 ng	6/6 (100%)	6/6 (100%)
0.5 ng	6/6 (100%)	6/6 (100%)
0.25 ng	6/6 (100%)	6/6 (100%)
0.1 ng	5/6 (83%)	6/6 (100%)

To assess the LoB, synthetic WT fragments for 5 markers were tested in the following groups: WT (0), WT (+3) with 3 nt addition, and WT (−3) with 3 nt deletions in the mononucleotide repeat region. Across 12 repeats of each WT group, no WT genotypes demonstrated positive signals for any markers, indicating 100% correct negative scoring within the WT allele length range (Table 3, online Supplemental Table 10 and online Supplemental Figs. 38–39).

The LoD was evaluated using variants with 7 nt deletions in the mononucleotide repeat regions of all 5 markers (MSI-7 variant pool) at VAFs ranging from 1% to 50%. These variants were designed to be near the WT allele length range and difficult to distinguish from WT. The results showed the VarTrace assay could detect and score all experimental repeats containing the −7 variant at 5% VAF as positive (Table 3, online Supplemental Table 11 and online Supplemental Figs. 40–43). This is consistent with detection of the BAT-25, NR-24, and MONO-27 variants with 6 nt deletions present at 5% VAF in the Seraseq reference sample from the analytical study.

The LoI was assessed using the Seraseq Positive Reference (5% VAF) and MSI-7 variant pool (5% VAF), which are samples that have low tumor content and short microsatellite deletions. Input amounts of 1 ng to 0.1 ng per tube of these 2 samples were tested 6 times. The VarTrace assay yielded positive results for all 5 markers in all 6 repeats with as little as 1 ng input (0.5 ng per tube) for both samples at 5% VAF (Table 3, online Supplemental Table 12 and online Supplemental Figs. 44–45). Evaluating MSI classification, the VarTrace assay detected MSI-H status with occasional marker dropout at all repeats in both samples down to 0.5 ng input (0.25 ng per tube).

Assay robustness and reproducibility were further evaluated by repetitive testing of the positive control sample across 41 different PCR plates, 2 reagent batches, and 5 operators. All 5 markers achieved high ΔCq concordance of less than one cycle compared to the median ΔCq value in more than 95% of the total 96 tests analyzed (Supplemental Results and Supplemental Figs. 46–47).

Discussion

VarTrace is a qPCR assay that can detect MSI status using 5 Bethesda Guideline-recommended markers directly from tumor sample DNA without matching normal samples. Our study demonstrated 100% sensitivity and 98% specificity for the VarTrace assay compared to the OncoMate assay across 101 FFPE samples. VarTrace also performed better than OncoMate when benchmarked to the vendor pathology reports. Beyond comparable accuracy and sample requirement benefit, VarTrace also showed advantages in industrial adaptability over OncoMate. Contrary to the two-step process required by OncoMate, involving pre-amplification followed by a CE readout, the VarTrace assay platform incorporates qPCR directly as the readout method for samples. This simplification of the work flow reduces instrumentation requirements, as shown through comparison of the timelines in Fig. 2E. This enabled VarTrace to be completed in 8 h, which is faster than the 10+ h OncoMate timeframe and better suited for shift work. The higher throughput scalability of qPCR vs CE also enables VarTrace to process 3-fold more samples per run (Table 2).

Studies of tumor percentage dilution and analytical performance using synthetic samples showed the VarTrace MSI assay could detect MSI with 97% positivity at tumor content as low as 10% and DNA input as little as 0.5 ng.

However, when markers were considered individually, the VarTrace sensitivity and specificity for NR-24 was lower at 80% and 90%, respectively, compared to the OncoMate. Discordant NR-24 results showed deletions of 4 to 7 nt in the mononucleotide repeats. In an additional LoD study using an MSI-5 10% variant pool (Supplemental Table 11), VarTrace also demonstrated 2 false negatives for NR-24 out of 6 repeats. This indicates less sensitive detection of short deletions <7 nt by NR-24 in the VarTrace assay. However, NR-24 did show positive amplification without reaching the calling threshold in these discrepant samples. Re-evaluating the NR-24 calling threshold could potentially improve the accuracy of this marker.

In addition, MSI variants with short indels (i.e., MSI-7) have weaker amplification efficiency compared to long-indel variants, which can obscure VAF quantitation.

NGS panels can concurrently detect MSI status and alterations in MMR genes from solid tumor samples. FoundationOne CDx (25), as a representative NGS panel, became the first FDA-approved NGS test for pan-cancer biomarker profiling and tumor mutational burden, and includes the analysis of multiple microsatellite markers to determine MSI status. It also enables discovery of novel microsatellite markers across cancer types (3, 4, 21, 25, 27, 29), ovarian cancer (25) for instance. While NGS provides additional benefits, its use for first-line MSI testing in colorectal cancer may be prohibitively expensive and has limited clinical utility beyond determination of MSI status.

The VarTrace assay platform has potential for broader applications through validation of additional MSI markers identified by NGS, as well as alternative sample types like cell-free DNA (cfDNA) (26, 36, 37). cfDNA enables minimally invasive MSI testing compared to tissue biopsies, but detecting MSI in cfDNA requires even higher sensitivity due to limited tumor DNA in plasma. As demonstrated by sensitive detection of each marker individually in single-plex validation (Supplemental Fig. 2), the VarTrace assay shows strong potential for detecting MSI variants from samples with low tumor content as well as cfDNA. VarTrace can better enrich and amplify minimal MSI markers in cfDNA by simply suppressing GAPDH amplification to minimize consumption of PCR reagents by this control gene. With further optimization and calibrated calling thresholds, the same platform may be adapted for sensitive and specific MSI detection from cfDNA pending clinical validation. Expanding the validated markers and sample types would extend the utility of the rapid, robust VarTrace assay for MSI analysis across clinical settings.

In conclusion, our findings demonstrate that the VarTrace assay is a promising new tool for fast, accurate, and broadly applicable MSI testing across clinical settings. This assay stands poised to redefine the standard for MSI detection, extending its relevance not only to colorectal cancer but potentially to other applications as well. The implementation of this assay holds the potential to facilitate Lynch syndrome screening on a grand scale, significantly reduce turnaround times for surgical decisions, and make possible the assessment of MSI status even in constrained biopsy samples or cfDNA. Further clinical validation and implementation studies will help translate these analytical performance improvements into meaningful patient benefits.

Author Declaration

A version of this paper was previously posted as a preprint on medRxiv as https://doi.org/10.1101/2023.11.07.23298217.

Supplemental Material

Supplemental material is available at Clinical Chemistry online.

Nonstandard Abbreviations

MSI, microsatellite instability; CE, capillary electrophoresis; qPCR, quantitative PCR; FFPE, formalin-fixed paraffin-embedded; PCR-CE, PCR-capillary electrophoresis; MSI-H, microsatellite instability high; MMR, mismatch repair; IHC, immunohistochemistry; NGS, next-generation DNA sequencing; MSS, microsatellite stable; MSI-L, microsatellite instability low; BDA, blocker displacement amplification; VAF, variant allele frequency; WT, wild-type; LoD, limit of detection; MSI-L, microsatellite low; nt, nucleotide; cfDNA, cell-free DNA.

Human Genes

GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Author Contributions

The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list.

Wei Chen (Data curation-Lead, Formal analysis-Lead, Methodology-Equal, Writing—review & editing-Equal), Yan Yan (Conceptualization-Lead, Data curation-Equal, Methodology-Equal), Blake Young (Data curation-Supporting, Methodology-Equal), Alessandro Pinto (Investigation-Equal), Qi Jiang (Methodology-Equal, Visualization-Supporting), Nanjia Song (Data curation-Supporting, Visualization-Supporting), Adam Yaseen (Data curation-Supporting, Validation-Supporting, Visualization-Supporting), Weijie Yao (Methodology-Supporting), David Zhang (Project administration-Lead, Resources-Lead, Writing—review & editing-Supporting), and Jinny Zhang (Software-Lead, Supervision-Lead, Visualization-Lead, Writing—original draft-Lead, Writing—review & editing-Lead)

Authors’ Disclosures or Potential Conflicts of Interest

Upon manuscript submission, all authors completed the author disclosure form.

Research Funding

This work is fully funded by NuProbe USA.

Disclosures

All authors are employed by NuProbe USA. Y.H. Yan is listed as inventor and NuProbe USA as the patent applicant on a pending patent on using BDA to detect MSI biomarker under U.S. Patent Application No. 17/499,536. D.Y. Zhang declares a competing interest in the form of consulting for and significant equity ownership in NuProbe Global, Torus Biosystems, Biostate.AI, Pupil Bio, and Pana Bio.

Role of Sponsor

The funding organization played a direct role in the design of study, the choice of enrolled patients, review and interpretation of data, preparation of manuscript, and final approval of manuscript.

Acknowledgments

The authors thank Dr. Mingjie Dai for valuable suggestions for the manuscript. The authors wish to acknowledge the assistance of Jana Harvey and Jonathan Tran in experiment execution and material organization for this study.

Author notes

Wei Chen and Yan Helen Yan contributed equally.