| Objectives: Lung cancer is one of the most common and fatal malignant tumors in clinic.Recurrence and metastasis of lung cancer significantly increase the risk of death.The determination of metastatic foci is of great guiding value for patients with primary lung cancer to make treatment plan and judge prognosis.This study aimed to establish an nomogram prediction model to predict the metastasis probability of newly diagnosed lung cancer patients based on their tumor markers levels,so as to provide reference for clinicians to help each patient develop an individual treatment plan.Methods: A total of 936 cases of newly diagnosed and pathologically confirmed lung cancer from June 1,2019 to June 1,2020 in our hospital were retrospectively analyzed.Their gender,age,past medical history,family history,smoking history,maximum tumor diameter,pathological biopsy method,pathological diagnosis,tumor stage,whether or not they metastasized and the site of metastasis,serum tumor markers(CEA,SF,CYFRA21-1,NSE,SCCA,CA125,CA199,CA153),and imaging characteristics(chest CT,brain CT/ MRI,abdomen CT/ ultrasound,lymph node ultrasound,ECT/ PET-CT)were collected.Single factor and multi-factor logistic regression analysis was used to calculate and determine the optimal risk factors for predicting metastasis and the nomogram model for predicting the risk of lung cancer metastasis was established.The consistency coefficient and calibration curve are used to evaluate the consistency and accuracy of the prediction model.The Bootstrap method was used for internal verification of the model.The clinical decision curve was used to evaluate the clinical utility of the model.Results: In this study,431 patients with lung cancer combined with metastasis at the time of initial diagnosis were investigated,among which 333 cases(36.2%)had distant metastasis,415 cases(45.1%)had lymph node metastasis,154 cases(16.7%)of contralateral lung metastasis,172 cases(18.7%)of pleural metastasis,51 cases(5.5%)of brain metastasis,175 cases(19.0%)of bone metastasis,42cases(4.6%)of adrenal metastasis,41 cases(4.5%)of liver metastasis,172 cases(48.5%)of single organ metastasis and 183 cases(51.5%)of multiple organ metastasis.In univariate analysis,gender,smoking history,pathological type,CEA,CA125,CA199,CA153,SF,CYFRA21-1,NSE and SCCA were correlated with the metastasis,and the differences were statistically significant(P<0.05).Multivariate logistic regression analysis showed that CEA,CA125,CYFRA21-1,CA153 and CA199 were risk factors for lung cancer metastasis(P<0.05).Increased CEA was a risk factor for bone metastasis(OR = 1.059,95% CI: 1.027–1.092,P < 0.05).The increases in CYFRA21-1 were lymph node metastasis(OR = 1.161,95% CI 1.062–1.269,P < 0.05),contralateral lung metastasis(OR = 1.171,95% CI 1.070–1.280,P <0.05),pleural metastasis(OR = 1.218,95% CI 1.044–1.421,P < 0.05),bone metastasis(OR=1.114,Risk factors for: 95% CI: 1.033–1.201,P< 0.05);adrenal metastasis(OR = 1.185,95% CI: 1.066–1.317,P <0.05);liver metastasis(OR = 1.147,95% CI: 1.060–1.241,P < 0.05).Increased CA125 was a risk factor for brain metastasis(OR = 1.248,95%CI: 1.052–1.524,P < 0.05).The increased CA153 and CA199 were the risk factors for lymph node metastasis(OR = 1.065,95% CI:1.037–1.094,P < 0.05),(OR = 1.024,95% CI: 1.010–1.037,P < 0.05).Based on the results of logistic regression,CEA,CYFRA 21–1,CA125,CA153 and CA199 were included as the characteristic variables of lung cancer metastasis to establish an nomographic prediction model(AUC:0.950,95% CI: 0.936–0.965,Mean Absolute Error: 0.034),suggesting that this model had good prediction ability.Conclusions: 1.Serum tumor markers CEA,CA125,CYFRA21-1,CA153 and CA199 are predictive factors for lung cancer metastasis.2.The increased CEA indicates an increased risk of bone metastasis.The increased CA125 index indicates an increased risk of brain metastasis.The increased CYFRA21-1 index suggests an increased risk of lymph node,contralateral lung,pleura,bone,adrenal gland and liver metastasis.The increased indexes of CA153 and CA199 suggested that the risk of lymph node metastasis was increased.3.The nomogram model of lung cancer metastasis risk based on serum tumor markers has good prediction value,and is conducive to making a more accurate assessment of the metastasis risk. |
