Development and Validation of a Clinical Prediction Model for Axillary Lymph Node Metastasis in Patients with an Early-stage Breast Cancer

Abstract

Background: Thailand has a high incidence of breast cancer, leading to the development of screening programs to detect the disease at an early stage. Surgery is the main treatment option during this stage. Advanced techniques, such as minimally invasive procedures like breast-conserving surgery, oncoplastic surgery, and sentinel lymph node biopsy, have been introduced to enhance breast cancer surgery, aiming to reduce complications and improve cosmetic outcomes. However, the shift towards favoring sentinel lymph node biopsy over axillary lymph node dissection presents challenges due to resource limitations in many hospitals, particularly the absence of expert pathologists and modern equipment. Therefore, there is a need for reliable tools or scoring systems to identify suitable candidates for this surgery. Objective: To develop and validate a clinical prediction model for predicting the probability of axillary lymph node metastases in patients with early-stage breast cancer. Methods: This study involved analyzing data from 132 early breast cancer patients at Ubonratchatani Cancer Hospital to develop a prediction scoring system. The system was created using a multivariable logistic regression model. Results: Tumor size, size of axillary nodes, and presence of lymphovascular invasion have been shown as significant factors in predicting the probability of axillary lymph node metastasis. The performance of our model was excellent as evidenced by an AuROC score of 0.88. Furthermore, the positive predictive values (PPVs) for low, moderate, and high risks were 2.30%, 30.80%, and 75.90% respectively. Conclusions: This scoring system assists surgeons in confidently selecting early-stage breast cancer patients for either sentinel lymph node biopsy or axillary node dissection, which is important in hospitals that lack expert pathologists.