Genomic Variation Predicts Esophageal Cancer Many Years Before Diagnosis

Esophageal cancer, illustration

Research suggests that the likelihood of progressing to esophageal cancer can be predicted using genetic analysis up to 10 years before an official cancer diagnosis is made.

Scientists based at the European Molecular Biology Laboratory and European Bioinformatics Institute (EMBL-EBI) in the UK, as well as colleagues at the University of Cambridge, found that genomic changes picked up by shallow sequencing of cell biopsy samples in people with some abnormal cells could reliably predict who would go on to develop cancer in the future.

Some people experience an abnormal change to cells in the lining of the esophagus — a condition called Barrett’s esophagus – which is not cancer, but can moderately increase a person’s risk (0.3% per year) of developing esophageal cancer.

Current methods of monitoring people with this condition are unpleasant, such as regular endoscopy, and often have to continue for years. If genetic analysis could be reliably used to stratify individual’s with Barrett’s esophagus into high and low risk then it could substantially improve quality of life for low-risk individuals and allow those at high risk to access early diagnosis and cancer treatment.

Earlier research by the same researchers showed that mutations, such as those that result in chromosome number changes, and driver mutations that help cancer to grow faster, can occur in cells around the body years before cancer actually begins to manifest.

In a continuation of this earlier work, the team, led by Moritz Gerstung, a researcher from the EMBL-EBI in Hinxton, and Rebecca Fitzgerald, a researcher at the Medical Research Council Cancer Unit at the University of Cambridge, tested whether these early mutations could accurately predict progression from Barrett’s esophagus to esophageal cancer.

As reported in the journal Nature Medicine, the researchers tested 777 cell biopsy samples taken from 88 people with Barrett’s esophagus over a period of 15 years. Using a statistical model, they found that genomic changes (largely copy number variation) could accurately predict progression up to 10 years in the future. They also checked and validated their methods in two independent groups of 76 and 248 patients.

The team found that genomic instability increased as time to diagnosis got closer, but even at 8 years or more before diagnosis 50% of biopsy samples taken from those who progressed showed genetic markers predicting this progression. This went up to 70% or more of the samples at 2 years prior to diagnosis. Importantly, the model also highlighted people at lower risk of progression.

“Our data and analysis showed that the propensity to become cancerous is often apparent in the genome many years before transformation,” commented Gerstung. “This helps make more rapid treatment decisions and reduce the need for monitoring in low risk cases. Overall, we’d estimate that 50% of patients could receive fewer endoscopies without compromising safety.”

These kinds of genomic changes can be seen across many cancers, but pinning down the cells with these changes can be tricky.

“It is easiest to test these hypotheses on cancers with known precursors or which are easily accessible,” says Gerstung.

“For example, we made similar observations in acute myeloid leukemia, where we detected genetic precursor cells in blood samples taken around 10 years prior to diagnosis.”

At present the model has only been tested on a relatively small sample set of patients with Barrett’s esophagus. It has a specificity, or ability to pick up negative cancer cases, of 83% and a sensitivity, or ability to pick up positive cancer cases, of 82%. While pretty good, this can still be improved by testing more samples.

“The next steps are to perform further testing in another independent set of patient samples and to see if we can improve the predictions further,” says Fitzgerald. “We would then plan to do a randomized clinical trial as final proof before it could be brought into clinical practice.”

This site uses Akismet to reduce spam. Learn how your comment data is processed.