A Study Of Intraoperative Molecular Diagnosis For Detecting Sentinel Lymph Nodes Metastasis And Prediction Of Non-sentinel Lymph Node Metastasis In Breast Cancer Patients

Objective:The aim of the study is to evaluate the value of One-Step Nucleic Acid Amplification(OSNA)as an intraoperative diagnostic method of sentinel lymph node(SLN)for breast cancer patient and to establish a nomogram for the intraoperative prediction of non-sentinel lymph node(NSLN)metastasis in breast cancer patients by analyzing the results of OSNA techniques so as to guide the following treatment efficiently.Methods:552 patients of 2010 OSNA clinical trial and 348 patients of this prospective study were enrolled.870 of 900 patients were valid cases.If the node weighed less than 100 mg,it was only tested by histopathological examination.If the node weighed 100-1200 mg,it can be cut perpendicular to the short axis.If the length of short axis was ?4mm,the node was cut into two blocks(a,b).Block a was performed by OSNA and block b was performed by histopathological examination.If the length of short axis was>4mm,the node was cut into four blocks(a,b,c,d).Blocks a and c were performed by OSNA.Blocks b and d were prepared for histopathological examination.All blocks(a,b,c,d)were examined by touch print cytology(TIC)intraoperatively.On one hand,we made a comparison of the OSNA assay and postoperative histopathological examination.On the other hand,we established a model for the intraoperative prediction of NSLN metastasis in breast cancer patients according to the results of OSNA assay.Results:The diagnostic value of OSNA compared to FS and TICAccording to the analysis of patients,sensitivity for OSNA was calculated as 88.7%,accuracy as 89.0%,specificity as 89.1%%,posive predict value(PPV)as 72.3% and negative predict value(NPV)was 96.1%.As for Frozen Section(FS),the sensitivity,accuracy,specificity,PPV and NPV were 77.9%,94.0%,100.0%,100.0%,92.3%,respectively.The sensitivity of TIC was 83.0%,the accuracy was 93.5%,the specificity was 96.8%,the PPV was 89.3% and the NPV was 94.7%.The sensitivity of OSNA was significantly better than FS(88.7% vs.77.9%,P=0.0095)and has no difference with that of TIC(88.7% vs.83.0%,P=0.1254).According to the analysis of lymph nodes,the sensitivity,accuracy,specificity,PPV and NPV of the OSNA compared with the histopathological examination were 85.4%,91.4%,92.6%,68.4%,97.1%,respectively.The sensitivity and accuracy of FS were 68.6%,94.3% with a specificity of 100.0%.PPV and NPV were 100.0%,93.5%.For TIC,81.0%,94.4%,96.9%,83.2%,96.5%.The sensitivity of OSNA was significantly better than FS(85.4% vs.68.6%,P=0.0021)and has no difference with that of TIC(85.4% vs.81.0%,P=0.82).With regard to SLN with macro metastases,the sensitivity of OSNA was 93.1%.For FS,88.5% and for TIC,92.2%.There was no difference between them(P=0.1907,P=0.8543).When referring to SLN with micro metastases,the sensitivity of OSNA was superior to FS and TIC(P=0.0121,P=0.0139).Intraoperative model for predicting NSLN metastasis103 with SLN metastasis and treated with axillary lymph node dissection(ALND)were assessed to establish a nomogram for intraoperative prediction of non-sentinel lymph node metastasis based on the molecular diagnosis.Univariate analysis revealed that primary tumor size(P=0.001),lymphovascular invasion(P=0.007),total tumor load(P=0.000),metastasis size in positive SLNs(P=0.000),the number of positive SLNs(P=0.000)and negative SLNs(P=0.000)and the ratio of number of metastatic SLN on total number of SLN(P=0.000)were associated with the presence of NSLN metastasis.Primary tumor size(P=0.002),total tumor load(P=0.013),the number of positive SLNs(P=0.000)and negative SLNs(P=0.001)were identified as independent predictors of NSLN metastasis based on the multivariable logistic regression results and a nomogram was then established with these variables.A validation cohort of 61 patients met the similar criteria subsequently validated it.The area under the ROC curve was 0.814 for the predictive model and it was 0.842 in the re-validation cohort.Then,we applied our model to the patients to assess whether or not it could differentiate the patients with pN1 and ? pN2.We found that the AUC of the ROC curve was 0.861(P<0.0001)and the cut-off value was 45.4%.So as to predict intraoperatively,the tumor size assessed by the postoperative histological examination was replaced by the size evaluated by the imaging examination,and the area under the ROC curve was 0.838 in this circumstance.There was no statistical difference in the accuracy compared with the former validation data(P=0.7406).What's more,the AUC that could differentiate the patients with pN1 and ? pN2 was 0.839(P<0.0001)and the cut-off value was 33.5%.We chose three predictive systems(M.D.Anderson Cancer Center,Tenon model,Isabel model)which were adopted for the validation using our data.The AUC of the three models were 0.745,0.623,0.834,respectively.Statistically significant difference was observed between our model and MDA and Tenon model.However,there was no significant difference between our model and Isabel model.Conclusions:The OSNA is an accurate,rapid intraoperative assay.It has greatly reduced subjectivity and workload of the pathologist as a objective technology.We can regard it as the preferred intra/post-operative diagnosis for SLN and it is suited to be applicated in medical practice of China.The predictive nomogram based on the molecular diagnosis can predict the NSLN metastases intra/post-operatively.It appears to be obviously superior to other predictive models and may help to guide the axillary management and to make decisions about radiation target region.