Evaluation of whole-body MRI for cancer early detection in Li-Fraumeni Syndrome

Peter Sodde*, Zerin Hyder, Sarah Pugh, Fiona Lalloo, Richard Martin, Calvin Soh, Jawad Naqvi, Richard Whitehouse, Dafydd Evans, Emma Woodward

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Li-Fraumeni syndrome (LFS) is a high-risk hereditary cancer predisposition syndrome affecting 1 in 5000 individuals. Current standard of care in adults includes annual whole-body MRI (WB-MRI) and MRI brain (MRB) surveillance to enable early cancer detection. We performed a retrospective single-centre study of adults with TP53 pathogenic germline variants (PGVs) or proven somatic mosaicism undergoing annual WBMRI surveillance between January 2012 and January 2024, and MRI brain surveillance between August 2017 and January 2024. Three hundred and twenty-five WB-MRI scans were performed in 75 individuals.
Seventeen cancers were diagnosed in 16 individuals. Nine out of seventeen cancers were WB-MRI detected (7/9 had stage 1/2 disease). Benign incidental findings were identified in 89/325 (27.4%) of WB-MRI scans prompting 53 additional investigations. As a stand-alone surveillance tool WB-MRI demonstrated a pancohort specificity of 95.5%, negative predictive value of 97.4% and sensitivity of 42.9%. Thirty-two individuals underwent 53 MRB scans detecting one cancer. We report the findings from the longest and largest singlecentre experience of WB-MRI surveillance for cancer early detection in adults with LFS, demonstrating a high and acceptable level of cancer exclusion but modest sensitivity with WB-MRI prompting a significant number of additional investigations.
Original languageEnglish
JournalJournal of Medical Genetics
Early online date30 Jun 2025
DOIs
Publication statusPublished - 30 Jun 2025

Keywords

  • Early Diagnosis
  • heredity

Fingerprint

Dive into the research topics of 'Evaluation of whole-body MRI for cancer early detection in Li-Fraumeni Syndrome'. Together they form a unique fingerprint.

Cite this