Carcinoembryonic antigen: a useful prognostic marker in small-cell lung cancer.

Abstract

Plasma carcinoembryonic antigen (CEA) was determined in 180 patients with small-cell lung cancer (SCLC) before treatment. An abnormal level (greater than or equal to 6 ng/mL) was found in 34% of patients tested. Patients with extensive disease (39/83) had a significantly higher frequency of abnormal CEA (P = .001) than those with limited disease (22/97). There was a strong correlation between obtaining an objective response--particularly a complete response (P = .00003)--and the absence of an elevated CEA. Patients with an abnormal CEA also had a shorter survival time (P = .0007) and the difference remained statistically significant after logrank adjustment for extent of disease and ECOG (Eastern Cooperative Oncology Group) performance status. There was also a negative correlation between survival time and the quantitative level of CEA. In this series, only the group of patients with normal initial CEA levels included all survivors beyond 2.5 years. We conclude that CEA is a useful prognostic factor in SCLC.

The prognostic influence of serum neuron specific enolase in small cell lung cancer

“…A negative correlation was described between initial high or low levels of NSE and survival (Akoun et al, 1985). Contradictory results have been found for CEA (Sculier et al, 1985, Bucceri et al, 1987, while a normalization of serum AGP level during chemotherapy involved a longer disease-free survival (Ganz et al, 1984). These investigations on tumour markers are based on univariate analysis which limits the conclusions to be drawn.…”

Tumour markers for prediction of survival and monitoring of remission in small cell lung cancer

“…The use of CEA as a prognostic factor has yielded inconclusive results with some studies reporting a relationship between presentation level and survival (Sculier et al, 1985;Laberge et al, 1987;Krischke et al, 1988), although multivariate analyses were not performed and others found no correlation (Lokich, 1982;Waalkes et al, 1982;Jaques et at., 1988), a result confirmed in this study. Studies of serial measurements of CEA have suggested that these correlate closely with the clinical course (Woo et al, 1981;Havemann et al, 1985;Shinkai et al, 1986) although this was not the finding in this study.…”

“…Several previous studies have examined the usefulness of CEA as a marker for small cell lung cancer, but as in this case the proportion of patients with elevated levels at presentation has generally been reported as less than half (Goslin et al, 1981;Lokich, 1982;Sculier et al, 1985), limiting its applicability. The use of CEA as a prognostic factor has yielded inconclusive results with some studies reporting a relationship between presentation level and survival (Sculier et al, 1985;Laberge et al, 1987;Krischke et al, 1988), although multivariate analyses were not performed and others found no correlation (Lokich, 1982;Waalkes et al, 1982;Jaques et at., 1988), a result confirmed in this study.…”

“…Enzymes (Carney et al, 1982;Bork et al, 1988), secretory peptides Sobol et al, 1986;North et al, 1988), hormones (Hansen et al, 1980) and cell surface molecules (Sculier et al, 1985;Jaques et al, 1988) have all been assayed in the circulation in the hope of identifying those which might contribute to the management of patients, either by screening for early diagnosis, defining prognosis or guiding treatment. To date however none has proven overwhelmingly superior and the need remains to examine potential candidates for their applicability in a variety of patient groups. This report details the results of an analysis of neurone specific enolase (NSE) …”

An immunohistochemical investigation of diagnostic biopsy material taken from short and long term survivors with small cell lung cancer

“…Serum markers such as NSE (Harding et al, 1990) are related to prognosis but appear to be indicators of mass of disease rather than independent predictors of outcome. The same has been suggested for CEA (Sculier et al, 1985;Laberge et al, 1987) In the present study we have tried to avoid some of these difficulties. The groups of patients have been taken at the extremes of prognosis (survival < 3 months or > 2 years) since, if no difference is detectable, it seems unlikely that one will be detected from a population with a wide distribution of prognosis.…”