Background : Although a ‘go to’ marker for the diagnosis and recurrence detection in pancreatic adenocarcinoma, the utility carbohydrate antigen (CA) 19-9 to detect recurrences in resected ampullary carcinomas (ACs) is unknown.

Methods : Histopathologically proven ACs post pancreaticoduodenectomy from January 2012 to January 2020 were evaluated. Survival, recurrence patterns, factors associated with recurrence and the utility of follow-up serum CA 19-9 levels in predicting recurrences was assessed.

Results : Out of the 572 patients who underwent a pancreaticoduodenectomy for AC, 251(43.38%) patients had recurrences[local-28(4.89%), 223(38.98%)]. Optimal cut-off of post-operative serum CA 19-9 was found to be 77.85 U/ml for predicting recurrence (sensitivity-61.22%, specificity-77.67%, AUC–0.711). A post-operative rise in CA 19-9 of >200 U/ml was 97.94% specific for diagnosing a recurrence. A serial rise in serum CA 19-9 during follow-up was a more accurate predictor of recurrence (sensitivity-71.05%, specificity-91.67%). The optimal cut off of relative rise in serum CA 19.9 in diagnosing a recurrence was established at 2.79x (sensitivity–46.26%, specificity–83.33%, AUC–0.614). The median duration between the first rise in CA 19-9 (>37 U/ml) and radiological evidence of recurrence was 4.04 months. In patients who had early recurrence detection post rise in serum CA 19-9 levels, the median survival after recurrence detection was found to be superior (12.8 months versus 7.6 months, p = 0.005).

Conclusions : Serum CA 19-9 has the potential to diagnose early recurrences in AC, leading to prompt institution of therapy, potentially culminating into longer OS and therefore should be recommended as a routine test in the follow-up evaluation of resected ACs.