Survival outcomes for post colonoscopy colorectal cancers and post endoscopy upper gastrointestinal cancers in England

Aims Post colonoscopy colorectal cancers (PCCRC) are diagnosed within 5 years of a colonoscopy without a colorectal cancer (CRC) diagnosis and post endoscopy upper gastrointestinal cancers (PEUGIC) with 3 years of an endoscopy without upper gastrointestinal tract cancer (UGIC) diagnosis. The impact on survival of PCCRC is uncertain and unknown following PEUGIC. This study assessed the survival of PCCRC and PEUGIC in national cohorts and provides recommendations for reporting such outcomes in future.

Methods A population-based retrospective cohort analysis was conducted. The study population included all patients with CRC or UGIC diagnosed between January 2009 and December 2018 by the English National Health Service. CRC diagnoses were categorised into detected CRC and PCCRC, which were detected at 12-47, 48-87, and 88-120 months following the index colonoscopy that did not diagnose cancer. UGIC diagnoses was categorised into detected UGIC and PEUGIC, which were detected at 6–11, 12-23, and 24-36 months following the index endoscopy. Patients with a pre-existing diagnosis of Barrett’s oesophagus were also analysed and the UGIC cohort excluding this patient group. Kaplan-Meier all-cause survival curves were used to investigate survival, with conditional survival reported. [as1]Needs changing to mirror PEUGIC cohort

Results 150 507 CRC (9 465 PCCRC) and 98 710 UGI cancers (9 068 PEUGIC) were studied, including 9 975 UGIC (2 583 PEUGIC) where patients had a previous Barrett’s oesophagus diagnosis. For CRC, survival curves showed improved survival for detected CRC and worse survival for PCCRC as time from index colonoscopy increased. Less patients presented with stage 4 disease among detected CRC, compared with PCCRC (12.0% versus. 15.0% p<0.05). For PEUGIC survival curves showed improved survival for PEUGIC, which worsened as time from index endoscopy increased and the lowest survival in detected UGIC. In patients with pre-existing Barrett’s oesophagus there was much better survival for PEUGIC compared to detected UGIC with more stage I disease (41.4% versus 20.0% p<0.05) Excluding patients with Barrett’s, survival in PEUGIC and detected UGIC was very similar.

Conclusions Compared with detected CRC, PCCRC results in worse outcomes. The apparent better survival in the PEUGIC cohort is related to the effect of surveillance in Barrett’s oesophagus. A number of epidemiological biases, including selection, lead time, immortal time, and healthy user bias, make reporting overall survival for these cancers challenging and recommendations for future reporting are made.

Publication History

Article published online:
27 March 2025