Through the research of this project, we expect to validate the clinical utility of the TuFEst-LN model in assessing axillary lymph node status in breast cancer patients. Specifically, we aim to prospectively validate its ability to identify pathologically node-negative patients among clinically and radiologically assessed cN0 patients undergoing upfront surgery ，and explore the predictive value of the TuFEst-LN model combined with preoperative MRI for ypN status assessment in initially node-positive patients following neoadjuvant therapy.
Histologically confirmed invasive breast cancer.
Clinical stage cT1-3, cN0, M0 disease at enrollment.
No clinically palpable suspicious metastatic axillary lymph nodes identified by physical examination.
No radiologically suspicious metastatic axillary lymph nodes identified by preoperative axillary imaging (at least ultrasound examination). Patients with suspicious lymph nodes must have negative cytological or histological findings confirmed by fine-needle aspiration or core needle biopsy.
Pregnancy or breastfeeding.
Prior surgical removal of the primary breast lesion before enrollment, resulting in inability to obtain the required preoperative blood samples.
Presence of confirmed distant metastasis.
Presence of supraclavicular, internal mammary, or other lymph node lesions that cannot be adequately assessed by planned surgery and pathological evaluation.
Breast cancer patients receiving care across the six participating centers who fulfill the eligibility criteria for either the cN0 cohort undergoing upfront surgery or the initially cN+ cohort after neoadjuvant systemic therapy.
Cohort
Prospective
No Intervention: Observational Cohort
Other interventionNo Intervention: Observational Cohort
Pathological axillary lymph node positivity rate among TuFEst-LN-negative patients
Defined as the proportion of patients with pathological axillary lymph node positivity (pN1mi or higher) among patients predicted as negative by the locked TuFEst-LN model in Cohort 1. FOR = FN/(TN+FN) = 1-NPV.
Negative Predictive Value (NPV) of the TuFEst-LN Model
The proportion of patients with negative pathological axillary lymph node status (pN0 or pN0(i+)) among patients classified as negative by the locked TuFEst-LN model.
Sensitivity and False Negative Rate (FNR) for Detecting Pathological Axillary Lymph Node Positivity
The ability of the locked TuFEst-LN model to identify patients with pathological axillary lymph node metastasis (pN1mi or higher). FNR is defined as 1-sensitivity.
Specificity of the TuFEst-LN Model
The proportion of patients with true pathological node-negative status (pN0 or pN0(i+)) who are correctly classified as negative by the locked TuFEst-LN model.
Positive Predictive Value (PPV), Overall Accuracy, and Discrimination Performance
Evaluation of the predictive performance of the locked TuFEst-LN model, including PPV, overall accuracy, area under the receiver operating characteristic curve (ROC-AUC), area under the precision-recall curve (PR-AUC), and likelihood ratios.
Predictive Performance of the TuFEst-LN Model for ypN Status
Using final surgical pathology as the reference standard, the predictive performance of the TuFEst-LN model for post-neoadjuvant pathological axillary lymph node status (ypN) will be evaluated, including area under the receiver operating characteristic curve (ROC-AUC), sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and calibration performance.
Predictive Performance of Preoperative MRI for ypN Status
Evaluation of the ability of preoperative magnetic resonance imaging (MRI) clinical and quantitative imaging features after neoadjuvant therapy to predict residual pathological axillary lymph node status (ypN0 versus ypN-positive).
Performance of Combined cfDNA Fragmentomics and MRI Model
Evaluation of the incremental value of integrating TuFEst-LN cfDNA fragmentomic features with MRI-based prediction models by comparing changes in ROC-AUC, Brier score, calibration performance, and decision curve analysis.
Changes in cfDNA Fragmentomic Features From T0 to T1
Exploration of the association between longitudinal changes in cfDNA fragmentomic scores or features before and after neoadjuvant therapy and pathological outcomes, including ypN status, breast pathological complete response (pCR), residual cancer burden (RCB), and residual tumor burden.
Click on the lead sponsor to view all of their trials.
Second Affiliated Hospital, Zhejiang University, School of Medicine
Lead sponsor
Shandong Cancer Hospital and Institute
Collaborator
Suzhou Municipal Hospital
Collaborator
Taizhou Hospital of Zhejiang Province affiliated to Wenzhou Medical University
Collaborator
First People's Hospital of Hangzhou
Collaborator
Ningbo Medical Center Lihuili Hospital
Collaborator
