基于表观扩散系数图影像组学的乳腺癌新辅助化疗疗效预测

打开文本图片集

Prediction neoadjuvant chemotherapy for breast cancer based on ADC radiomicsChen Ti，Su Xuejuan， Wang Li， Liu Lina.*Department Imaging，the ， ，

【Abstract】 ObjectiveTo explore the appropriate radiomics modeling apparent difusion coeffcient （ADC） images to predict the efcacy neoadjuvant chemotherapy（NAC）for breast cancer.MethodsAretrospective study was conducted on 8O patients with pathologically confirmed breastcancer undergoing neoadjuvant chemotherapyfrom May 2022 to October 2O23 in the City.All patients underwent routine MRI examinations before chemotherapy and after neoadjuvant chemotherapy. predictive eficacy NAC treatment breast cancer beforeand after treatment was evaluated by four detection methods：traditional MRI imaging analysis， imaging biomarker random forest （RF）model，logistic regression （LR） model and support vector machine （SVM） model evaluation based on ADC maps.Compare the prediction performance four methods for breast cancer NAC ffcacy with pathological resultsas thegold standard.predictive performance the imaging biomarkermodel for the efficacy NAC treatment breast cancer before and after treatment was evaluated basedon the accuracy，sensitivity， and specificity the four detection methods. Results ① efficacy traditional MRI methods to predict the NAC treatment breast cancer before and after treatment： accuracy： O.71，sensitivity： O.67，specificity： 0.88. ② efficacy the imaging biomarkermodel to predicttheeficacy NAC treatmentbreastcancerbefore andafter treatment： When the RF model was trained，the test set：area under the curve （AUC）： O.924，accuracy：0.82，sensitivity：0.89，specificity： 0.91，andthetraningset：AUC：10，accuracy：0.99，sensitivity：0.99，specificity：0.99.WhentheLRmodelwas traied， thetestet： AUC： 0.827，accuracy：0.79，sensiiviy：0.82，specificity：0.87，andthetraningset： AUC： 0.884，accuracy：0.3， sensitivity：0.85，specificity：0.88.WhentheSVMmodelwas traned，thetestset：AUC： 0.779，accuracy：0.72，sesitiity： 0.73，specificity：0.75，and the training set： AUC： 0.892，accuracy：0.78，sensitivity：0.80，specificity：0.81.Conclusion ① Based on ADC radiomics is superior to MRI traditional predictive methods in predicting the eficacy neoadjuvant chemotherapy for breast cancer. ② RF model based on ADC radiomics is superior to the LR modeland SVM model. 【Kev words】Rreast nennlasms： Chemoradintheranv.adiuvant：Maonetic resonance imaoing：ADC man

乳腺癌是女性中最常见和第二致命的疾病，约占所有女性肿瘤疾病数量的 30% 。（剩余8900字）