Are serial measurements of CA19-9 useful in predicting response to chemotherapy in patients with inoperable adenocarcinoma of the pancreas?

Thirty-nine patients with inoperable adenocarcinoma of the pancreas were studied (27 male, 12 female; median age 60 years, range 39-75 years). All patients received chemotherapy with continuous infusion 5-fluorouracil with intravenous bolus epirubicin followed by cisplatin, repeated every 21 days for a total of six cycles and were evaluable for response. Serum CA19-9 concentrations were obtained at baseline and before each cycle. A rise or fall in the tumour marker was defined as a greater than 15% increase or decrease in the marker on two consecutive occasions 3 weeks apart. A plateau in the tumour marker was defined as a less than 15% decrease or increase on two occasions. Changes in marker expression were compared with serial computerized tomography scanning before treatment and after the third and sixth cycle of chemotherapy. Thirty-five of 39 patients had an elevated CA19-9 (87.9%). Thirteen (36.2%) exhibited a decrease, seven (19.4%) a plateau and 16 (44.4%) patients had a progressive rise in serum CA19-9. The sensitivity of CA19-9 was 67% for predicting a partial response and 86% for progressive disease. The median survival for the 13 patients exhibiting a reduction was 333 days, for the seven patients exhibiting a plateau 253 days and for those who had a progressive rise 185 days. The difference in median survival between the group of patients with > 15% decrease and those with > 15% increase of CA19-9 was significant (P = 0.001). In the cohort of patients who exhibited a reduction in CA19-9, no tumour progression was seen, and the reduction occurred during the first three cycles of treatment. Thus, interval scanning may be avoided in this group of patients.