This tool can be used to help newly diagnosed breast cancer patients assess the likelihood that their breast cancer has spread to the sentinel lymph nodes (SLN). It integrates clinical features (age, tumor characteristics, staging), imaging measurements (ultrasound, mammography, MRI), and pathology findings (tumor type, grade, receptor status) to generate a predicted probability. All fields marked with * are required.
Clinical Information
Imaging
Pathological Information
Select Negative if estrogen receptors stain positive in <1% of cells; select Positive if ≥1% of cells.
Select Negative if progesterone receptors stain positive in <1% of cells; select Positive if ≥1% of cells.

Paste Clinical Text to Auto-Fill

Paste a clinical report, pathology summary, or any text containing patient data below. The system will extract matching fields and fill the form automatically.
Supported formats: labeled text, tabular/spreadsheet rows, or mixed formats.
Recognition tips: Supports both labeled text and tab-separated spreadsheet rows. Unique keywords auto-matched: Others → Pathology, II/III → Grade, Invisible/Visible → LVI, >=20%/<20% → Ki67, standalone +/- → ER → PR → HER2. T1–T4 → cT Stage, Negative/Suspicious → Axillary US. Tumor size values (US Length, US Width, MG, MRI) are filled sequentially in order.

Tip: Works with most clinical text formats

Predicted Probability of SLN Metastasis

—
0 Low Risk 0.5 High Risk 1
—

About This Prediction

This tool helps clinicians and patients to accurately predict the likelihood that a patient's breast cancer has spread to the sentinel lymph nodes. The model was developed using a multimodal fusion deep learning architecture that integrates 35 clinical features and 20 pathological features extracted from a cohort of breast cancer patients. The calculated estimates include the risk of any tumor cells being found in the sentinel lymph nodes, including isolated tumor cells detected only on immunohistochemistry. Predictions are generated by processing clinical data (age, breast side, tumor count and location, cT stage, axillary ultrasound findings, and Chen-Dai distance), imaging measurements (ultrasound, mammography, and MRI tumor dimensions, calcification status, breast density), and pathological information (tumor type, histological grade, ER, PR, HER2, LVI, and Ki67 status) through the trained neural network. This tool is intended for academic and research use only and should not be used as a substitute for clinical judgment.

User Guide