Accurate assessment of axillary lymph nodes in patients with breast cancer is essential for prognosis and treatment planning. Staging and surgical management have evolved from axillary lymph node dissection to sentinel lymph node biopsy to minimize morbidity. However, sentinel lymph node biopsy has non-negligible morbidity, and more than 70% of biopsies are negative, calling into question its routine use. Magnetic resonance imaging (MRI) can be used to detect and stage lymph node metastases in situ, but its sensitivity and specificity are moderate to poor. Few studies have employed artificial intelligence to detect lymph node metastases on MRI images, and none have used an integrative multidata approach (IMA), defined as modeling the combination of clinical and laboratory data with multiparametric MRI.
The primary objective of this retrospective observational study is to improve the accuracy of detecting lymph node involvement in breast cancer using IMA. The secondary objective is to allow longitudinal monitoring of the effects of neoadjuvant therapy on lymph node involvement
Patients diagnosed with breast cancer who underwent MRI and biopsy or surgery of axillary lymph nodes as part of their diagnosis and treatment
Patients of all genders, ages, and stage I-III of breast cancer
Patients who underwent neoadjuvant therapy and have longitudinal imaging data available for analysis (for secondary outcome analysis)
Patients whose MRI images were of insufficient quality for analysis
Patients who had a previous history of breast cancer
Patients with a history of axillary surgery or lymph node dissection prior to the current diagnosis of breast cancer
Patients who received neoadjuvant therapy at another institution
The study population will consist of patients with breast cancer who underwent MRI and biopsy or surgery of axillary lymph nodes as part of their diagnosis and treatment at the involved centers. Patients of all genders, ages, and stages I-III of breast cancer will be included in the study. Patients with incomplete data or whose MRI images were of insufficient quality for analysis will be excluded.
Cohort
Retrospective
Not listed
The accuracy of detecting lymph node involvement using multiparametric MRI, clinical characteristics, and laboratory data, and to compare it to the accuracy of detecting lymph node involvement using MRI alone
Sensitivity, specificity, positive predictive value, and negative predictive value of the integrative model will be calculated to evaluate the diagnostic performance of the model.
Evaluate the changes in lymph node involvement over time
The effects of neoadjuvant therapy will be estimated as the difference in the rate of lymph node involvement before and after neoadjuvant therapy, with corresponding 95% confidence intervals.
Evaluate the association between changes in lymph node involvement and overall survival
Association between changes in lymph node involvement and overall survival (OS) will be reported as Hazard Ratio and relative 95% Confidence Interval (95% CI) OS will be calculated from neoajuvant treatment start to death or end of follow up whichever came first
Evaluate the association between changes in lymph node involvement and progression free survival (PFS)
Association between changes in lymph node involvement and overall survival (PFS) will be reported as Hazard Ratio and relative 95% Confidence Interval (95% CI) PFS will be calculated from neoajuvant treatment start to progression, death or end of follow up whichever came first
Click on the lead sponsor to view all of their trials.
