Comparison of five diagnostic models in the diagnosis of non-small cell lung cancer
CHEN Xiaohong YANG Yuwei▲ XU Peng YANG Wu
Departmant of Clinical Laboratory, Mianyang Central Hospital Affiliated to School of Medicine, University of Electronic Science and Technology of China, Sichuan Province, Mianyang 621000, China
Abstract：Objective To compare the application value of five combined patterns for diagnosing non-small cell lung cancer(NSCLC). Methods A total of 413 NSCLC patients in Mianyang Central Hospital Affiliated to School of Medicine,University of Electronic Science and Technology of China from January 11, 2015 to October 28, 2019 were retrospectively selected as the lung cancer group, a total of 723 patients with benign pulmonary disease and 282 healthy subjects were selected as benign group and healthy control group respectively. Serum carcinoembryonic antigen (CEA), alpha fetoprotein (AFP), carbohydrate antigen 125 (CA125), carbohydrate antigen 199 (CA199) and neuron-specific enolase(NSE) were measured, and were used to develop the models of multiplayer artificial neural network (MPL-ANN), radial basis function artificial neural network (RBF-ANN), decision tree, logistic regression and classical discriminant analysis(CDA). Then the diagnostic efficacy of each model was compared. Results Among the single indicators, CEA had the highest diagnostic value for NSCLC, with the AUC of 0.76 (95%CI: 0.74-0.78), but its specificity was low (64.9%). Of all the diagnostic models for the diagnosis of NSCLC, MPL-ANN model was the optimal with 0.91 (95%CI: 0.89-0.96)of AUC, 75.3% of sensitivity, 91.1% of specificity, as well as 71.7% (76/106) and 94.4% (305/323) of diagnosis accuracy for the lung cancer group and non-lung cancer group, respectively. Conclusion The MPL-ANN model based on tumor markers measurement can be helpful for the diagnosis of NSCLC, which provides a new idea for the differential diagnosis of NSCLC.
Bray F，Ferlay J，Soerjomataram I，et al.Global cancer statistics 2018：GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J].CA Cancer J Clin，2018，68（6）：394-424.
[2]
Siegel RL，Miller KD，Jemal A.Cancer statistics，2018[J].CA Cancer J Clin.2018，68（1）：7-30.
[3]
Jiang ZF，Wang M，Xu JL.Thymidine kinase 1 combined with CEA，CYFRA21-1 and NSE improved its diagnostic value for lung cancer[J].Life Sci，2018，194：1-6.
[4]
Wang J，Chu Y，Li J，et al.The clinical value of carcinoembryonic antigen for tumor metastasis assessment in lung cancer[J].Peer J，2019，7：e7433.
[5]
Bartholomai JA，Frieboes HB.Lung Cancer Survival Prediction via Machine Learning Regression，Classification，and Statistical Techniques[J].Proc IEEE Int Symp Signal Proc Inf Tech，2018，2018：632-637.
Lynch CM，Abdollahi B，Fuqua JD，et al.Prediction of lung cancer patient survival via supervised machine learning classification techniques[J].Int J Med Inform，2017，108：1-8.
[7]
Ganti AK，Wang X，Stinchcombe TE，et al.Clinical prognostic model for older patients with advanced non-small cell lung cancer[J].J Geriatr Oncol，2019，10（4）：555-559.
Hanai T，Yatabe Y，Nakayama Y，et al.Prognostic models in patients with non-small-cell lung cancer using artificial neural networks in comparison with logistic regression[J].Cancer Sci，2003，94（5）：473-477.
Feng F，Wu Y，Wu Y，et al.The Effect of Artificial Neural Network Model Combined with Six Tumor Markers in Auxiliary Diagnosis of Lung Cancer[J].J Med Syst，2012，36（5）：2973-2980.
京公网安备 11010502046607号