Richard Gallon, Rachel Phelps, Christine Hayes, Laurence Brugieres, Léa Guerrini-Rousseau, Chrystelle Colas, Martine Muleris, Neil A.J. Ryan, D. Gareth Evans, Hannah Grice, Emily Jessop, Annabel Kunzemann-Martinez, Lilla Marshall, Esther Schamschula, Klaus Oberhuber, Amedeo A. Azizi, Hagit Baris Feldman, Andreas Beilken, Nina Brauer, Triantafyllia Brozou, Karin Dahan, Ugur Demirsoy, Benoît Florkin, William Foulkes, Danuta Januszkiewicz-Lewandowska, Kristi J. Jones, Christian P. Kratz, Stephan Lobitz, Julia Meade, Michaela Nathrath, Hans-Jürgen Pander, Claudia Perne, Iman Ragab, Tim Ripperger, Thorsten Rosenbaum, Daniel Rueda, Tomasz Sarosiek, Astrid Sehested, Isabel Spier, Manon Suerink, Stefanie-Yvonne Zimmermann, Johannes Zschocke, Gillian M. Borthwick, Katharina Wimmer, John Burn, Michael S. Jackson, and Mauro Santibanez-Koref
BACKGROUND & AIMS: Constitutional mismatch repair deficiency (CMMRD) is a rare recessive childhood cancer predisposition syndrome caused by germline mismatch repair variants. Constitutional microsatellite instability (cMSI) is a CMMRD diagnostic hallmark and may associate with cancer risk. We quantified cMSI in a large CMMRD patient cohort to explore genotype-phenotype correlations using novel MSI markers selected for instability in blood.METHODS: Three CMMRD, 1 Lynch syndrome, and 2 control blood samples were genome sequenced to >120× depth. A pilot cohort of 8 CMMRD and 38 control blood samples and a blinded cohort of 56 CMMRD, 8 suspected CMMRD, 40 Lynch syndrome, and 43 control blood samples were amplicon sequenced to 5000× depth. Sample cMSI score was calculated using a published method comparing microsatellite reference allele frequencies with 80 controls.RESULTS: Thirty-two mononucleotide repeats were selected from blood genome and pilot amplicon sequencing data. cMSI scoring using these MSI markers achieved 100% sensitivity (95% CI, 93.6%-100.0%) and specificity (95% CI 97.9%-100.0%), was reproducible, and was superior to an established tumor MSI marker panel. Lower cMSI scores were found in patients with CMMRD with MSH6 deficiency and patients with at least 1 mismatch repair missense variant, and patients with biallelic truncating/copy number variants had higher scores. cMSI score did not correlate with age at first tumor.CONCLUSIONS: We present an inexpensive and scalable cMSI assay that enhances CMMRD detection relative to existing methods. cMSI score is associated with mismatch repair genotype but not phenotype, suggesting it is not a useful predictor of cancer risk.