Dataset Information

Identifying Optimal Loci for the Molecular Diagnosis of Microsatellite Instability.

ABSTRACT:

Background

Microsatellite instability (MSI) predicts oncological response to checkpoint blockade immunotherapies. Although microsatellite mutation is pathognomonic for the condition, loci have unequal diagnostic value for predicting MSI within and across cancer types.

Methods

To better inform molecular diagnosis of MSI, we examined 9438 tumor-normal exome pairs and 901 whole genome sequence pairs from 32 different cancer types and cataloged genome-wide microsatellite instability events. Using a statistical framework, we identified microsatellite mutations that were predictive of MSI within and across cancer types. The diagnostic accuracy of different subsets of maximally informative markers was estimated computationally using a dedicated validation set.

Results

Twenty-five cancer types exhibited hypermutated states consistent with MSI. Recurrently mutated microsatellites associated with MSI were identifiable in 15 cancer types, but were largely specific to individual cancer types. Cancer-specific microsatellite panels of 1 to 7 loci were needed to attain ≥95% diagnostic sensitivity and specificity for 11 cancer types, and in 8 of the cancer types, 100% sensitivity and specificity were achieved. Breast cancer required 800 loci to achieve comparable performance. We were unable to identify recurrent microsatellite mutations supporting reliable MSI diagnosis in ovarian tumors. Features associated with informative microsatellites were cataloged.

Conclusions

Most microsatellites informative for MSI are specific to particular cancer types, requiring the use of tissue-specific loci for optimal diagnosis. Limited numbers of markers are needed to provide accurate MSI diagnosis in most tumor types, but it is challenging to diagnose breast and ovarian cancers using predefined microsatellite locus panels.

SUBMITTER: Long DR

PROVIDER: S-EPMC7528407 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundMismatch repair (MMR)/microsatellite instability (MSI) and tumor mutational burden (TMB) are independent biomarkers that complement each other for predicting immune checkpoint inhibitors (ICIs) efficacy. Here we aim to establish a strategy that integrates MSI and TMB determination for colorectal cancer (CRC) in one single assay.MethodsSurgical or biopsy specimens retrospectively collected from CRC patients were subjected to NGS analysis. Immunohistochemistry (IHC) and polymerase chain reaction (PCR) were also used to determine MMR/MSI for those having enough tissues. The NGS-MSI method was validated against IHC and PCR. The MSI-high (MSI-H) or microsatellite stable (MSS) groups were further stratified based on tumor mutational burden, followed by validation using the The Cancer Genome Atlas (TCGA) CRC dataset. Immune microenvironment was evaluated for each subgroup be profiling the expression of immune signatures.ResultsTissues from 430 CRC patients were analyzed using a 381-gene NGS panel. Alterations in KRAS, NRAS, BRAF, and HER2 occurred at a significantly higher incidence among MSI-H tumors than in MSS patients (83.6% vs. 58.4%, p = 0.0003). A subset comprising 98 tumors were tested for MSI/MMR using all three techniques, where NGS proved to be 99.0 and 93.9% concordant with PCR and IHC, respectively. Four of the 7 IHC-PCR discordant cases had low TMB (1.1-8.1 muts/Mb) and were confirmed to have been misdiagnosed by IHC. Intriguingly, 4 of the 66 MSS tumors (as determined by NGS) were defined as TMB-high (TMB-H) using a cut-off of 29 mut/Mb. Likewise, 15 of the 456 MSS tumors in the TCGA CRC cohort were also TMB-H with a cut-off of 9 muts/Mb. Expression of immune signatures across subgroups (MSS-TMB-H, MSI-H-TMB-H, and MSS-TMB-L) confirmed that the microenvironment of the MSS-TMB-H tumors was similar to that of the MSI-H-TMB-H tumors, but significantly more immune-responsive than that of the MSS-TMB-L tumors, indicating that MSI combined with TMB may be more precise than MSI alone for immune microenvironment prediction.ConclusionThis study demonstrated that NGS panel-based method is both robust and tissue-efficient for comprehensive molecular diagnosis of CRC. It also underscores the importance of combining MSI and TMB information for discerning patients with different microenvironment.