Detection of Mismatch Repair Deficiency in Endometrial Cancer: Assessment of IHC, Fragment Length Analysis, and Amplicon Sequencing Based MSI Testing.

Simple Summary: Tumours of the colon and endometrium (lining of the womb) often have defects in genes that repair our DNA. Identifying these tumours is important for patient treatment and identifying Lynch syndrome (LS), an inherited cancer predisposition. However, the two methods most commonly used to identify these defects often give inconsistent results for endometrial tumours. Here, we investigate these inconsistencies by re-analysing 361 tumour samples from clinical trials with a third assay and establish that they are largely due to false positives in one assay (immunohistochemistry) and false negatives in the other (Promega MSI assay). In addition, the DNA repair defects seem to have a reduced impact within endometrial tumours. However, we find that tumours with defects in a gene called MSH6, especially when associated with LS, are only detected efficiently with immunohistochemistry. This supports current guidelines specifically recommending the use of this method for endometrial tumours. Background/Objectives: Mismatch repair (MMR) deficiency can be indicative of Lynch syndrome (LS) and guide treatment with immune checkpoint inhibitors. Colorectal cancers (CRCs) and endometrial cancers (ECs) are routinely screened to identify LS, primarily using immunohistochemistry (IHC) or microsatellite instability (MSI) testing, but concordance between these methods is variable in ECs. Here, we investigate this variability in 361 ECs from the Ohio OCCPI/OPTEC (n = 196) and Manchester PETALS (n = 165) trials, where concordance between assays differed significantly. Methods: Samples were re-tested using the amplicon-sequencing-based Newcastle MSI assay (NCL_MSI), and analysed with respect to existing IHC, MSI and MLH1 promoter hypermethylation data. Results: NCL_MSI showed consistency with the Ohio results (94% and 97% concordance with IHC and original MSI assays, respectively) and increased concordance within the Manchester cohort from 78% to 86% (MSI) and 84% (IHC). Among discordant Manchester samples, NCL_MSI was significantly associated with MLH1 promoter methylation status (p = 0.0028) and had the highest concordance with methylation, (62/69 samples, 90%), indicating utility as a screening tool in this tumour type. However, tumours with germline MSH6 defects were only detected efficiently with IHC; seven out of eight LS tumours classified as MSS by either MSI assay had isolated MSH6 loss, compared to four out of twelve classified as MSI-H by both (p = 0.028). Furthermore, reduced MSI signal was observed in tumours with isolated MSH6 loss (p = 0.009 Ohio, p = 6.2 × 10−5 Manchester) and in both ECs and CRCs with germline defects, although this only reached significance in CRCs (p = 0.002). Conclusions: These results provide further evidence that ECs with MSH6 loss in particular and LS tumours in general have an attenuated MSI signal, providing support for current guidelines specifically recommending IHC for LS detection and immune checkpoint therapy assessment in EC. [ABSTRACT FROM AUTHOR]