Prediction of histological grading in ductal carcinoma in situ based on mammographic signs and clinical information using machine learning models - Takeaways - MDSpire

Prediction of histological grading in ductal carcinoma in situ based on mammographic signs and clinical information using machine learning models

  • By

  • Jianyu Wang

  • Shilu Zhao

  • Liuying Zhao

  • Furong Huang

  • Hao Wu

  • Da Pang

  • July 2, 2026

  • 0 min

Share

  • 1

    The study developed machine learning models to predict histological grading in ductal carcinoma in situ using mammographic and clinical data.

  • 2

    A total of 243 patients with confirmed DCIS were analyzed, categorized into non-high-grade and high-grade groups based on histological results.

  • 3

    Three machine learning models—XGBoost, logistic regression, and multinomial Naive Bayes—were evaluated for their predictive performance.

  • 4

    The multinomial Naive Bayes model showed the highest AUC value of 0.806, but no significant differences were found among the models.

  • 5

    The findings suggest potential for integrating mammographic features and clinical information to enhance DCIS grading prediction accuracy.

Original Source(s)

Prediction of histological grading in ductal carcinoma in situ based on mammographic signs and clinical information using machine learning models

  • by Jianyu Wang, Shilu Zhao, Liuying Zhao, Furong Huang, Hao Wu, Da Pang

  • July 2, 2026

Related Content

ASCO26: Jennifer Ligibel, MD, BWEL

Dana-Farber Cancer Institute's Dr. Jennifer Ligibel shared updated results from the BWEL study showing a weight loss intervention led to significantly better physical function, global physical and mental health, and symptom in patients with early breast cancer.

  • July 1, 2026

  • 1 min
Combined treatment using repurposed synthetic peptide desmopressin and bevacizumab as a potential antiangiogenic strategy in osteosarcoma

Incidental thyroid carcinoma in surgically treated multinodular goiter: a retrospective study