| Purpose:To explore the adding value of applying acoustic radiation force impulse ultrasound(ARFI,one type of ultrasonic elastography technique)after the conventional one in assessing regional lymph nodes(LN)metastatic status of breast caner(BC)before radiotherapy(RT).Methods:The female patients with BC who were admitted at the Radiation Oncology Department of the Fourth Hospital of Hebei Medical University from January 2015 to February 2016 for RT with suspicious abnormal LN by US or simulation CT images were identified for this respective study.In total,29 abnormal LNs from 27 patients were brought for US(if not done before)and ARFI exams.The patient demographic and BC related measurement data were collected for analysis as well as the routine measures from US and ARFI.The pathological results of using fine needle aspiration(FNA)to abnormal LN were treated as the gold standard outcome for association analysis.To describe the data,mean and standard deviation(std)were used for the continuous variables;the percentage was used for the categorical variables including the derived ones.Chi-square test and univariate logistic regression were conducted in evaluating the associations of metastatic LN rates with other variables.At end,the multivariate logistic regression model analyses were undertook to asses the contributing value of ARFI measure(VTQ≥1.9m/s vs.VTQ<1.9m/s)in diagnosing the LN metastatic status.Odd ratio(OR)and its 95% confidence interval(CI)were calculated for presentation as well as its p values.Two-tailed P<0.05 was regarded as the significant level unless stated otherwise.Statistical tool used was SPSS 22.0.Results:1 Characteristics of study population:Twenty seven patients were enrolled in this study.They had the mean age(std)49.7(12.1)and range of 25-70 years old.The pre-menopause rate was 55.6%(15/27);2 Characteristics of US and ARFI measures:Twenty nine abnormal LNs were identified for ultrasound tests.Among them,2 LNs were from each of 2 patients.Among the US measures,the width/length of LN ratio(W/L ratio)of LNs had mean(std)of 0.57(0.20)and the rate of the poor-circumscribed of LNs was 41.4%(12/29).From ARFI,the continuous VTQ measures had mean(std)as of 2.50(1.48)m/s.The metastatic rate of LN was 34.5%(10/29);3 Univariate logistic regression analysis results.There were significant associations of age group(age ≥50 vs.<50)and boundary status of LN(poor-circumscribed vs.well-circumscribed)with LN metastatic rate.The corresponding ORs(P value)were 6.856(P=0.037)and 15.00(P=0.005),respectively.The statistically marginal associations with LN metastatic rate were found for the menopausal status(post-menopausal vs.pre-menopause)and hilar structure state(absence vs.presence)with ORs(P value)as of 5.506(P=0.056)and 8.100(P=0.069)respectively.Univariate logistic regression analysis had not showed the significant value of using VTQ value(≥1.90 m/s vs.<1.90 m/s)for LN assessment.Its OR(P value)was estimated as 1.864(P=0.511);4 Multivariate logistic regression analysis results: After the multiple failures of establishing sound multivariate logistic regression models with a list of potential covariates selected from univariate analysis screening and clinical perspective,two levels of multivariate logistic regression models were established and presented.At Level A model with two non-ultrasound measures(clinical T1-4 stage,lymphovascular invasion status)as the covariates only,the VTQ value(≥1.90 m/s vs.<1.90 m/s)was found to be insignificantly associated with metastatic LN rate(OR=1.785,P=0.619).At Level B model with two non-ultrasound measures(the same as above)plus two conventional US measures(binary boundary status of LN stated above,W/L ratio category ≥0.6 vs.<0.6)had not showed the significant role of the VTQ value(≥1.90 m/s vs.<1.90 m/s)either.Its OR(P value)was estimated as 1.083(P=0.957).However,both conventional US measures of binary boundary status of LN and W/L ratio binary category from Level B model indicated their associating ORs(P value)as 37.800(P =0.019)and 1.171(P = 0.896)with metastatic LN rate respectively.Larger sample size appeared to be required for more accurate evaluation of VTQ value further.Conclusion:1 The boundary status in ultrasound is an independent predictor of metastatic lymph nodes of breast cancer before radiotherapy.2 The value of additionally applying ultrasonic elastography(VTQ measure from ARFI technique)in assessing regional LN metastatic status of breast caner before radiotherapy was not proved yet.Larger sample size is required. |