Sixty-three patients with lung (34 small-cell, 18 squamous, 11 adeno-) carcinomas and 43 patients with benign lung diseases were characterized with seven tumor markers: neuron-specific enolase (NSE); cancer antigens CA 19-9, CA 125, CA 15-3, and CA 50; carcinoembryonic antigen (CEA); and tissue polypeptide antigen. Diagnosis had been established by histological examination after surgery and used for classification. After vector transformation, the tumor marker data were fed into neural networks (NNs) based on three different types of learning algorithms: backpropagation (BP), competitive learning (CL), and Hopfield (H). For comparison, the data were evaluated with multivariate stepwise discriminant analysis (MVSDA). BP-NNs are equal to (NSE, CA 19-9, CEA) or better than (100% correct classification when using all seven markers) MVSDA in assigning the correct diagnosis to the patients. Cross-validation data yielded shrinkage effects ranging from 0% to 12.5%. Quality-control (QC) data were evaluated by traditional QC algorithms and compared with the results obtained by a BP-NN. The results show that the BP-NNs could only partly fulfill the tasks of three QC algorithms regarding the violation of static borders but gave good results with respect to dynamic changes.