Value of intratumoral and peritumoral radiomics models based on ultrasound images combined with conventional ultrasound parameters in predicting pathological complete response to neoadjuvant therapy for triple-negative breast cancer
10.3760/cma.j.cn131148-20250122-00037
- VernacularTitle:瘤内及瘤周超声影像组学联合常规超声参数预测三阴性乳腺癌新辅助治疗病理完全缓解的价值
- Author:
Shuangxiu TAN
1
;
Xinyan QIN
;
Wentao KONG
;
Qiaoliang CHEN
Author Information
1. 南京大学医学院附属鼓楼医院超声医学科，南京 210008
- Publication Type:Journal Article
- Keywords:
Ultrasonography;
Triple negative breast cancer;
Peritumor;
Radiomics;
Predictive model
- From:
Chinese Journal of Ultrasonography
2025;34(4):295-302
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the predictive value of conventional ultrasound，intratumoral and peritumoral radiomics models based on ultrasound images for the efficacy of neoadjuvant therapy in triple negative breast cancer（TNBC），and to construct a combined model.Methods:A total of 122 patients diagnosed with TNBC and admitted to Nanjing Drum Tower Hospital，Affiliated Hospital of Medical School，Nanjing University between April 2020 and December 2023 were retrospectively selected as study subjects. All patients underwent surgical resection after neoadjuvant therapy and were categorized into two groups：50 cases in the pathological complete remission（pCR）group and 72 cases in the non-pCR group according to surgical pathology. A comprehensive array of clinical data，along with conventional ultrasound imaging findings，was systematically collected from patients prior to treatment and at the conclusion of the second cycle of treatment. The region of interest（ROI）was delineated on the baseline two-dimensional gray-scale ultrasound image using 3D Slicer software on the maximum long-axis section of the lesion. The peri-tumor area was obtained by circularly expanding outward by 5 mm，and the imaging histological features were extracted separately. The dataset was then randomly partitioned into a training set and a validation set，with a ratio of 7∶3. The χ2/ t/Mann-Whitney U tests were used for intergroup comparison of general information. Maximum correlation minimum redundancy and least absolute shrinkage and selection operator regression were used to screen the optimal radiomics feature set，respectively. Variables that exhibited statistically significant differences between group comparisons were then employed to construct combined models，incorporating intratumor and peritumor ultrasonographic models. The predictive efficacy，accuracy，and clinical utility of the models were assessed using ROC curves，calibration curves，and decision curve analysis（DCA），respectively. Results:Subsequent between-group comparisons and multifactorial Logistic regression analysis identified blood flow（ OR=0.213，95% CI=0.062-0.735）and change rate of length diameter（ OR=1.091，95% CI=1.013-1.175）as independent risk factors for predicting pCR. A total of seven and eight radiomics features from each of the intratumoral and peritumoral regions were screened for the construction of intratumoral imaging histology score（RS）and peritumoral RS. The Nomogram model was constructed by combining the blood flow，change rate of length diameter，intratumoral RS，and peritumoral RS，and its AUC values in the training and validation sets were 0.884（95% CI=0.815-0.953）and 0.841（95% CI=0.683-0.940），respectively. The calibration curves demonstrated the Nomogram model's exceptional precision，with a C-index of 0.860 and 0.782 for the training and validation sets，respectively. The DCA revealed that the Nomogram model exhibited the optimal net clinical benefit. Conclusions:Conventional ultrasound，intratumoral and peritumoral radiomics models based on ultrasound images have been shown to possess satisfactory predictive value for the efficacy of neoadjuvant therapy in TNBC，thereby facilitating clinical decision-making.
