Microsatellite instability (MSI) refers to nucleotide insertions or deletions in microsatellite loci, which are caused by the inactivation of proteins related to DNA mismatch repair (e.g., MLH1, MSH2, MSH6, and PMS2). Deficient mismatch repair (dMMR) from mutation, hypermethylation, or other epigenetic regulation of related genes leads to MSI. MSI occurs in various types of cancers, with the highest incidence in colon and endometrial cancers.[3–6] Approximately 15% of sporadic colorectal cancer (CRC) tumors progress through an MSI-dependent pathway. In addition, MSI is also a characteristic of hereditary Lynch syndrome-related cancer, which accounts for approximately 20% of dMMR CRC.
In the clinical settings, MSI is a key biomarker for cancer companion diagnosis and prognosis prediction for cancer patients deficient in mismatch repair (MMR). Tests for determining MSI or MMR status have been recommended for the diagnosis of Lynch syndrome, and treatment selection of chemotherapy or immunotherapy is based on the recommendations from Tumor Targeted Therapy Professional Committee of China Anti-Cancer Association, the Chinese Society of Clinical Oncology, the Committee of Colorectal Cancer, the Chinese Society of Clinical Oncology, the Genetics Group of the Committee of Colorectal Cancer, the National Comprehensive Cancer Network, and the European Society for Medical Oncology (ESMO) Translational Research and Precision Medicine Working Group. Some studies have indicated that advanced CRC patients with high-level MSI (MSI-H)/dMMR may benefit from immune checkpoint inhibitors therapy—such as antibodies to programmed cell death-1, to its ligand PD-L1. Two immune checkpoint block drugs, pembrolizumab and nivolumab, have shown efficacy in metastatic CRC patients with MSI-H/dMMR.[4,11] The prognosis of CRC patients with MSI-H/dMMR is relatively satisfactory, despite their poor response to fluorouracil-based chemotherapy.
Detection of MSI is usually performed indirectly by immunohistochemical (IHC) analysis of MMR proteins, or directly by MSI polymerase chain reaction (MSI-PCR), which reflect the MSI status. MMR IHC was performed on formalin fixed paraffin embedded specimens (FFPE) to detect the expression of MMR proteins such as MLH1, MSH2, MSH6, and PMS2. Tumors that have completely lost the expression of one or more proteins are defined as dMMR. For MSI-PCR, most assays used to require paired samples. An MSI-PCR panel with two mononucleotide (BAT-25 and BAT-26) and three dinucleotides (D5S346, D2S123, and D17S250) repeat sites is recommended by the National Cancer Institute, USA. Another PCR panel with five poly-A mononucleotide repeat sites (BAT-25, BAT-26, NR-21, NR-24, and NR-27) is recommended by ESMO. The sample is usually defined as MSI-H if two out of five, or more, microsatellite markers are unstable. An optimized PCR method has been developed in recent years, which enables MSI-PCR to use only tumor samples. However, this method has neither been widely incorporated into existing MSI-PCR assays nor is it widely used in China.
Recently, next-generation-sequencing (NGS)-based tests have been widely adopted in clinical applications for comprehensive genomic profiling (CGP) across different cancers. CGP could detect genomic alterations in actionable and druggable genes, including base substitutions, insertions and deletions, copy number alterations, and rearrangements or fusions. Detection of genome-wide markers, such as tumor mutation burden (TMB) and MSI, is also being incorporated into NGS panels. CGP is helpful for identifying therapeutic targets and combined treatment strategies. NGS-based methods or bioinformatical tools have been developed for MSI detection, including MSIsensor, MANTIS, mSINGS, MSI-ColonCore, and MSIsensor-pro.[13–17] An NGS-based MSI detection method could streamline the clinical testing processes by embedding the MSI detection into the genomic profiling test, which could save tissue sample, reduce turn-around time and cost, and provide MSI status and CGP in single test. Herein, we developed a novel NGS-based MSI calling model, Colorectal Cancer Microsatellite Instability (CRC-MSI) test, for CRC patients with only tumor specimens.
The study was approved by the Institutional Review Board at Tianjin Third Central Hospital (No. IRB2019-015-01). All participants provided written informed consent before undergoing any study-related procedures. All procedures in this study were performed in accordance with the Declaration of Helsinki.
Patient samples and study design
From January 2019 to December 2020, a total of 174 CRC patients from Tianjin Third Central Hospital were reviewed and enrolled in this study. The patient inclusion criteria were CRC patients who were examined by both NGS and MSI-PCR tests.
Among the 174 patients, 56 patients were assigned to the training set and paired tumor tissue and peripheral blood samples of these patients were analyzed with a 616-gene pan-cancer panel (USCI, Beijing, China) sequencing, which covers 2.2 Mb of the human genome, including exons and partial introns of cancer driver genes, as well as hereditary cancer-related genes and therapy guidance-related genes. The other 118 patients with only tumor samples were assigned to the validation set. The samples were subjected to sequencing by a 47-gene CRC-related panel (USCI), which included BRAF, KRAS, NRAS, PTEN, and other genes related to carcinogens and tumor development of CRC.
Genomic DNA was isolated from tissue and peripheral blood samples using QIAamp DNA FFPE Tissue Kit (Qiagen, Dusseldorf, Germany) and TIANamp Blood DNA Maxi Kit (Tiagen, Beijing, China) according to the manufacturer's instructions. Sheared genomic DNA (50–100 ng) was subjected to library construction with an MGIEasy Universal DNA Library Kit (MGI, Beijing, China), and then followed by hybrid capture using an xGen Hybridization and Wash Kit (IDT, IA, USA). The qualified libraries were sequenced with 2 × 100 bp paired-end reads on a MGISEQ-2000 (MGI) platform. The paired-end reads were aligned to the human reference genome GRCh37/hg19 using BWA-MEM (v0.7.17), and single nucleotide variants were determined by VarScan (v2.4.3). TMB was assessed as described by Chalmers et al.
MSI-PCR was performed as previously reported. An MSI Detection Kit (Microread Genetics Co. Ltd., Beijing, China; National Medical Product Administration approval number: 20213400936) with six mononucleotide repeat loci (NR-21, BAT-26, NR-27, BAT-25, NR-24, and MONO-27) was used in accordance with the manufacturer's instructions. The length of PCR fragments was detected on an ABI 3730xl Genetic Analyzer (Applied Biosystems, CA, USA) and analyzed with GeneMapper software version 4.0 (Applied Biosystems). Samples were considered as MSI-H when instability was observed in two or more of the six loci and as microsatellite stability (MSS) when instability in fewer than two loci was observed.
The difference between MSS and MSI-H cases clustered by MSI-PCR was determined by chi-squared (χ2) test or non-parametric Mann–Whitney U test. A two-sided P value <0.05 was considered to be statistically significant.
Development and performance of the CRC-MSI model
In this study, we aimed to detect the MSI status of CRC patients using only tumor tissue with a 47-gene NGS panel test. A total of 56 patients were assigned to training set and their tumor and blood samples were analyzed with a 616-gene pan-cancer panel. A total of 118 patients were assigned to validation group and their tumor samples were sequenced with the 47-gene panel [Figure 1].
To select microsatellite loci, we located the intersecting area of the covered genomic regions in the 616-gene panel and the 47-gene panel. Microsatellite loci identified by RepeatFinder in the shared regions were candidates for further modeling. A total of 42 common microsatellite loci were selected, including 23 mononucleotide repeat sites, 15 dinucleotide repeat sites, three trinucleotide repeat sites, and one pentanucleotide repeat site. The mononucleotide repeat site, BAT-25, within the KIT gene region, was also included. A baseline was built for each candidate locus based on the NGS data of the 56 control blood samples in the training set. We used multinomial distribution to assess changes in repeat length in each site in the tissue samples. If the change was greater than the threshold, the site was considered as unstable. The MSI score, the percentage of unstable sites within all sites that met the requirement of parameters, is used for MSI calling. A fraction of >0.2 (>20% unstable loci) was considered as MSI-H.
We then evaluated the performance of the model with 42 loci in the training and validation sets. The MSI status of CRC samples determined by NGS was 100% consistent with MSI-PCR in both training and validation sets [Figure 2A], and the area under the receiver operating characteristic curve (AUC) reached 1.000 [Figure 2B]. However, the MSI score was relatively continuous in the validation set, which might influence the judgment of samples in the gray area.
To further optimize the model, we investigated the contributions of MSI score of loci with different repeat length motifs. As shown in Figure 2C, the performance of the 23 mononucleotide repeat sites was as good as that of the 42 loci, and the AUC was up to 1.000 [Figure 2D]. The median number of unstable mononucleotide repeat sites was 17 (range: 3–19) in the MSI-H samples, which was significantly higher than that in the MSS samples (median: 0; range: 0–3; P < 0.01). Notably, the gap between the scores of MSI-H and MSS was bigger, which indicated a better performance of the 23 loci in the gray area. In contrast, the best concordance rate of the remaining 19 loci with longer repeat motifs was 60.9% compared with MSI-PCR, and the AUC reached only 0.699 [Figure 2E and 2F]. Thus, mononucleotide repeat sites made a major contribution on MSI status calling, which was consistent with a previous study. Therefore, the MSI calling model with 23 loci was named CRC-MSI and used in further analysis.
As shown in Tables 1 and 2, the CRC-MSI achieved a 100% sensitivity and 100% specificity compared with MSI-PCR. The MSI status of locus BAT-25 was consistent in MSI-PCR and CRC-MSI. We also investigated the concordance of CRC-MSI with IHC. For 11 patients with IHC results (5 dMMR and 6 proficient MMR [pMMR]), only 1 dMMR sample with loss of MSH2 was predicted as MSS by the CRC-MSI [Table 1]. The overall concordance rate was 90.9%.
Table 1 -
Correlation of CRC-MSI, MSI-PCR, IHC, and MMR genes sequencing.
||MSI-PCR (n = 174)
||IHC (n = 11)
||MMR genes sequencing (n = 56)
MMR genes were limited to MLH1, MSH6, MSH2, and PMS2. CRC-MSI: Colorectal cancer microsatellite instability; dMMR: deficient mismatch repair; IHC: Immunohistochemical analysis; MMR: Mismatch repair; MSI: Microsatellite instability; MSI-H: MSI-high; MSS: Microsatellite stability; NPV: Negative predictive value; PCR: Polymerase chain reaction; pMMR: proficient MMR; PPV: Positive predictive value.
Table 2 -
Performance of CRC-MSI when compared with MSI-PCR, IHC, and MMR genes sequencing.
|MMR genes sequencing
CRC-MSI: Colorectal cancer microsatellite instability; IHC: Immunohistochemical analysis; MMR: Mismatch repair; NPV: Negative predictive value; PPV: Positive predictive value.
To evaluate the performance of CRC-MSI on low tumor purity, gradient dilutions of tumor DNA with paired normal DNA were conducted. As shown in Table 3, the 23-loci and 42-loci models detected MSI-H mixtures with a tumor content as low as 5.88% and 9.80%, respectively. Along with the dilution, the number of unstable loci in the 42-loci model and that in the 23-loci model tended to be equal, which showed that the unstable loci with a motif length more than one nucleotide could not be detected. These results explain why the CRC-MSI model outperforms the original 42-loci model.
Table 3 -
Performance of CRC-MSI on low tumor content samples.
||Unstable loci/total loci (n/n)
||Unstable loci/total loci (n/n)
The 23-loci and 42-loci models could detect MSI-H mixtures with tumor content down to 5.88% and 9.80%. Along with the dilution, the unstable loci with the motif length more than one nucleotide could not be detected, which may illustrate that the CRC-MSI model outperforms the original 42-loci model. CRC-MSI: Colorectal cancer microsatellite instability; MSI: Microsatellite instability; MSI-H: MSI-high; MSS: Microsatellite stability; TP: tumor purity.
Correlation between MSI status and patients’ clinical characteristics
The clinical characteristics of patients are summarized in Table 4. The median age of all patients was 50 years, ranging from 32 to 91 years. There were 31 patients with MSI-H and 143 patients with MSS. MSI-H patients were much younger than MSS patients (P < 0.01). Additionally, more percentage of female patients had an MSI-H status (P < 0.01). MSI-H was significantly enriched in samples with tumor stage II. Approximately 0% (0/6), 39.4% (13/33), 6.7% (3/45), and 1/9 of patients with tumor stages I, II, III, and IV, respectively, showed MSI-H. Most MSI-H tumors were located at the right hemicolon. The incidences of right hemicolon cancer and left hemicolon cancer in MSI-H patients were 76.5% (13/17) and 23.5% (4/17), respectively, compared with 26.9% (25/93) and 73.1% (68/93) in MSS patients.
Table 4 -
Characteristics of patients in this study.
||MSI-H (n = 31)
||MSS (n = 143)
| Right hemicolon
| Left hemicolon
MSI-H and MSS were grouped by MSI-PCR. MSI: Microsatellite instability; MSI-H: MSI-high; MSS: Microsatellite stability.
Correlation between MSI status, MMR gene mutations, and other genomic alterations
We investigated alterations in MMR genes covered by the 616-gene panel, including MLH1, MSH6, MSH2, and PMS2, in the 56 training samples [Figure 3]. The results showed that at least one of the MMR genes was altered in 8/10 (80%) MSI-H samples and 1/46 (2.1%) MSS samples. The most frequently altered MMR gene was MLH1 (5/56), followed by MSH6 (4/56) and PMS2 (4/56). Four MSI-H samples harbored alterations in two or more of the MMR genes. All the alterations in these MMR genes were somatic. For the two MSI-H patients without alterations in exons, we speculated that gene alteration might be caused by epigenetic inactivation. Among the 11 patients with IHC results, eight patients (2 dMMR and 6 pMMR) were subjected to the 616-gene panel NGS test. All six pMMR samples showed no alteration in exons of the four MMR genes. For the discordant dMMR patient who was mistakenly predicted as MSS by CRC-MSI, no alteration was found in the four MMR genes, which might also be caused by epigenetic inactivation.
We also profiled other CRC-related hot genes, such as AKT1, ATM, BRAF, EGFR, ERBB2, FGFR1, KIT, KRAS, NRAS, PIK3CA, POLE, and PTEN, in all the 174 samples [Supplementary Figure 1, https://links.lww.com/CM9/B107]. The most frequently mutated gene was KRAS (16/31 in MSI-H patients vs. 65/143 MSS patients). Six genes were more frequently altered much more in MSI-H than those in MSS samples: ATM (6/31 vs. 3/143, P < 0.01), BRAF (9/31 vs. 7/143, P < 0.01), ERBB2 (5/31 vs. 2/143, P < 0.01), PIK3CA (11/31 vs. 17/143, P < 0.01), POLE (5/31 vs. 0/143, P < 0.01), and PTEN (4/31 vs. 1/143, P < 0.01. Although alterations in AKT1, BRCA2, EGFR, FGFR1, KIT, and NRAS were only identified in MSS samples, there was no significant difference between MSS and MSI-H samples. The number of altered genes in each sample was significantly higher for MSI-H samples than for MSS samples (P < 0.01, Supplementary Figure 2A, https://links.lww.com/CM9/B108).
Furthermore, we calculated the TMB scores for samples using the 616-panel NGS test. The median TMB score of the MSI-H samples (66.6 mutations/Mb; range 37.8–101.1 mutations/Mb) was much higher than that of the MSS samples (5.8 mutations/Mb; range 0–16 mutations/Mb; P < 0.01; Supplementary Figure 2B, https://links.lww.com/CM9/B108).
Although MSI-PCR test is the gold standard for MSI assessment in the clinical setting, along with the NGS-panel-based profiling test, NGS-based MSI calling methods independent of normal samples would simplify testing processes, reduce cost, and provide more genomic information in a single test. Here, we developed an NGS-based MSI calling model, CRC-MSI, to determine MSI status along with a CRC NGS test using only tumor samples. With effective selection of candidate loci, the model with 23 mononucleotides repeat sites achieved a 100% concordance rate with MSI-PCR in all samples and was robust to low tumor content.
Several bioinformatical tools evaluating MSI have been developed in recent years, including tools requiring tumor/normal-paired samples, such as MSI sensor, MANTIS, and MSI-ColonCore, and tools requiring only tumor samples, such as mSINGS and MSIsensor-pro. mSINGS exhibited a comparable sensitivity (range: 96.4%–100%) and specificity (range: 97.2%–100%) in three cohorts in a total of 108 cases and a 97.1% sensitivity and 100% specificity in another study of 78 cases. MSIsensor-pro was reported to have the best AUC compared with MANTIS, mSINGS, and MSIsensor in 1532 samples from the TCGA database (AUC: 0.99). We also tested the performance of our 23 mononucleotide repeat sites in mSINGS and MSIsensor-pro using our 174 CRC samples. Both mSINGS and MSIsensor-pro achieved a 100% concordance, which indicated the importance of loci selection (data not shown).
Our study confirmed that mononucleotide repeat sites are more sensitive and specific for detecting MSI status than sites with longer motifs (repeat unit with a length longer than one). Mononucleotide repeat loci contributed most to the MSI score, and loci with longer repeat motifs simply increased the denominator in the fraction used for calculating the MSI score. This may explain the difference of the cutoff scores of the two models in our study.
The mSINGS and MSI-ColonCore methods also only adopted mononucleotide microsatellite loci. Because of the difference in underlying algorithms, mSINGS had a better performance with fewer loci. Although MSI sensor-pro is capable of evaluating >10,000 loci with 1 to 5 nucleotide repeat motifs, approximately half of the loci are mononucleotide repeat sites. Therefore, we propose that mononucleotide repeat sites are the first choice, instead of sites with longer repeat motifs, to determine MSI along with a small NGS panel.
The CRC-MSI was 100% consistent with MSI-PCR and 91% consistent with IHC. Previous studies reported that the concordant rate of IHC for MMR proteins and MSI-PCR ranged from 84.5% to 98.6%.[26–28] With limited IHC data, we also analyzed alterations in MMR genes. In the analysis of the commonly tested MMR genes (MLH1, MSH2, MSH6, and PMS2), 80% of the MSI-H samples and 2.1% of the MSS samples were altered in at least one gene. The overall consistency was 94.6%. In another study, 80.43% of the MSI-H samples and 17.8% of the MSS samples were altered in these four MMR genes, and the overall consistency was 81.8%. However, alterations in these four MMR genes were identified in 67% of 25 dMMR cases in another study. These differences may be because of alterations not studied in this study, such as MLH1 promoter methylation, alterations in MLH3 genes, and loss of heterozygosity of MMR genes or complex structural rearrangements. We also analyzed another two MMR genes covered by the 616-gene-panel, MSH3 and PMS1. MSH3 was altered in half of the tested MSI-H samples, along with other MMR genes or alone, while PMS1 was normal in all tested samples. Germline variations in MSH3 and PMS1 have been shown to be correlated with sporadic CRC or Lynch syndrome.[31,32] Our study showed somatic variation in these two genes. We also confirmed that high MSI was relevant to high TMB in CRC. Both MSI and TMB are emerging as biomarkers for immunotherapy response.
In conclusion, we developed a robust NGS MSI calling model for CRC patients that only requires tumor samples and that shows satisfactory sensitivity and specificity. The performance of mononucleotide repeat sites surpasses loci with a longer repeat motif in MSI calling.
Author thanks all patients for consenting to participate.
Conflicts of interest
1. Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, et al. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci U S A
1998; 95:68706875. doi: 10.1073/pnas.95.12.6870.
2. Ligtenberg MJ, Kuiper RP, Chan TL, Goossens M, Hebeda KM, Voorendt M, et al. Heritable somatic methylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3’ exons of TACSTD1. Nat Genet
2009; 41:112117. doi: 10.1038/ng.283.
3. Hause RJ, Pritchard CC, Shendure J, Salipante SJ. Classification and characterization of microsatellite instability
across 18 cancer types. Nat Med
2016; 22:13421350. doi: 10.1038/nm.4191.
4. Le DT, Durham JN, Smith KN, Wang H, Bartlett BR, Aulakh LK, et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science
2017; 357:409413. doi: 10.1126/science.aan6733.
5. Pino MS, Chung DC. Microsatellite instability
in the management of colorectal cancer
. Expert Rev Gastroenterol Hepatol
2011; 5:385399. doi: 10.1586/egh.11.25.
6. Bonneville R, Krook MA, Kautto EA, Miya J, Wing MR, Chen HZ, et al. Landscape of microsatellite instability
across 39 cancer types. JCO Precis Oncol
2017; 2017: O.17.00073. doi: 10.1200/po.17.00073.
7. Pino MS, Mino-Kenudson M, Wildemore BM, Ganguly A, Batten J, Sperduti I, et al. Deficient DNA mismatch repair
is common in Lynch syndrome-associated colorectal adenomas. J Mol Diagn
2009; 11:238247. doi: 10.2353/jmoldx.2009.080142.
8. Le DT, Uram JN, Wang H, Bartlett BR, Kemberling H, Eyring AD, et al. PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med
2015; 372:25092520. doi: 10.1056/NEJMoa1500596.
9. Committee of Colorectal Cancer
, Chinese Society of Clinical Oncology; Genetics Group of The Committee of Colorectal Cancer
, China Anti-Cancer Association; Genetics, Committee of The Committee of Colorectal Cancer
, Chinese Medical Doctor Association. Consensus on the detection of microsatellite instability
in colorectal cancer
and other related solid tumors in China (in Chinese). Chin J Oncol
2019; 41:734741. doi: 10.3760/cma.j.issn.0253-3766.2019.10.003.
10. Luchini C, Bibeau F, Ligtenberg MJL, Singh N, Nottegar A, Bosse T, et al. ESMO recommendations on microsatellite instability
testing for immunotherapy in cancer, and its relationship with PD-1/PD-L1 expression and tumour mutational burden: a systematic review-based approach. Ann Oncol
2019; 30:12321243. doi: 10.1093/annonc/mdz116.
11. Nivolumab plus ipilimumab achieves responses in dMMR/MSI-H tumors. Cancer Discov
2018; 8:263doi: 10.1158/2159-8290.CD-RW2018-017.
12. de Rosa N, Rodriguez-Bigas MA, Chang GJ, Veerapong J, Borras E, Krishnan S, et al. DNA mismatch repair
deficiency in rectal cancer: benchmarking its impact on prognosis, neoadjuvant response prediction, and clinical cancer genetics. J Clin Oncol
2016; 34:30393046. doi: 10.1200/jco.2016.66.6826.
13. Niu B, Ye K, Zhang Q, Lu C, Xie M, McLellan MD, et al. MSIsensor: microsatellite instability
detection using paired tumor-normal sequence data. Bioinformatics
2013; 30:10151016. doi: 10.1093/bioinformatics/btt755.
14. Kautto EA, Bonneville R, Miya J, Yu L, Krook MA, Reeser JW, et al. Performance evaluation for rapid detection of pan-cancer microsatellite instability
with MANTIS. Oncotarget
2017; 8:74527463. doi: 10.18632/oncotarget.13918.
15. Zhu L, Huang Y, Fang X, Liu C, Deng W, Zhong C, et al. A novel and reliable method to detect microsatellite instability
in colorectal cancer
by next-generation sequencing
. J Mol Diagn
2018; 20:225231. doi: 10.1016/j.jmoldx.2017.11.007.
16. Salipante SJ, Scroggins SM, Hampel HL, Turner EH, Pritchard CC. Microsatellite instability
detection by next generation sequencing. Clin Chem
2014; 60:11921199. doi: 10.1373/clinchem.2014.223677.
17. Jia P, Yang X, Guo L, Liu B, Lin J, Liang H, et al. MSIsensor-pro: fast, accurate, and matched-normal-sample-free detection of microsatellite Instability
. Genom Proteom Bioinform
2020; 18:6571. doi: 10.1016/j.gpb.2020.02.001.
18. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. ArXiv
19. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, et al. VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res
2012; 22:568576. doi: 10.1101/gr.129684.111.
20. Chalmers ZR, Connelly CF, Fabrizio D, Gay L, Ali SM, Ennis R, et al. Analysis of 100,000 human cancer genomes reveals the landscape of tumor mutational burden. Genome Med
2017; 9:34doi: 10.1186/s13073-017-0424-2.
21. Zheng K, Wan H, Zhang J, Shan G, Chai N, Li D, et al. A novel NGS-based microsatellite instability
(MSI) status classifier with 9 loci for colorectal cancer
patients. J Transl Med
2020; 18:215doi: 10.1186/s12967-020-02373-1.
22. Volfovsky N, Haas BJ, Salzberg SL. A clustering method for repeat analysis in DNA sequences. Genome Biol
2001; 2: RESEARCH0027. doi: 10.1186/gb-2001-2-8-research0027.
23. Buhard O, Suraweera N, Lectard A, Duval A, Hamelin R. Quasimonomorphic mononucleotide repeats for high-level microsatellite instability
analysis. Dis Markers
2004; 20:251257. doi: 10.1155/2004/159347.
24. Hempelmann JA, Scroggins SM, Pritchard CC, Salipante SJ. MSIplus for integrated colorectal cancer
molecular testing by next-generation sequencing
. J Mol Diagn
2015; 17:705714. doi: 10.1016/j.jmoldx.2015.05.008.
25. Suraweera N, Duval A, Reperant M, Vaury C, Furlan D, Leroy K, et al. Evaluation of tumor microsatellite instability
using five quasimonomorphic mononucleotide repeats and pentaplex PCR. Gastroenterology
2002; 123:18041811. doi: 10.1053/gast.2002.37070.
26. Vanderwalde A, Spetzler D, Xiao N, Gatalica Z, Marshall J. Microsatellite instability
status determined by next-generation sequencing
and compared with PD-L1 and tumor mutational burden in 11,348 patients. Cancer Med
2018; 7:746756. doi: 10.1002/cam4.1372.
27. Yan WY, Hu J, Xie L, Cheng L, Yang M, Li L, et al. Prediction of biological behavior and prognosis of colorectal cancer
patients by tumor MSI/MMR in the Chinese population. Onco Targets Ther
2016; 9:74157424. doi: 10.2147/ott.S117089.
28. Yuan L, Chi Y, Chen W, Chen X, Wei P, Sheng W, et al. Immunohistochemistry and microsatellite instability
analysis in molecular subtyping of colorectal carcinoma based on mismatch repair competency. Int J Clin Exp Med
29. Kim JE, Chun SM, Hong YS, Kim KP, Kim SY, Kim J, et al. Mutation burden and I index for detection of microsatellite instability
in colorectal cancer
by targeted next-generation sequencing
. J Mol Diagn
2019; 21:241250. doi: 10.1016/j.jmoldx.2018.09.005.
30. Pang J, Gindin T, Mansukhani M, Fernandes H, Hsiao S. Microsatellite instability
detection using a large next-generation sequencing
cancer panel across diverse tumour types. J Clin Pathol
2020; 73:8389. doi: 10.1136/jclinpath-2019-206136.
31. Berndt SI, Platz EA, Fallin MD, Thuita LW, Hoffman SC, Helzlsouer KJ. Mismatch repair polymorphisms and the risk of colorectal cancer
. Int J Cancer
2007; 120:15481554. doi: 10.1002/ijc.22510.
32. Peltomäki P. Lynch syndrome genes. Fam Cancer
2005; 4:227232. doi: 10.1007/s10689-004-7993-0.