Molecular screening of urine for Mismatch Repair deficient urothelial tumours; an under-appreciated cancer in Lynch syndrome


Session type:

Rachel Phelps1, John Burn1, Christine Hayes1, Richard Gallon1, Jo Pethick2, Fiona McRonald2, Steven Hardy2, Peter Gawthorpe1, Gillian Borthwick1, Mauro Santibanez-Koref1, Michael Jackson1
1Newcastle University, 2Public Health England



Urothelial cancers (UCs), involving the bladder, ureter or kidney, are the third most common Lynch Syndrome (LS) malignancy, but are under-recognised and not subject to surveillance.  The >25% lifetime-cumulative UC incidence in MSH2 carriers calls for surveillance, but current methods (urine cytology) are not easily scalable and are insensitive.  Increased microsatellite instability (MSI) has been reported in LS-associated UC, prompting us to test our highly sensitive MSI assay as a non-invasive, urine-based UC screening tool.


NHS Digital National Disease Registration Service data was used to investigate LS detection in UCs in England. To develop a urine screening tool, we used twelve 7-12bp mononucleotide repeats from our published, sequencing-based MSI assay in a multiplex PCR to facilitate cfDNA analysis, and validated it using colorectal cancer samples. We assayed peripheral blood leukocyte DNA, tumour DNA, and cfDNA from pre- and post-operative urine samples from an LS patient (germline MSH2 variant carrier) with a right kidney UC.  Amplicons were sequenced using a MiSeq (Illumina), and variants in microsatellite length were assessed.


Our diagnostic laboratory review revealed 10,243 UCs in England in 2018.  The estimated frequency of LS in UC is ~2%.  At least 5% should be tested to identify LS cases.  However, only 43 patients (0.4%) had undergone germline testing. A clear MSI signal involving 6 out of 12 markers was observed in cfDNA from pre-operative urine and tumour DNAs of the LS patient with an Upper Tract UC.  Variant allele frequencies (VAFs) of the markers showing instability ranging from ~25% to 78%, with the same markers showing even higher frequencies within DNA extracted from the resected tumour.  This MSI signal disappeared in post-op samples and was absent from blood. We estimated that 70% of the urinary cfDNA in pre-op samples was tumour derived. 


Our assay accurately detected MSI in patient pre-operative urine, offering the potential to become a low-cost routine surveillance method for UC in LS patients.

Impact statement

Urothelial Cancers are not widely recognised as an important part of Lynch Syndrome, the commonest inherited cancer disorder; we have developed an effective non-invasive screening technique based on the demonstration of MSI in urinary DNA.