BackgroundGastro-oesphageal is one of the most common cancers worldwide. Evidence suggested that increased awareness of symptoms and earlier diagnosis could help improve treatment options and improve survival.
AimTo derive and validate an algorithm to estimate the absolute risk of having gastro-oesophageal cancer in patients in primary care with and without symptoms.
Design and settingCohort study of 375 UK QResearch ® general practices for development, and 189 for validation.
MethodIncluded patients were aged 30-84 years, free at baseline of a diagnosis of gastro-oesophageal cancer, and without dysphagia, haematemesis, abdominal pain, appetite loss, or weight loss recorded in previous 12 months. The primary outcome was incident diagnosis of gastrooesophageal cancer recorded in the next 2 years. Risk factors examined were age, body mass index, alcohol status, smoking status, deprivation, family history of gastrointestinal cancer, dysphagia, previous diagnosis of cancer apart from gastrooesophageal cancer, haematemesis, abdominal pain, appetite loss, weight loss, tiredness, and anaemia. Cox proportional hazards models were used to develop risk equations. Measures of calibration and discrimination assessed performance in the validation cohort.
ResultsThere were 2527 incident cases of gastrooesophageal cancer from 4.1 million person-years in the derivation cohort. Independent predictors were age, smoking, dysphagia, haematemesis, abdominal pain, appetite loss, weight loss, and anaemia. On validation, the algorithms explained 71% of the variation in females and 73% in males. The receiver operating curve statistics were 0.89 (females) and 0.92 (males). The D statistic was 3.2 (females) and 3.3 (males). The 10% of patients with the highest predicted risks included 77% of all gastro-oesophageal cancers diagnosed over the next 2 years.
ConclusionThe algorithm has good performance and could potentially be used to help identify those at highest risk of gastro-oesophageal cancer, to facilitate early referral and investigation.