Loading
In a randomised trial, a high CA 19-9 before treatment predicted shorter survival, but a 50% fall during chemotherapy did not mean living longer once the analysis was corrected, so a falling marker is not a valid substitute for survival.
CA 19-9 was measured at baseline and every 3 weeks in patients with histologically proven advanced pancreatic carcinoma in a randomised trial of gemcitabine versus gemcitabine plus capecitabine; those with normal or missing baseline values were excluded. Of 319 randomised patients, 247 were assessable for baseline and 175 for marker response, with comparisons corrected for guarantee-time bias by the landmark method. Median overall survival was 5.8 months for baseline at or above the median (59 times the upper limit of normal) against 10.3 months below it (P < 0.0001). An early decrease of at least 50% after two cycles was not associated with longer survival (10.1 versus 8.6 months; hazard ratio 1.11), nor was a 50% decrease at nadir (7.8 versus 6.7 months; hazard ratio 0.95). CA 19-9 response is therefore not a valid surrogate endpoint for survival.
It stops CA 19-9 response being used as a trial endpoint or as the reason to change treatment, while leaving baseline level as a prognostic factor.
Shares Capecitabine, Gemcitabine, Pancreatic ductal adenocarcinoma.
Shares Capecitabine, Gemcitabine, Pancreatic ductal adenocarcinoma.
Shares CA 19-9, Pancreatic ductal adenocarcinoma.
Shares CA 19-9, Metastatic pancreatic ductal adenocarcinoma, Pancreatic ductal adenocarcinoma.
Shares Capecitabine, Gemcitabine, Pancreatic ductal adenocarcinoma.
Shares Capecitabine, Gemcitabine, Pancreatic ductal adenocarcinoma.
Shares Capecitabine, Gemcitabine, Pancreatic ductal adenocarcinoma.
Shares Capecitabine, Gemcitabine, Metastatic pancreatic ductal adenocarcinoma.