The diagnostic value of neuron-specific enolase and carcino-embryonic antigen analyses in patients with carcinoma of the lung.

Neuron-specific enolase and carcino-embryonic antigen were quantified simultaneously in sera of 135 patients attending the Department of Respiratory Diseases for diagnostic bronchoscopy. Fifteen small cell lung carcinomas, 24 non-small cell lung carcinomas and 96 benign pulmonary diseases were investigated. Lung biopsies or bronchial washings were obtained from about 75% of the patients, including all patients with neoplastic diseases. Serum neuron-specific enolase was measured by a recently introduced enzyme-immuno assay (WaKo NS-Enolase EIA-II testkit). The results obtained with this kit were similar to those based on RIA assays. Receiver Operating Characteristic curves (ROC curves) were constructed for comparison of the discriminating ability of neuron-specific enolase and carcino-embryonic antigen in small cell lung carcinomas and non-small cell lung carcinomas. For small cell lung carcinomas the sensitivity and the specificity of neuron-specific enolase (cutoff value: 10 micrograms/l) were 87% and 88%, respectively, and for carcino-embryonic antigen values 60% and 77% were obtained. There was no correlation between neuron-specific enolase and carcino-embryonic antigen in small cell lung carcinoma patients. The diagnostic value of neuron-specific enolase and carcino-embryonic antigen in non-small cell lung carcinomas is illustrated by sensitivities of 13% and 58%, respectively. An extensive literature survey is included to allow comparison with other studies. The use of ROC curves is recommended for the determination of optimal cutoff values for the assays employed.

Laboratory standardizatin in an international study: the Kaunas--Rotterdam intervention study (KRIS).

Laboratory results obtained in different laboratories over lengthy periods of time usually are difficult to compare. In cooperative long-term studies where such results must be pooled, thorough standardization of methods is vital. We describe a program in which comparable plasma cholesterol and glucose analyses have been obtained, by simple methods. In the Netherlands and the Soviet Union in close collaboration with the Center for Disease Control, Atlanta, Ga., U.S.A. The two laboratories produced glucose values (direct o-toluidine reaction) within 2% of the target reference values and cholesterol results (direct Liebermann-Burchard reaction) with a consistent 6-8% positive bias over the reference method values. Intralaboratory precision was subject to preset acceptance limits. The use of common control materials, exchange of patient samples, and on-site comparison of all details of laboratory procedures are vital tools in standardization efforts. A laboratory protocol that included quality requirements and rejection criteria was developed and proved to be indispensable. The experience gained should be useful in standardizing inter-laboratory results in similar studies.