The goal of this observational study is to develop and validate an artificial intelligence (AI) model for predicting the risk of distant metastasis in patients with primary breast cancer. The main question it aims to answer is: Can a multimodal AI model, trained on routinely available histopathological images, accurately predict the long-term risk of breast cancer metastasis? Researchers will analyze existing hematoxylin and eosin (H\&E) and immunohistochemistry (IHC) stained tissue slides from patients who underwent surgery between 2015 and 2025. Clinical data will be used to train the AI model and evaluate its performance in predicting metastasis.
Study Type
OBSERVATIONAL
Enrollment
400
This is an observational study with no therapeutic or procedural interventions. The "intervention" refers to the analytical method applied to existing data. Archived tissue samples (H\&E and IHC stained slides) will be digitally scanned and analyzed by a multimodal artificial intelligence (AI) model to develop a risk prediction tool for distant metastasis. Patients' clinical data will be collected for model training and validation. No direct interaction with patients occurs, and no treatment decisions are influenced by this study.
Jilin Cancer Hospital
Changchun, Jilin, China
NOT_YET_RECRUITINGCancer Institute and Hospital, Tianjin Medical University, China
Tianjin, Tianjin Municipality, China
NOT_YET_RECRUITING2nd Affiliated Hospital, School of Medicine, Zhejiang University, China
Hangzhou, Zhejiang, China
RECRUITINGThe Fourth Affiliated Hospital of Zhejiang University School of Medicine
Hangzhou, Zhejiang, China
NOT_YET_RECRUITINGPredictive accuracy for distant metastasis risk assessed by Time-dependent Area Under the Receiver Operating Characteristic Curve (Time-dependent AUC)
The Area Under the Receiver Operating Characteristic Curve (AUC) will be used to evaluate the model's binary classification performance in discriminating between patients with and without distant metastasis at the 5-year post-operative time point. This metric reflects the model's classification accuracy at a specific time.
Time frame: From the date of initial surgery up to 5 years post-operatively, with the occurrence of distant metastasis defined as the event of interest.
Sensitivity and Specificity
Sensitivity and Specificity will be calculated at the optimal cut-off point of the model's risk score to evaluate its binary classification performance. Sensitivity measures the model's ability to correctly identify patients who develop distant metastasis (true positive rate), while Specificity measures its ability to correctly identify patients who do not (true negative rate).
Time frame: Assessed at the 5-year post-operative time point.
Concordance Index (C-index)
Harrell's Concordance Index (C-index) will be employed to assess the model's overall prognostic discrimination ability throughout the follow-up period. It evaluates the consistency of the model's risk scores in correctly ranking the time to distant metastasis-free survival among individual patients.
Time frame: From the time of the initial surgical treatment until distant metastasis occurs or until the end of the follow-up (the longest duration can be up to 10 years).
Model calibration assessed by calibration curve
The agreement between the model-predicted probability of distant metastasis and the observed actual incidence will be visualized and assessed using a calibration curve.
Time frame: From the time of the initial surgical treatment until distant metastasis occurs or until the end of the follow-up (the longest duration can be up to 10 years).
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.