Clinical utility of serum CEA, CA72-4 and CA19-9 in the diagnosis and staging of gastric carcinoma

Objective To derive and evaluate diagnostic determinants of gastroesophageal malignancy as a means to help family physicians select patients for 'open-access' gastroscopy.Patients In a multicenter study, 861 consecutive patients were investigated with first-time gastroscopy (study population), followed by 1153 patients studied during the next six years (validation population).Main Outcome Measures Univariate and multivariate analyses resulted in four relevant symptoms that were then compared with alarm symptoms from the literature.ROC analysis was used to evaluate the chance of finding a malignancy using these symptoms.Results Four symptoms, i.e., weightloss, obstruction, pain-free nights, and the absence of heartburn, predicted malignancy with an AUC of 0.90 in the study population.'Alarm symptoms' performed less well in the study population (AUC 0.85), although the reproducibility was better in the validation population (0.71 vs 0.63).If exclusion of malignancy had been chosen as the only valid reason for requesting gastroscopy, then preselection with the help of these symptoms reduced the amount of requests by 10%.Conclu- sion Thorough evaluation of alarm symptoms in dyspeptic patients might help minimize unnecessary gastroscopy requests by GPs aiming at a safe treatment decision.