分子影像学杂志2025，Vol.48Issue(12)：1464-1474,11.DOI:10.12122/j.issn.1674-4500.2025.12.02
基于多模态超声和血清炎性指标构建列线图预测浸润性乳腺癌组织学分级
Development of a nomogram based on multimodal ultrasound and serum inflammatory indicators for predicting histological grading of invasive breast cancer
摘要
Abstract
Objective To investigate the predictive value of a nomogram combining multimodal ultrasound features and serum inflammatory markers for histological grading of invasive breast cancer.Methods Multimodal ultrasound images and immune inflammatory markers were retrospectively analyzed from 174 patients with histologically confirmed invasive breast cancer admitted to Inner Mongolia Autonomous Region People's Hospital from January 2022 to June 2025.Data were randomly divided into a training set(n=121)and a validation set(n=53)at a 7:3 ratio.Independent predictors associated with breast cancer histological grade were identified via multivariate logistic regression.A regression model was established to generate a nomogram,which was validated using the validation cohort and calibration curves.Predictive performance was assessed using ROC curves and decision curve analysis(DCA).Clinical utility was evaluated through clinical impact curves(CIC).Results Multivariate logistic regression analysis revealed perfusion defects(OR=3.743,95%CI:1.342-10.439,P=0.012),AUC(OR=1.002,95%CI:1.000～1.003,P=0.043),Emax(OR=1.021,95%CI:1.003-1.040,P=0.023),and LMR(OR=0.721,95%CI:0.572-0.909,P=0.006)were predictive factors for breast cancer histological grade.A nomogram prediction model was constructed based on these four indicators.The training set AUC was 0.856(95%CI:0.790-0.922),with a C-index of 0.856.Calibration curves demonstrated good agreement between predicted and actual probabilities.The Hosmer-Lemeshow test showed no statistically significant differences(P=0.231),indicating good model fit.The DCA curve indicated that intervention measures yielded high clinical net benefit when the probability threshold ranged from 0%to 87%.The CIC curve demonstrated that when the probability threshold exceeded 70%,the number of breast cancer patients predicted by the nomogram to have a certain histological grade highly matched the actual patient count.The validation set achieved an AUC of 0.846(95%CI:0.739-0.954)and a C-index of 0.846.The calibration curve demonstrated good agreement between the nomogram's predictions and actual outcomes.Both the DCA and CIC curves indicated the model possesses significant clinical utility.Conclusion The nomogram model for predicting breast cancer histological grade,constructed based on multimodal ultrasound features combined with serum inflammatory markers,provides important reference information for clinical diagnosis,subsequent treatment,and prognosis assessment,demonstrating significant clinical application value.关键词
乳腺癌/浸润性乳腺癌/多模态超声/血清炎性指标/列线图Key words
breast cancer/invasive ductal carcinoma/multimodal ultrasound/serum inflammatory markers/nomogram引用本文复制引用
刘萌儒,刘强,王玉敏..基于多模态超声和血清炎性指标构建列线图预测浸润性乳腺癌组织学分级[J].分子影像学杂志,2025,48(12):1464-1474,11.基金项目
内蒙古自治区科技计划项目(2021GG0125) （2021GG0125）
内蒙古医学科学院公立医院科研联合基金科技项目(2024GLLH0078) （2024GLLH0078）
内蒙古医科大学百万工程联合项目[KD2020KJBW(LH)044] （LH）
