MMR testing is performed to screen Lynch Syndrome, evaluate the prognosis of colorectal cancer (CRC) and predict the efficacy of PD1/PDL1 blockade in all tumor types. Two methods are available: immunohistochemistry (IHC) using antibodies against MMR proteins and molecular biology (MB) for assessing microsatellite instability (MSI). Classically, dMMR tumor corresponds to loss of expression of two proteins (MLH1 and PMS2 or MSH2 and MSH6) associated with MSI. Atypical profiles of dMMR tumors have sporadically been described. The aim of our study was to describe the frequency and characteristics of these atypical cases.
All MMR testing performed in our center between 2007 and 2017 were checked to select cases with both available IHC and MB. Then, all dMMR cases were reviewed to identify atypical cases which were defined by: isolated loss of expression of one protein, loss of expression of two proteins without MSI, normal expression of the four proteins with MSI, aberrant loss of proteins, or MSI-low. Biological data of atypical cases were controlled and clinical data were collected for each case.
4948 MMR tests were performed, 3800 had both available IHC and MB data, and 585 were dMMR (15 %). Among them, 97 cases were atypical and after biological control, 8 cases were re-classified typical; allowing to finally identify 89 atypical cases: 60 CRC, 10 endometrial carcinoma, 8 digestive non CRC and 11 others types of cancers. A strong correlation with genetic syndromes was observed for those atypical profiles. 2015P (* MSS: microsatellite stability)Isolated PMS2 or MSH6 loss n = 53 Expression of the four proteins n = 5 MSH2/MSH6 or MLH1/PMS2 loss n = 16 Aberrant loss of proteins n = 15 MSI 43 3 - 13 MSI low 1 2 8 - MSS * 9 - 8 2 Clinical characteristics Predominantly CCR Genetic predisposition syndrome (73%) Exclusively CCR or endometrial Genetic predisposition syndrome (≥40%) Predominantly Non CRC (63%) None
Even using controlled IHC and MB, 15% of dMMR tumors have an atypical profile. These atypical cases mainly involve non CRC cancer with a strong prediction for Lynch syndrome. Their therapeutic impact particularly for immunotherapy should be now evaluated.
The authors.
Has not received any funding.
All authors have declared no conflicts of interest.