Analysis Of Factors Influencing The Efficacy Of Neoadjuvant Chemotherapy For Breast Cancer And The Construction And Evaluation Of The Predictive Models

Background:Breast cancer is currently the malignant tumor with the highest morbidity and mortality rate among female patients.It is mainly due to the continuous malignant transformation of normal breast epithelial cells,and its strong invasive and metastatic ability is one of the major causes of death among contemporary patients.Neoadjuvant chemotherapy(NAC)is becoming more and more important in the treatment of breast cancer as the concept of treatment continues to evolve.Since NAC for breast cancer has many advantages and achieving pathologic complete response(pCR)is strongly associated with better prognosis,it is crucial to find people who are sensitive to NAC,especially for the Luminal type of breast cancer.There is no single clinical factor that can directly predict the efficacy of neoadjuvant chemotherapy,so it is important to use clinical indicators to find patients sensitive to NAC treatment.Objective:In this study,we comprehensively analyzed the relationship between serological indexes,imaging indexes,pathological data and postoperative pCR and NAC effectiveness,so as to establish a regression prediction model of comprehensive data.In addition,this study focused on analyzing the influencing factors related to NAC efficacy of Luminal type breast cancer,aiming to provide some guidance significance for individualized and precise comprehensive treatment of breast cancer in clinical practice.Methods:A retrospective analysis was performed on 387 patients who received neoadjuvant chemotherapy from December 2017 to June 2021 in our hospital,randomly assigned to 283 in the modeling group and 104 in the validation group.The hematological,imaging and clinicopathological data of the patients were collected and recorded retrospectively.The study endpoints were effectiveness and pCR after NAC.the clinicopathological data of the study subjects in the modeling group were included in the univariate analysis to screen out the meaningful variables,and these variables were included in the binary logistic regression analysis to screen out the independent influencing factors of NAC effectiveness and pCR,based on which the clinical prediction model to predict the effectiveness and pCR of NAC was constructed.And the predictive performance of the model was evaluated using the receiver opera-ting characteristic curve(ROC).The clinicopathological data of patients in the validation group were substituted into the prediction model separately,and the diagnostic efficacy of the prediction model was tested using the Kappa consistency test.In addition,this study conducted the analysis of factors influencing the efficacy of NAC in the modeling group for Luminal type breast cancer,and the clinicopathological data of the study subjects were included in the univariate analysis to screen out the significant variables,and these variables were included in the binary logistic regression analysis to screen out the independent factors influencing the efficacy of NAC.The ROC curve was also used to evaluate the predictive value of each index,and the optimal cut-off value of the corresponding index was calculated.The difference was considered statistically significant at P<0.05.Results1.Comparison of general information between modeling and validation groups:Statistical analysis of clinicopathological data between modeling and validation groups was done,and it was found that all P values were>0.05.As a result,there was no significant difference in overall data between the two groups,and the baseline comparison between the two groups was balanced and comparable.2.Screening out factors affecting the effectiveness of NAC:In the modeling group,factors related to the effectiveness of NAC were screened out by one-way chi-square test or Fisher’s exact test:platelet level,magnetic resonance gland background parenchymal enhancement(BPE)type,ER,PR,HER-2,Ki67 and Molecular typing,the difference was statistically significant(P<0.05).3.Construction of logistic regression model for predicting the effectiveness of NAC:The meaningful variables screened by univariate analysis were inclu ded in the binary logistic regression analysis with an introduction criterion of 0.05 and an exclusion criterion of 0.10.The stepwise method was used to ana lyze the relevant influencing factors of the effectiveness of NAC and to establi sh the prediction model,and the following factors were finally screened into t he model,and according to the SPSS output The results can be written the m odel formula as:valid model P=ex/(1+ex)X=0.097+0.303*mild reinforcement+2.003*moderate reinforcement+2.660*significant reinforcement-2.081*HER-2-0.119*LuminalB type-0.096*HER-2 positive type+1.847*triple negative type4.Evaluation of the model:Based on the output prediction model,the ROC curve was plotted,and the area under the curve was 0.827(95%CI:0.779-0.875,P<0.001).The results showed that the accuracy and predictive ability of this NAC validity prediction model were good,with the best cut-off value of 0.546,sensitivity of 62.1%,and specificity of 92.5%.The validation group data were substituted into the established prediction model formula for Kappa consistency test,and the results showed that Kappa=0.550,P<0.001,the accuracy was 78.85%,the sensitivity was 79.41%,and the specificity was 77.78%,indicating that the model prediction results were generally consistent with the actual results.5.Screening out factors affecting neoadjuvant chemotherapy pCR:In the model group,the factors associated with effective NAC were screened out by one-way chi-square test or Fisher exact test:platelet level,BPE,ER,PR,HER-2,Ki67 and molecular typing,and the differences were statistically significant(P<0.05).6.Construction of logistic regression model for predicting neoadjuvant che motherapy pCR:The meaningful variables screened out by univariate analysis were included in the binary logistic regression analysis with an introduction cri terion of 0.05 and an exclusion criterion of 0.10.The stepwise method was us ed to analyze the relevant influencing factors of NAC effectiveness and to esta blish a prediction model,and the following factors were finally screened out a nd included in the model,according to the SPSS The model formula can be written as:P=ex/(1+ex)X=0.026+0.862*mild reinforcement+2.312*moderate reinf orcement+2.089*significant reinforcement+2.205*ER-3.327*HER-2-0.010*Luminal B type+0.863*HER-2 positive type+1.281*triple negative type.7.Evaluation of the model:According to the prediction model,the ROC curve was plotted,and the area under the curve was 0.808(95%CI:0.758-0.858,P<0.001).The results showed that the accuracy and predictive ability of this pCR prediction model were good,with the best cut-off value of 0.491,sensitivity of 87%and specificity of 62.1%.The validation group data were substituted into the established prediction model formula for Kappa consistency test,and the results showed that Kappa=0.675,P<0.01,accuracy was 88.46%,sensitivity was 72%,and specificity was 93.67%,indicating that the model prediction results were in general agreement with the actual results.8.For Luminal type breast cancer,the results showed that platelet level,BPE,ER,PR,HER-2,and Ki67 were the influencing factors of NAC efficacy by one-way chi-square test or rank sum test.The results of multifactorial analysis showed that HER-2 positivity was an independent beneficial factor for NAC efficacy,and ER and Ki67 were independent influencing factors for NAC efficacy with best cut-off values of 75%(sensitivity 76.5%and specificity 56.9%)and 27.5%(sensitivity 74.5%and specificity 74.3%),respectively,and the area under the ROC curve was 0.683(95%CI:0.587-0.780,P<0.001),0.776(95%CI:0.693-0.895,P<0.001).Conclusion:1.BPE,HER-2,and molecular typing were independent influencing factors of neoadjuvant chemotherapy effectiveness,and BPE,ER,HER-2,and molecular typing were independent influencing factors of neoadjuvant chemotherapy pCR.2.In this study,two prediction models were constructed.The model for NAC effectiveness has average predictive ability,and the model for pCR of NAC has good predictive ability,which can be used to clinically screen out sensitive patients and provide reference for guiding patients’ precise treatment.3.It was concluded from the analysis of the study that ER,HER-2 and Ki67 were independent influencing factors of NAC efficacy in Luminal type,and the efficacy of NAC was better when ER was<75%,HER-2-positive and Ki67>27.5%.