Journal of Translational Medicine (Jan 2018)

Accuracy of four mononucleotide-repeat markers for the identification of DNA mismatch-repair deficiency in solid tumors

  • Yuko Takehara,
  • Takeshi Nagasaka,
  • Akihiro Nyuya,
  • Tomoko Haruma,
  • Junko Haraga,
  • Yoshiko Mori,
  • Keiichiro Nakamura,
  • Toshiyoshi Fujiwara,
  • C. Richard Boland,
  • Ajay Goel

DOI
https://doi.org/10.1186/s12967-017-1376-4
Journal volume & issue
Vol. 16, no. 1
pp. 1 – 11

Abstract

Read online

Abstract Background To screen tumors with microsatellite instability (MSI) arising due to DNA mismatch repair deficiency (dMMR), a panel of five quasi-monomorphic mononucleotide-repeat markers amplified in a multiplex PCR (Pentaplex) are commonly used. In spite of its several strengths, the pentaplex assay is not robust at detecting the loss of MSH6-deficiency (dMSH6). In order to overcome this challenge, we designed this study to develop and optimize a panel of four quasi-monomorphic mononucleotide-repeat markers (Tetraplex) for identifying solid tumors with dMMR, especially dMSH6. Methods To improve the sensitivity for tumors with dMMR, we established a quasi-monomorphic variant range (QMVR) of 3–4 bp for the four Tetraplex markers. Thereafter, to confirm the accuracy of this assay, we examined 317 colorectal cancer (CRC) specimens, comprising of 105 dMMR [45 MutL homolog (MLH)1-deficient, 45 MutS protein homolog (MSH)2-deficient, and 15 MSH6-deficient tumors] and 212 MMR-proficient (pMMR) tumors as a test set. In addition, we analyzed a cohort of 138 endometrial cancers (EC) by immunohistochemistry to determine MMR protein expression and validation of our new MSI assay. Results Using the criteria of ≥ 1 unstable markers as MSI-positive tumor, our assay resulted in a sensitivity of 97.1% [95% confidence interval (CI) = 91.9–99.0%] for dMMR, and a specificity of 95.3% (95% CI = 91.5–97.4%) for pMMR CRC specimens. Among the 138 EC specimens, 41 were dMMR according to immunohistochemistry. Herein, our Tetraplex assay detected dMMR tumors with a sensitivity of 92.7% (95% CI = 80.6–97.5%) and a specificity of 97.9% (95% CI = 92.8–99.4%) for pMMR tumors. With respect to tumors with dMSH6, in the CRC-validation set, Tetraplex detected dMSH6 tumors with a sensitivity of 86.7% (13 of 15 dMSH6 CRCs), which was subsequently validated in the EC test set as well (sensitivity, 75.0%; 6 of 8 dMSH6 ECs). Conclusions Our newly optimized Tetraplex system will help offer a robust and highly sensitive assay for the identification of dMMR in solid tumors.

Keywords