Diagnostic value of an artificial neural network model based on tumor markers for colorectal carcinoma
Keng Li · 2009
Objective To investigate the value of an artificial neural network(ANN) model based on tumor marker in serum for diagnosing colorectal carcinoma.Methods A diagnostic model for colorectal carcinoma based on 7 tumor markers which were assayed combinedly was constructed with ANN.Results The sensitivity of the ANN diagnostic model for colorectal carcinoma was 0.914,the specificity was 0.964,and its accuracy was 0.939.All the above three indexes were higher than that of any single tumor marker(P=0.000).Conclusion Compared with a single tumor marker,the ANN model based on multiple tumor markers in serum can markedly increase the sensitivity,specificity and accuracy for detecting colorectal carcinoma.It has a higher value in early diagnosing colorectal carcinoma.