Author(s)

FeiFei Shan¹, JunQi Sun^2*, Wei Zhou², ChengYuan Zheng²

Affiliation(s)

¹Department of Ultrasound, Affiliated Yuebei People's Hospital of Shan University, Shaoguan 512026, Guangdong, China.

²Department of Radiology, Affiliated Yuebei People's Hospital of Shan University, Shaoguan 512026, Guangdong, China.

Corresponding Author

JunQi Sun

ABSTRACT

Background: Accurately predicting pathological complete response (PCR) to neoadjuvant chemotherapy (NAC) in patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer is crucial for tailoring individualized treatment strategies. This study seeks to develop and validate a radiomics-based nomogram utilizing pretreatment ultrasound (US) data to predict PCR on an individual basis. Methods: In this retrospective analysis, patients diagnosed with HER2-positive breast cancer who underwent NAC were included and randomly assigned to either a training cohort (70%) or a validation cohort (30%) between January 2020 and December 2024. Radiomics features were extracted from the entire tumor volume as delineated on pretreatment US images. Feature selection was conducted using a combination of stability analysis, the Least Absolute Shrinkage and Selection Operator (LASSO)to establish a comprehensive radiomics signature (Rad-score). A predictive nomogram was then constructed by integrating the Rad-score with independent clinical predictors through multivariate logistic regression. The model's performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA). Results:The radiomics signature, consisting of 10 features derived from first-order statistics and texture matrices, exhibited significant predictive capability, achieving an area under the curve (AUC) of 0.705 in the validation cohort. The final nomogram, integrating the Rad-score, progesterone receptor (PR), demonstrated enhanced performance, with AUCs of 0.749 and 0.744 in the training and validation cohorts, respectively. Calibration curves and decision curve analyses confirmed the model's robust calibration and clinical applicability. Conclusion: The pretreatment ultrasonography-based radiomics nomogram offers a non-invasive and effective tool for the individualized prediction of pathological complete response (PCR) in HER2-positive breast cancer patients undergoing neoadjuvant chemotherapy (NAC), potentially aiding in the customization of therapeutic strategies. 

KEYWORDS

Breast cancer; HER2-Positive; Neoadjuvant chemotherapy; Ultrasonography; Radiomics; Machine learning

CITE THIS PAPER

FeiFei Shan, JunQi Sun, Wei Zhou, ChengYuan Zheng. Pretreatment ultrasound-based radiomics nomogram for predicting pathological complete response to neoadjuvant chemotherapy in HER2-positive breast cancer. Journal of Pharmaceutical and Medical Research. 2025, 7(3): 71-80. DOI: https://doi.org/10.61784/jpmr3059.

REFERENCES

[1] Wagle NS, Nogueira L, Devasia TP, et al. Cancer treatment and survivorship statistics. CA: A Cancer Journal for Clinicians, 2025, 75(4): 308-340.

[2] Bardia A, Jhaveri K, Im SA, et al. Datopotamab Deruxtecan Versus Chemotherapy in Previously Treated Inoperable/Metastatic Hormone Receptor-Positive Human Epidermal Growth Factor Receptor 2-Negative Breast Cancer: Primary Results From TROPION-Breast01. Journal of Clinical Oncology, 2025, 43(3): 285-296.

[3] Gillies R J, Kinahan P E, Hricak H, et al. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016, 278(2): 563-577.

[4] Lin Z, Zheng M, Li Z, et al. Development and validation of a delta ultrasomics model for predicting treatment response to neoadjuvant chemotherapy in breast cancer. Translational Cancer Research, 2025, 14(11): 7967-7979.

[5] Moore-Palhares D, Alberico D, Chan AW, et al. Quantitative ultrasound imaging for predicting response and guiding personalized neoadjuvant chemotherapy in breast cancer: randomized phase 2 clinical trial results. NPJ Precision Oncology, 2025, 9(1): 390.

[6] Liu J, Leng X, Yuan Z, et al. Predicting breast cancer response to neoadjuvant chemotherapy with ultrasound-based deep learning radiomics models -- dual-center study. BMC Cancer, 2025, 25(1): 1737.

[7] Wang M, Huang Z, Tian H, et al. Longitudinal Ultrasound Delta Radiomics for Early Stratified Prediction of Tumor Response to Neoadjuvant Chemotherapy in Breast Cancer. Academic Radiology, 2025, 32(12): 7119-7133.

[8] Peng Q, Ji Z, Xu N, et al. Prediction of neoadjuvant chemotherapy efficacy in patients with HER2-low breast cancer based on ultrasound radiomics. Cancer Imaging, 2025, 25(1): 112.

[9] Wei C, Jia Y, Gu Y, et al. Predictive Analysis of Neoadjuvant Chemotherapy Efficacy in Breast Cancer Using Multi-Region Ultrasound Imaging Features Combined With Pathological Parameters. Ultrasound in Medicine and Biology, 2025, 51(12): 2205-2216.

[10] Moore-Palhares D, Sannachi L, Chan AW, et al. Validation of a Quantitative Ultrasound Texture Analysis Model for Early Prediction of Neoadjuvant Chemotherapy Response in Breast Cancer: A Prospective Serial Imaging Study. Cancers (Basel), 2025, 17(15): 2594.

[11] Liu J, Xue X, Yan Y, et al. Prediction of breast cancer HER2 status changes based on ultrasound radiomics attention network. Computer Methods and Programs in Biomedicine, 2025, 271: 108987.

[12] Valizadeh P, Jannatdoust P, Moradi N, et al. Ultrasound-based machine learning models for predicting response to neoadjuvant chemotherapy in breast cancer: A meta-analysis. Clinical Imaging, 2025, 125: 110574.

[13] Chen M, Hong T, Wang Y, et al. Effect of a machine learning prediction model on the false-negative rate of sentinel lymph node biopsy for clinically node-positive breast cancer after neoadjuvant chemotherapy. Breast, 2025, 83: 104543.

[14] Tenghui W, Xinyi L, Ziyi S Y, et al. Combination of ultrasound-based radiomics and deep learning with clinical data to predict response in breast cancer patients treated with neoadjuvant chemotherapy. Frontiers in Oncology, 2025, 15: 1525285.

[15] Zhang H, Lang M, Shen H, et al. Machine learning-based fusion model for predicting HER2 expression in breast cancer by Sonazoid-enhanced ultrasound: a multicenter study. Frontiers of Medicine (Lausanne), 2025, 12: 1585823.

[16] Fan Y, Sun K, Xiao Y, et al. Deep learning predicts HER2 status in invasive breast cancer from multimodal ultrasound and MRI. Biomolecules and Biomedicine, 2025, 25(10): 2243-2251.

[17] Yao J, Zhou W, Jia X, et al. Machine learning prediction of pathological complete response to neoadjuvant chemotherapy with peritumoral breast tumor ultrasound radiomics: compare with intratumoral radiomics and clinicopathologic predictors. Breast Cancer Research and Treatment, 2025, 212(2): 325-336.

[18] Chan AW, Sannachi L, Moore-Palhares D, et al. Validation of Quantitative Ultrasound and Texture Derivative Analyses-Based Model for Upfront Prediction of Neoadjuvant Chemotherapy Response in Breast Cancer. Journal of Imaging, 2025, 11(4): 109.

[19] Feng X, Shi Y, Wu M, et al. Predicting the efficacy of neoadjuvant chemotherapy in breast cancer patients based on ultrasound longitudinal temporal depth network fusion model. Breast Cancer Research, 2025, 27(1): 30.

[20] Zhou P, Qian H, Zhu P, et al. Machine learning for predicting neoadjuvant chemotherapy effectiveness using ultrasound radiomics features and routine clinical data of patients with breast cancer. Frontiers in Oncology, 2025, 14: 1485681.

[21] Xie J, Wei J, Shi H, et al. A deep learning approach for early prediction of breast cancer neoadjuvant chemotherapy response on multistage bimodal ultrasound images. BMC Med Imaging, 2025, 25(1): 26.