A Nomogram for Predicting Occult Axillary Lymph Node Metastasis: Performance in US-diagnosed N0 Breast Cancer Patients

Abstract Background To develop a nomogram model combining gray-scale ultrasound and virtual touch tissue imaging quantification (VTIQ) characteristics to predict axillary lymph node metastasis (ALNM) in ultrasound (US)-diagnosed N0 breast cancer patients. Methods A total of 567 patients enrolled in this study from April 2017 to May 2022, including 395 patients in the primary cohort and 172 patients in the external validation cohort. These are patients who are undergoing upfront surgery (no neoadjuvant treatment). Their preoperative gray-scale ultrasound images and VTIQ parameters were collected and used to develop a nomogram with internal and external validation for the prediction of occult ALNM. Results Three gray-scale ultrasound characteristics (age, margin, and distance from the nipple) and one VTIQ parameter (E max ) were identified as independent risk factors in univariate and multivariate analyses. The nomogram showed an area under the curves of 0.843 and 0.869 in the training and external validation cohorts, respectively, indicating good calibration. Conclusions The nomogram model can predict occult ALNM with relatively high accuracy. It is expected to be a non-invasive, easy, quick, and affordable supplement to traditional axillary ultrasound (AUS), which can help to determine appropriate axillary treatment for US-diagnosed N0 breast cancer patients.