Construction of a risk prediction model of postoperative recurrence of breast cancer based on multimodal ultrasound and MRI
Objective To construct a risk prediction nomogram model for postoperative recurrence of breast cancer based on multimodal ultrasound and MRI.Methods The clinical data of 330 patients with breast cancer who received surgical treatment in Yulin First Hospital from January 2018 to October 2020 were retrospectively analyzed.Among them,there were 3 males and 327 females;they were(60.59±11.53)years old.They were divided into a modeling group(231 cases)and a validation group(99 cases)according to the ratio of 7:3.The multimodal ultrasound and MRI detection results were recorded.The patients were followed up by telephone and clinic recheck until October 2023 or recurrence.The data were compared between the groups by the independent-sample t test and x2 test.The modeling group data were used to analyze the influencing factors of postoperative recurrence by the Cox regression analysis.A nomogram model was constructed for predicting postoperative recurrence risk based on multimodal ultrasound and MRI.The calibration curve and receiver operating characteristic curve(ROC)were used to validate the effectiveness of the model in predicting postoperative recurrence.Results There were no statistical differences in general information between the modeling group and the validation group(all P＞0.05).The follow-up time was 22-55 months,with a median of 41 months.The recurrence rate was 23.64％(78/330).The recurrence rate in the modeling group was 23.81％(55/231),and that in the validation group 23.23％(23/99).Stage Ⅲ-Ⅵ(HR=1.970,95％CI 1.228-3.159),undifferentiated/poorly differentiated(HR=2.044,95％CI1.116-3.746),triple negative breast cancer(HR=1.919,95％CI1.035-3.559),edge burr sign(HR=2.217,95％CI 1.214-4.046),hard ring sign(HR=2.252,95％CI 1.130-4.490),radial enhancement(HR=2.232,95％CI 1.017-4.899),edge blurring after enhancement(HR=2.109,95％CI 1.106-4.018),Eratio(HR=2.206,95％CI 1.187-4.097),time to peak(TtoPK)(HR=2.177,95％CI 1.193-3.974),and apparent diffusion coefficient(ADC)(HR=0.636,95％CI 0.417-0.969)were influential factors for postoperative recurrence of breast cancer(all P＜0.05).A nomogram model of postoperative recurrence of breast cancer was constructed based on multimodal ultrasound and MRI findings and parameters;the nomogram model predicted that the risk of postoperative recurrence in the modeling group and the validation group was basically consistent with the actual recurrence risk(P＞0.05).The sensitivity,specificity,and area under the curve(AUC)(9 5％CI)of the nomogram model for predicting postoperative recurrence of breast cancer in the modeling group were 96.36％,97.16％and 0.994(0.973-0.999),respectively;those of the prediction validation group were 95.65％,72.37％,and 0.906(0.831-0.956),respectively.Conclusions Edge burr sign,hard ring sign,radial enhancement,blurred edge after enhancement,Eratio,TtoPK,and ADC are influencing factors for postoperative recurrence of breast cancer;the nomogram model for predicting the risk of postoperative recurrence of breast cancer based on the above multimodal ultrasound and MRI manifestations and parameters is highly effective.
Breast cancerMultimodal ultrasoundMRIRecurrenceRisk predictionNomogram model
万方数据
维普
