Carcinoembryonic antigen in monitoring of response to systemic chemotherapy in patients with metastatic colorectal cancer

The response to chemotherapy of solid tumors is generally assessed by measu ring tumors visualized by imaging. However, the response assessment based o n imaging is not always feasible because patients often have disease not me asurable by imaging, such as diffuse peritoneal dissemination. We evaluated the correlation between the change on imaging and change in CEA levels for assessing chemotherapeutic response of patients with metastatic colorectal cancer. Between July 1993 and August 1999 we retrospectively examined 136 patients with metastatic colorectal carcinoma, all of whom had measurable l esions. Forty patients received oral tegafur-uracil (300 mg/m(2)/day) plus folinic acid (60 mg/day) for 4 weeks, repeated every 5 weeks, as the first- line treatment. Another 96 patients received either a weekly intravenous bo lus injection of 5-fluorouracil (400 mg/m(2)) plus folinic acid (20 mg/m(2) ), or an intravenous bolus injection of 5-fluorouracil (425 mg/m(2)) plus f olinic acid (20 mg/m(2)) for 5 consecutive days every month. Responders, ba sed on CEA assessment, were defined as those with a greater than 50% drop i n CEA level for more than 4 weeks. The pretreatment CEA levels were elevate d beyond the normal cutoff value in 110(81%)patients. A response rate of 18 .4% (95% CI, 11.9-24.9%), including 8 complete remissions and 17 partial re missions, was achieved according to imaging studies. The response rate asse ssed by CEA was 25% (34/136), Sixteen responders (47%) based on CEA had no remission on imaging. The sensitivity of change in CEA levels in the predic tion of true responders and progressive diseases on imaging were 72% and 81 %, respectively. In terms of the positive predictive value, change in CEA l evels in the prediction of true responders and progressive disease on imagi ng were 53% and 85%, respectively. Patients with remarkable falls on CEA le vels survived significantly longer than nonresponders (P <0.001, log-rank t est). At follow-up of 48 months the median survival for responders and nonr esponders assessed by CEA was 28 months and 13 months, respectively. These data suggest that measurement of CEA levels might be helpful in monitoring chemotherapeutic response when imaging study is unsuitable for assessing th e response in clinical practice. Furthermore, measurement of CEA levels may be helpful in determining the prognosis of patients with metastatic colore ctal cancer receiving chemotherapy.