| Objective: to apply the clinical commonly used indicators to establish the diagnosis model(nomograph)for the diagnosis of gastric cancer and gastric cancer with distant metastasis,it can provide certain guiding significance to the diagnosis and treatment of gastric cancer.Methods : this study is a retrospective study,divided into two parts.1.Screening for gastric cancer research;200 cases for research group select randomly from October 2013 to October 2016 patients diagnosed with stomach cancer during the period,among which 155 were male,female 45 cases,32 ~ 85 years old.Control group for a random sample of our department in October 2013 to October 2016 hospitalized during the period of patients with gastric benign lesions(polyps)120 cases,55 cases of men,women,65 cases,aged 32 ~ 80 as control group.2.Research on distant metastasis;60 cases selected randomly from October 2013 to October 2016 patients diagnosed gastric cancer with distant metastasis as research group,among which 42 were male,female 18 cases.as control group.120 cases selected randomly from October 2013 to October 2016 patients diagnosed gastric cancer as control group,among which 110 were male,female 40 cases.Records researchers' gender,age,preoperative serum CEA,CA199,CA724 values,the FIB density,preoperative CT tumor diameter as a result,NLR,defecate occult blood and anemia,distant organ metastasis and clinical pathologic TNM staging,etc.Distant organ metastasis include: liver,peritoneum,lungs,bones,and other parts.comparing gastric cancer with gastric polyps in gender,age,preoperative serum CEA,CA199,CA724,FIB,gastric cancer metastasis group comparing with gastric cancer without metastasis in serum CEA,CA199,CA724,FIB,tumor diameter,NLR.Choosing the significative index in the study to set up ROC curve for the diagnosis of distant metastases and gastric cancer.Choose high diagnostic value,namely the larger area under ROC curve index into the nomograph diagnosis model.The nomogram of each index in the corresponding scores into research group and control group.Result:The difference of CEA in gastric cancer and gastric polyps group is statistically significant(P < 0.001).The difference of CA724 in gastric cancer and gastric polyps group is statistically significant(P = 0.002),The difference of fibrinogen(FIB)in gastric cancer and gastric polyps group differences is statistically significant(P = 0.003,CA199 in gastric cancer and gastric polyp groups,the difference is not statistically significant(P = 0.836).Gender distribution in gastric cancer and gastric polyps group has significant difference(P < 0.001),namely the gastric cancer group now for gastric polyp male proportion is much higher than the proportion of women.Proportion of anemia in patients with gastric cancer group and far higher than that of gastric polyps group and has obvious difference(P < 0.001),There is no statistically significant difference of the presence of occult blood in stool gastric polyps and gastric cancer group(P = 0.498).2.ROC curves show the CEA in the area under the curve is AUC = 0.714,diagnostic value is the middle,the best cut-off point for the CEA is 4.0 ng/ml,the sensitivity was 49.25%,specificity of 87%,the area under the curve of CA724 dentify gastric cancer and gastric polyp is 0.61,the area under the curve of FIB dentify gastric cancer and gastric polyp is 0.69.FIB best cut-off point of 3.6 g/ml,the sensitivity of 51.5%,specificity of 85%.the area under the curve of sex dentify gastric cancer and gastric polyp is 0.686,the area under the curve of anemia dentify gastric cancer and gastric polyp is 0.701.the area under the curve of four index dentify gastric cancer and gastric polyp is 0.861.3.Through the selected four kinds of indexes to establish a logistic regression model Logit(P)= 7.11 + 0.589 FIB CEA sex + 2.06 + 0.572 + 2.171 hyp.establishing a nomograph found CEA scores 100 points,as for fibrinogen 50 points,anemia,scoring 90 points,gender score 50 points.Fibrinogen,carcinoembryonic antigen,anemia,sex represented by the score test after gastric cancer and gastric polyps group respectively,according to different index combination to produce predictive value of cancer of the gastric,the high score,the higher the risk.When a score of 190 points or more,the risk over 92%,the specific degree of almost 100%,it indicates that there is almost no gastric polyps in patients with a score of 190 points,also found that when the score is 150 points,its specificity is 92.5%,the sensitivity is 84.4%,the correct index achieve maximum to 76.9%,Best curve is 140 points,the risk is 78%,therefore,we can clinically basic judgment when the score is 140 points,namely the cancer risk is greater than 78%,we can diagnose the patient with gastric cancer,the misdiagnosis rate of 7.5%,19.7% missed diagnosis.At the same time to reduce the misdiagnosis rate can continue to improve the score,but missed diagnosis will further increase,to improve the sensitivity we can reduce fractions.4,The difference of CEA in gastric cancer and gastric cancer with distant metastases group is statistically significant(P < 0.001),The difference of CA724 in gastric cancer and gastric cancer with distant metastases group is statistically significant(P = 0.002),The difference of FIB in gastric cancer and gastric cancer with distant metastases group is not statistically significant(P = 0.166),The difference of CA199 in gastric cancer and gastric cancer with distant metastases group is not statistically significant(P = 0.836).The difference of NLR in gastric cancer and gastric cancer with distant metastases group is statistically significant P < 0.001,The difference of Dia in gastric cancer and gastric cancer with distant metastases group is statistically significant(P < 0.001)5.ROC curves show the CEA distinct distant metastasis and gastric cancer without metastasis is 0.707,the optimal cut-off point of 6.6 ng/ml,at this time of the sensitivity of 46.5%,specificity of 83.7%.CA724 distinct gastric cancer metastasis and gastric cancer without metastasis is 0.649,the best cut-off point of 6.8 IU/ml,at this time of the sensitivity of 42%,specificity of 77.6%.NLR distincts gastric cancer metastasis and gastric cancer without metastasis of AUC is 0.704,the best cut-off point of 3.2 IU/ml,at this time of the sensitivity of 75%,specificity of 56.4%.diameter(Dia)distinct gastric cancer metastasis and gastric cancer without metastasis is 0.839,the best cut-off point is 4.5 cm,the sensitivity of 92.3%,specificity of 64%.When the NLR and CA124,CEA,tumor diameter(Dia)joint use of four index,area under the curve more than any single indicator area under the curve in the application,its AUC reached 0.91.6.Through the selected four kinds of indexes to establish a logistic regression model Logit(P)= 4.999 + 0.25 NLR CEA CA724 + 0.497 + 0.1 + 0.026 Dia,through the establishment of nomograph found elevated CEA group score 52 points,CA724 rising group score 70 points,higher NLR group score 82 points,the tumor diameter greater than 4.5 cm score 100 points.Each index scores represented respectively inspection gastric cancer metastasis group and gastric cancer metastasis,different indicators combination different gastric cancer metastasis prediction value,the overall results show that the patients with gastric cancer nomograph score is higher,the greater of the the risk.When the four indexes for elevated status,their risk of metastatic gastric cancer for more than 90%.found that all of the patients with gastric cancer metastasis in nomograph model calculating the risk of gastric cancer metastasis is greater than 9%,93.3% of patients with gastric cancer metastasis in the nomograph model is more than 26%,90% of patients with gastric cancer metastasis transfered risk is greater than 45% in the nomograph model calculation.70% of patients with gastric cancer metastasis in nomograph model the transfered risk is greater than 65%,53.3% of patients with gastric cancer metastasis in nomograph model transferred risk is greater than 70%,36.7% of patients with gastric cancer metastasis in transfered risk is greater than 81%,26.7% of patients with gastric cancer metastasis in the nomograph model of metastasis risk is greater than 92%.Also found that when the risk of metastasis is 45%,the score is 182,as the diagnosis of gastric carcinoma metastasis optimum truncation points,has the highest accuracy,sensitivity of 90%,91%,the accuracy reaches 81%.So when score is greater than or equal to 182 gastric cancer patients,the risk is greater than or equal to 45% we can diagnosis of gastric cancer metastasis,the missed diagnosis rate was 10%,the misdiagnosis rate was 9%.We also can according to the actual demand change score,but missed diagnosis and misdiagnosis rate will also along with the change.Conclusion: 1.The application of CEA,FIB,gender and anemia in nomograph diagnostic model for gastric cancer screening has a good value.2.The application of CEA,NLR,CA724 and Dia in nomograph diagnostic model for gastric cancer distant metastasis has a good value. |
PDF Full Text Request