Abstract
Background: Telomeres have long been found to be involved in cancer development, while little was known about the dynamic changes of telomere length in carcinogenesis process. Methods: The present study longitudinally investigated telomere alterations of cell-free DNA (cfDNA) in 86 gastric cancer (GC) subjects recruited through a 16-year prospective cohort with 2-4 serums collected before each GC-diagnosis from baseline and three follow-up time-points (a total of 276 samples). As the control, 86 individual-matched cancer-free subjects were enrolled with 276 serums from the matched calendar year. Results: In the 73 pairs of baseline serums from GC and control subjects, shortened telomeres showed increased subsequent GC risk [odds ratio (OR) = 9.17, 95% CI: 2.72-31.25 for 1 unit shortening]. In each baseline gastric lesion category, higher risks of GC progression were also found with shortened cfDNA telomeres; ORs per 1 unit shortening were 6.99 (95% CI: 1.63-30.30) for mild gastric lesions, 6.06 (95% CI: 1.89-19.61) for intestinal metaplasia and 15.63 (95% CI: 1.91-125.00) for dysplasia. With all measurements from baseline and follow-up time-points, shortened telomeres also showed significant association with GC risk (OR = 7.37, 95% CI: 2.06-26.32 for 1 unit shortening). In temporal trend analysis, shortened telomeres were found in GC subjects compared to corresponding controls more than 3 years ahead of GC-diagnosis (most P < 0.05), while no significant difference was found between two groups within 3 years approaching to GC-diagnosis. Conclusion: Our findings suggest that telomere shortening may be associated with gastric carcinogenesis, which supports further etiological study and potential biomarker for risk stratification.