Deep learning–based non-invasive prediction of axillary lymph node metastasis in breast cancer: performance of the YOLO-v11 object detection algorithm

This study proposes a deep learning-based decision support system for the non-invasive prediction of axillary lymph node (ALN) metastasis in breast cancer using ultrasound (USG). Both localisation and benign–malignant classification of ALNs were performed using YOLO-v11, a single-stage object detection approach. The dataset created from labelled USG images was split into training/validation/test sets at a ratio of 70%/20%/10%, and the model produced a class label and confidence score for each prediction. Performance was found to be high with [email protected] = 0.904s (benign: 0.866, malignant: 0.942); F1-, precision-, and sensitivity-confidence curves showed that thresholds in the 0.52–0.60 range provided a balance between sensitivity and specificity. Confidence scores for the malignant class were observed to be concentrated in the 0.82–0.87 band. The confusion matrix includes an auxiliary ‘background’ row/column representing unmatched predictions (false positives) and missed ground-truth detections (false negatives); background was not modelled as a semantic class during training. The findings point to a rapid and reproducible artificial intelligence approach that could reduce operator dependency in USG. The proposed model has the potential to provide radiologists and clinicians with non-invasive, real-time decision support in ALN assessment. Axillary lymph node (ALN) metastasis plays a decisive role in the clinical course and treatment planning of breast cancer following diagnosis. Therefore, the accurate and early detection of metastatic lymph nodes directly impacts treatment success. Ultrasonography (USG) stands out as the most common and effective imaging method for evaluating axillary lymph node metastasis. The main objective of this study is to develop an artificial intelligence-supported USG analysis system to provide a decision support system for radiologists and clinicians in the detection of ALN metastasis. This study included a total of 471 patients: 248 patients diagnosed with breast cancer between 2024 and 2025 and reported as malignant based on axillary lymph node biopsy results, and 223 control cases with benign lymph nodes detected in breast ultrasound scans. Using a dataset created from the acquired ultrasound images, a single-stage object detection model (YOLO-v11) was employed to detect ALN metastasis. The data were split into a 70% training, 20% validation, and 10% testing ratio. The developed model produces a confidence score and class label for each detection. The developed model has demonstrated high accuracy performance in benign and malignant lymph node samples. The [email protected] value calculated under the Precision–Recall (PR) curve was obtained as 0.904 (benign: 0.866, malignant: 0.942). The precision and sensitivity values were found to be 0.866 in benign samples and 0.942 in malignant samples, respectively. Confidence score values for the malignant class were concentrated in the range of 0.82–0.87. When evaluated from a clinical application perspective, the threshold value in the range of 0.5–0.6 provides a balanced trade-off between sensitivity and specificity. The proposed artificial intelligence model can localise axillary lymph nodes with high accuracy on USG images and distinguish between benign and malignant conditions. In this respect, the model has the potential to guide radiologists and clinicians as a decision support tool in the diagnostic process.