BACKGROUND: The serum kinetics of carcinoembryonic antigen (CEA) after resection of lung carcinoma are not well characterized. Its prognostic implications remain unclear. This study was designed to clarify the correlation between postoperative CEA time-course and patient prognosis. METHODS: The authors analyzed early postoperative CEA time-course in 31 lung carcinoma patients using nonlinear least square analysis with the following three equations: Equation 1: C(t) = C(0); Equation 2: C(t) = C(0)exp(-kt); and Equation 3: C(t) (C(0)PLAT)exp(-kt) + PLAT, in which t: postoperative day; C(t): postoperative CEA; PLAT: postoperative CEA at plateau; C(0): CEA at postoperative Day zero; and k: rate constant of elimination. The equation that yielded the least Akaike's information criterion was adopted as the best fitting regression equation for each patient. When Equation 3 was adopted, postoperative CEA production (PROD) was calculated as PLAT multiplied by k. RESULTS: Equations 1, 2, and 3 were adopted for 16 (Group 1), 0, and 15 (Group 2) patients, respectively. In Group 1, no decrease in serum CEA level after surgery was detected and CEA production appears to have been little or none. In Group 2, biologic CEA half-life was 1.1 +/- 0.7 days and was not useful for predicting patient prognosis. Tumor recurrences were observed in 9 of the Group 2 patients 19 +/- 9 months postoperatively and there was no significant difference in PLAT or PROD between patients with and without recurrence. Early recurrence within 6 months after surgery was recognized in 5 (early-REC) of the 15 Group 2 patients, in whom there was a tendency for PLAT to be higher than in the other 10 patients without recurrence (early-NON) (early-REC: 3.8 +/- 4.9 ng/mL; early-NON: 1.5 +/- 1.1 ng/mL; P = 0.08). PROD was significantly higher in early-REC than in early-NON (early-REC: 3.5 +/- 4.2 ng/mL/day; early-NON: 0.8 +/- 0.4 ng/mL/day; P < 0.05). CONCLUSIONS: Although PROD is more sensitive than PLAT, both parameters appear to be useful as prognostic tools for predicting early recurrence after resection of lung carcinoma. This is probably because they represent the number of residual tumor cells immediately after surgery.
BACKGROUND: The serum kinetics of carcinoembryonic antigen (CEA) after resection of lung carcinoma are not well characterized. Its prognostic implications remain unclear. This study was designed to clarify the correlation between postoperative CEA time-course and patient prognosis. METHODS: The authors analyzed early postoperative CEA time-course in 31 lung carcinomapatients using nonlinear least square analysis with the following three equations: Equation 1: C(t) = C(0); Equation 2: C(t) = C(0)exp(-kt); and Equation 3: C(t) (C(0)PLAT)exp(-kt) + PLAT, in which t: postoperative day; C(t): postoperative CEA; PLAT: postoperative CEA at plateau; C(0): CEA at postoperative Day zero; and k: rate constant of elimination. The equation that yielded the least Akaike's information criterion was adopted as the best fitting regression equation for each patient. When Equation 3 was adopted, postoperative CEA production (PROD) was calculated as PLAT multiplied by k. RESULTS: Equations 1, 2, and 3 were adopted for 16 (Group 1), 0, and 15 (Group 2) patients, respectively. In Group 1, no decrease in serum CEA level after surgery was detected and CEA production appears to have been little or none. In Group 2, biologic CEA half-life was 1.1 +/- 0.7 days and was not useful for predicting patient prognosis. Tumor recurrences were observed in 9 of the Group 2 patients 19 +/- 9 months postoperatively and there was no significant difference in PLAT or PROD between patients with and without recurrence. Early recurrence within 6 months after surgery was recognized in 5 (early-REC) of the 15 Group 2 patients, in whom there was a tendency for PLAT to be higher than in the other 10 patients without recurrence (early-NON) (early-REC: 3.8 +/- 4.9 ng/mL; early-NON: 1.5 +/- 1.1 ng/mL; P = 0.08). PROD was significantly higher in early-REC than in early-NON (early-REC: 3.5 +/- 4.2 ng/mL/day; early-NON: 0.8 +/- 0.4 ng/mL/day; P < 0.05). CONCLUSIONS: Although PROD is more sensitive than PLAT, both parameters appear to be useful as prognostic tools for predicting early recurrence after resection of lung carcinoma. This is probably because they represent the number of residual tumor cells immediately after surgery.