Multiple Factors Logistic Regression Analysis Of Axillary Level â…¢ Lymph Nodes Metastases And Micrometastases In Patients With Breast Cancer

Objective To investigate the relationship between clinical and pathological factors and axillary levelⅢlymph nodes metastases and micrometastases in patients with breast cancer.To establish the logistic regression equation and to evaluate the accuracy,sensitivity and specificity for predicting axillary levelⅢlymph nodes metastases and micrometastases.Thus help to determine whether levelⅢlymph nodes dissection was adopted.Methods 100 female breast cancer patients underwent modified radical mastectomy therapy between 2005 and 2007 aged 26-76 years were investigated. The tumors were identified by postoperative histopathological examination. After the resection of mammary and the dissection of axillary levelⅠ,Ⅱlymph nodes,axillary levelⅢlymph nodes were dissected by distracting the major and minor pectoral muscles.Of the patients without pathologically verified metastasis,series sections HE staining,cytokeratin 19(CK19)and epithelial membrane antigen(EMA)immunohistochemical staining were performed respectively to detect micrometastases.The clinical and pathological factors that may potentially influence the levelⅢlymph nodes metastases and micrometastases were collected.The statistic analysis was performed by using SPSS 13.0 software.Single and multiple factor logistic regression analysis were used to analyze the data.Discriminant equations were established to predict axillary levelⅢlymph nodes metastases and micrometastases:P=exp∑BiXi/1+exp∑BiXi.The odds ratio was ealeulated:OR=exp(B),and the accuracy,sensitivity,specificity of the equation were evaluated respectively.Results1.Of 100 patients with breast cancer,322 levelⅢlymph nodes were collected,the average number was 3.22.The number of the patients whose axillary lymph nodes were positive confirmed by routine histopathological examination was 42,and of the 42 patients there were 25 cases whose levelⅢlymph nodes were positive.Of the 42 patients,35(83.33%)cases' axillary lymph nodes metastases followed the regularity of levelⅠ→levelⅡ→levelⅢ.The axillary lymph nodes leap metastases were found in 7 patients (16.7%),four patients with levelⅠand levelⅢpositive,one patient with levelⅡand levelⅢpositive,one patient with levelⅡpositive and one patient with levelⅢpositive.2.Of 235 levelⅢlymph nodes from 75 patients whose levelⅢlymph nodes were negative comfirmed by routine histopathological examination,6 micrometastases(0.2mm＜φ＜2mm)from 6 patients were found by series sections HE staining.The micrometastases check rate of series sections HE staining was 2.55%(6/235)and the patients check rate was 8.00%(6/75).By IHC staining,3 micrometastases(φ≤0.2mm)from 3 patients were found manifested as single or cell clusters mainly in the marginal sinus,besides the 6 micrometastases(0.2mm＜φ＜2mm).The micrometastases check rate of IHC staining was 3.83%(9/235)and the patients check rate was 12.00% (9/75).The incidence of levelⅢlymph nodes micrometastases wasl2.00% (9/75).The incidence of levelⅢlymph nodes metastases and micrometastases was 34.00%(34/100). 3.The results of single factor regression of levelⅢlymph nodes metastases showed:of 14 clinical and pathological variables,4 were selected: tumor size,tumor location,the level of CerbB-2 expression,the grade of the number of positive levelⅠlymph nodes and levelⅡlymph nodes metastases status(OR＞1,P＜0.05).By multiple factors logistic regression analysis,it was revealed that there were 3 significant factors:tumor location(especially located in the lateral side,OR=6.18,P＜0.05),the grade of number of positive levelⅠlymph nodes(OR=5.32,P＜0.05)and tumor size(OR=4.02,P＜0.05).Logistic regression equation:P= 1/1+e^{-(-12.59+1.82X tumor location+1.67X the grade of number of positive levelⅠlymph nodes+1.39X tumor size)}The accuracy,sensitivity and specificity of the logistic regression equation were 91.00%(91/100),84.00%(21/25),93.33%(70/75).4.The results of single factor regression of levelⅢlymph nodes metastases and micrometastases showed:of 14 clinical and pathological variables,6 were selected:tumor size,tumor location,the level of CerbB-2 expression,the grade of the number of positive levelⅠlymph nodes,levelⅡlymph nodes metastasis status and pathological stage(OR＞1,P＜0.05).By multiple factors logistic regresion analysis,it was revealed that there were 4 significant factors:pathological stage(OR=10.64,P＜0.05),tumor size (OR=4.01,P＜0.05),tumor location(especially located in the lateral side, OR=3.28,P＜0.05),and the grade of number of positive levelⅠlymph nodes(OR=2.41,P＜0.05).Logistic regression equation:P=1/1+e^{-(-9.67+2.36X pathological stage +1.39X tumor size+1.19X tumor location+0.88Xthe grade of number of positive levelⅠlymph nodes)}.The accuracy,sensitivity and specificity of the logistic regression equation were 90.00%(90/100),85.29%(29/34),92.42%(61/66).Conclution1.It was practicable to expose and dissect axillary levelⅢlymph nodes by distracting the major and minor pectoral muscles,the average number of dissected axillary levelⅢlymph nodes was 3.22.The rate of axillary levelⅢlymph nodes metastases was 25.00%.83.33%of the axillary lymph nodes metastases followed the regularity of levelⅠ→levelⅡ→levelⅢ.Axillary lymph nodes leap metastases were found in a few patients(16.67%).Besides levelⅠ,levelⅡmetastases status,other clinical and pathological factor should be considered when determine the dissection of levelⅢlymph nodes while there are leap metastases.2.By series sections HE staining and by the combined application of EMA and CK19,the check rates of levelⅢlymph nodes micrometastases is 3.83%(9/235).It was practicable to detect micrometastases by series sections HE staining and the combined application of EMA and CK19.The incidence of levelⅢlymph nodes micrometastases was 12.00%(9/75).The incidence of levelⅢlymph nodes metastases and micrometastases increased to 34.00% (34/100).3.The results of multiple factor logistic regression analysis showed the independent risk factors of levelⅢlymph nodes metastases were tumor location(especially located in the lateral side),the grade of tumor size and the grade of the number of positive levelⅠlymph nodes(OR＞1,P＜0.05).The logistic regression equation based on these factors was reliable in predicting the levelⅢlymph nodes metastases.Therefore it can help to make rational decision for the dissection of levelⅢlymph nodes.4.The results of multiple factor logistic regression analysis showed the independent risk factors of levelⅢlymph nodes metastases and micrometastases were tumor location(especially located in the lateral side),the grade of tumor size,the grade of the number of positive levelⅠlymph nodes and pathological stage(OR＞1,P＜0.05).The logistic regression equation based on these factors was reliable in predicting the levelⅢlymph nodes metastases and micrometastases.Therefore it can help to make rational decision for the dissection of levelⅢlymph nodes.When considering whether micrometastases exist,pathological stage can not be neglected.