Objective To analyze the clinical and pathological features of lung cancer with metastasis, explore the regularity and characteristics of the location of metastasis, and provide reference for future clinical treatment. Methods A total of 658 patients with lung cancer treated in West China Hospital of Sichuan University from January 2008 to December 2014 were enrolled in this study. The effect of different clinical and pathological characteristics on different locations of metastasis was analyzed by χ2 test and logistic regression. Results Adenocarcinoma was the main pathological type (342 cases, 52.0%), and bone (150 cases, 22.8%) and pleura (118 cases, 18.0%) were the most common distant metastasis. Compared with patients with no corresponding metastasis, patients with age <60 years took bigger proportions in patients with bone, brain and mediastina metastasis ( P<0.05). Furthermore, logistic regression analysis showed that the younger patients were more likely to have brain metastasis (P=0.024). Besides, the elder patients were more common in those with liver metastasis (P<0.001). The proportion of males was higher in the patients with lymph node metastasis than those without lymph node metastasis (P=0.010); however, the proportion of females was higher in patients with bone or pleural metastasis than those without bone or pleural metastasis (P<0.05). There was no significant difference in gender among patients with brain, lung, liver, adrenal and mediastinal metastases (P>0.05). Conclusions Bone and pleura are the most common sites of metastasis of lung cancer. The age structure of brain metastasis tends to be younger.
ObjectiveTo explore the prognostic value of fasting blood glucose concentration in patients with newly diagnosed lung cancer.MethodsThe clinical data of 956 patients with lung cancer who were first diagnosed at West China Hospital of Sichuan University between January 2008 and December 2011 were retrospectively analyzed. The patients were followed up for more than 5 years. Using the fasting blood glucose concentration of 6.1 mmol/L as the cut-off value, the patients were divided into the hyperglycemia group and the control group. Kaplan-Meier method was used for survival analysis, and log-rank test was used to analyze the survival of different groups. Univariate and multivariate Cox proportional hazard models were used to evaluate the prognostic variables.ResultsThere were 166 patients in the hyperglycemia group with a 5-year overall survival rate of 23.5%, and 790 patients in the control group with a 5-year survival rate of 30.8%, and the difference between the two groups was statistically significant (P=0.008). Univariate Cox proportional hazard analysis found that blood glucose concentration, gender, age, smoking history, staging, and whether surgery were factors that affected the 5-year survival rate of patients (P<0.05); multivariate Cox proportional hazard analysis showed that blood glucose concentration [hazard ratio (HR)=1.235, 95% confidence interval (CI) (1.013, 1.504), P=0.036], age [HR=1.305, 95%CI (1.110, 1.534), P=0.001], smoking history [HR=1.210, 95%CI (1.033, 1.418), P=0.018], staging [HR=1.546, 95%CI (1.172, 2.040), P=0.002], and whether surgical treatment [HR=0.330, 95%CI (0.257, 0.424), P<0.001] were independent factors which influenced 5-year survival rate. Blood glucose concentration, age, smoking history, and staging were independent risk factors.ConclusionFasting blood glucose concentration is able to be a prognostic factor for patients with newly diagnosed lung cancer.
ObjectiveTo explore the prognostic value of modified Glasgow Prognostic Score (mGPS) in lung cancer patients.MethodsThe clinical data and follow-up information of patients with lung cancer diagnosed for the first time in West China Hospital of Sichuan University from August 2008 to May 2013 were retrospectively analyzed. Overall survival (OS) of patients with different mGPS were compared by Kaplan-Meier test and log-rank test. Univariate and multivariate Cox proportional hazard analysis were performed, and hazard ratio (HR) and 95% confidence interval (CI) were counted to evaluate the predictive value of different prognostic factors in patients with lung cancer.ResultsA total of 289 patients were included. According to the mGPS score, 127 patients had 0 point, 90 patients had 1 point, and 72 patients had 2 points. The OS of lung cancer patients with mGPS=0 was better than that of patients with mGPS=1 and mGPS=2 (P<0.001). Cox proportional hazards of univariate analysis revealed that age< 65 (P=0.022), stage for Ⅰand Ⅱ (P<0.001), surgery (P<0.001), chemotherapy (P=0.018), and mGPS=0 (1 vs. 0, P=0.008; 2 vs. 0, P<0.001) were the protective factors for lung cancer patients (P<0.05). Multiple-factor analysis showed that age [HR=0.680, 95%CI (0.508, 0.911), P=0.010], stage [HR=0.580, 95%CI (0.359, 0.939), P=0.027], operation [HR=0.254, 95%CI (0.140, 0.459), P<0.001], chemotherapy [HR=0.624, 95%CI (0.435, 0.893), P=0.010], mGPS (1 vs. 0) [HR=1.548, 95%CI (1.101, 2.176), P=0.012] and mGPS (2 vs. 0) [HR=1.425, 95%CI (1.003, 2.024), P=0.048] were independent predictors of OS in patients with lung cancer.ConclusionmGPS could be considered as an independent prognostic factor in lung cancer.
ObjectiveTo identify differences in blood routine indicators between lung cancer patients and healthy controls, and between different subgroups of lung cancer patients, so as to improve the early detection of lung cancer prognosis, and provide a basis for risk stratification and prognostic judgment for patients with lung cancer.MethodsThis study enrolled 1 227 patients pathologically diagnosed with lung cancer from December 2008 to December 2013 and 2 454 healthy controls 1∶2 matched by sex and age. The blood routine data of lung cancer patients were collected when they were first diagnosed with lung cancer. Gender and age stratified analysis of blood routine indicators between lung cancer patients and controls were conducted. Comparisons of blood routine indicators among lung cancer patients with different pathological types, stages, and prognosis were performed, followed by Cox regression survival analysis. Normally distributed quantitative variables were presented as mean ± standard deviation and non-normally distributed quantitative variables as medium (lower quartile, upper quartile).ResultsCompared to healthy controls, the counts of platelet [(206.84±80.47) vs. (175.27±55.74)×109/L], white blood cells [(7.04±2.29) vs. (6.08±1.40)×109/L], neutrophil [(4.90±2.08) vs. (3.61±1.07)×109/L], monocyte [0.42 (0.30, 0.54) vs. 0.33 (0.26, 0.42)×109/L], and eosinophil [0.14 (0.07, 0.24) vs. 0.12 (0.07, 0.19)×109/L], as the well as neutrophil-lymphocytes ratio (3.91±2.82 vs. 2.03±0.89) and platelet-lymphocyte ratio (160.35±96.06 vs. 96.93±38.02) in lung cancer patients increased significantly, while the counts of red blood cells [(4.41±0.58) vs. (4.85±0.51)×1012/L] and lymphocyte [(1.49±0.60) vs. (1.93±0.59)×109/L] in lung cancer patients decreased, and the differences were statistically significant (P<0.05). The counts of platelet, red blood cells, white blood cells, neutrophil, and monocyte differed among patients with different pathological types, tumor stages, and prognosis (P<0.05). Neutrophil-lymphocytes ratio and platelet-lymphocyte ratio were higher in squamous cell carcinoma patients than those in other pathological patients, higher in advanced lung cancer patients than those in early stage patients, and higher in dead lung cancer patients than those in survival patients (P<0.05). Neutrophil-lymphocyte ratio was an independent factor affecting the prognosis of lung cancer [hazard ratio=1.077, 95% confidence interval (1.051, 1.103), P<0.001].ConclusionsThe inflammatory index of blood routine indicators are higher in lung cancer patients than those in healthy controls, which indicates that lung cancer is closely related to chronic inflammation. There are significant differences in blood routine inflammation index among lung cancer patients with different pathological types, stages, and prognosis, which reflects the heterogeneity and complexity of lung cancer. Neutrophil-lymphocytes ratio inverse correlates with the prognosis of lung cancer.
Objective To systematically review the cost-effectiveness of gefitinib for advanced non-small cell lung cancer (NSCLC), in order to provide the economics values of gefitinib for clinical application. Method We electronically searched databases including PubMed, Ovid, Embase, Cochrane Library, Medline, China National Knowledge Internet, VIP, and Wanfang database for articles about the cost-effectiveness of gefitinib for advanced NSCLC patients from January 1946 to October 2017, and then performed a systematic literature review of economic evaluations of geftinib. Results A total of 20 independent studies were included in the present systematic review, in which 8 were the first-line treatment, 9 were the second-line treatment, 1 was the third-line treatment, and 2 were maintenance treatment. The most common comparison was gefitinib vs. chemotherapy (n=7), and other comparisons were gefitinib vs. erlotinib (n=4), gefitinib vs. docetaxel (n=3), gefitinib vs. placebo (n=2), gefitinib vs. icotinib (n=2), gefitinib vs. afatinib (n=1), and gefitinib vs. other treatments (n=1). For the advanced NSCLC patients, the first- or second-line treatment with gefitinib compared to chemotherapy was considered to be more cost-effective, especially in patients with mutated epidermal growth factor receptor gene. As the second-line treatment, gefitinib was considered to be more economical than erlotinib and docetaxel. Conclusion Gefitinib is considered to be a cost-effective strategy for the advanced NSCLC patients as the first- or second-line therapy.
Objective To analyze the clinical features and survival of lung cancer with pleural effusions. Methods A total of 982 consecutive patients with a newly diagnosed lung cancer from January 2008 to December 2014 were retrospectively reviewed. To analyze the clinical features and survival differences, the total patients were divided into the following two groups: with (n=204) or without (n=778) pleural effusions. Results Lung cancer comprised 682 (69.5%) males and 300 (30.5%) females, with an average age of 59.74 years (19–93 years). There were 487(49.6%) squamous carcinoma, 254 (25.9%) adenocarcinoma and 166 (16.9%) small cell lung cancer; 113 (11.5%) lung cancer at early stage (Ⅰ–Ⅱ), 247 (25.2%) cases at stage Ⅲ and 567 (57.7%) at stage Ⅳ. The median survival time of all patients was 12 months. Patients with pleural effusions had a worse prognosis compared to patients without (median survival time: 11 vs.12 months, P=0.003), the median survival time could be reduced by 1 month in males (P=0.004), 3 months in elder patients over 60 years (P<0.001), 4 to 8 months in carcinoma and small cell lung cancer (P≤0.001), and 2 to 3 months in advanced lung cancer (stage Ⅲ and Ⅳ) (P<0.05). Any or combined treatment of surgery, radiotherapy, chemotherapy and targeted therapy was associated with an improved overall survival of about 2 months (P=0.009), and targeted therapy could even improve the median survival time by 1 to 8 months (P=0.002). Conclusions About 20.8% of the patients developed pleural effusion at the same time during the course of lung cancer. Pleural effusion is a poor prognostic factor of lung cancer.
Objective To develop a machine learning (ML) model to predict the risk of death in intensive care unit (ICU) patients with chronic obstructive pulmonary disease (COPD), explain the factors related to the risk of death in COPD patients, and solve the "black box" problem of ML model. Methods A total of 8088 patients with severe COPD were selected from the eICU Collaborative Research Database (eICU-CRD). Data within the initial 24 hours of each ICU stay were extracted and randomly divided, with 70% for model training and 30% for model validation. The LASSO regression was deployed for predictor variable selection to avoid overfitting. Five ML models were employed to predict in-hospital mortality. The prediction performance of the ML models was compared with alternative models using the area under curve (AUC), while SHAP (SHapley Additive exPlanations) method was used to explain this random forest (RF) model. Results The RF model performed best among the APACHE IVa scoring system and five ML models with the AUC of 0.830 (95%CI 0.806 - 0.855). The SHAP method detects the top 10 predictors according to the importance ranking and the minimum of non-invasive systolic blood pressure was recognized as the most significant predictor variable. Conclusion Leveraging ML model to capture risk factors and using the SHAP method to interpret the prediction outcome can predict the risk of death of patients early, which helps clinicians make accurate treatment plans and allocate medical resources rationally.