Metastatic Lymph Node Detection on Ultrasound Images Using YOLOv7 in Patients with Head and Neck Squamous Cell Carcinoma

Simple Summary Cervical lymph node (LN) metastasis is a critical prognostic factor for patients with head and neck squamous cell carcinoma (HNSCC), rendering accurate diagnosis of LN metastasis crucial for improving patient outcomes. Our study aimed to develop deep learning models for metastatic LN detection using YOLOv7, the fastest single-stage object detection model, on B-mode and power Doppler (D-mode) ultrasonography in patients with HNSCC and investigate their utility in supporting the diagnosis by comparing their performance to that of highly experienced radiologists and less experienced residents. A total of 462 B- and D-mode ultrasound images were used to train, validate, and test the B- and D-mode models, respectively. The detection performances of the B- and D-mode models for metastatic LNs were higher than those of less experienced residents; the performance of the D-mode model was comparable to that of highly experienced radiologists, suggesting that YOLOv7-based models are useful for supporting the diagnosis.Abstract Ultrasonography is the preferred modality for detailed evaluation of enlarged lymph nodes (LNs) identified on computed tomography and/or magnetic resonance imaging, owing to its high spatial resolution. However, the diagnostic performance of ultrasonography depends on the examiner's expertise. To support the ultrasonographic diagnosis, we developed YOLOv7-based deep learning models for metastatic LN detection on ultrasonography and compared their detection performance with that of highly experienced radiologists and less experienced residents. We enrolled 462 B- and D-mode ultrasound images of 261 metastatic and 279 non-metastatic histopathologically confirmed LNs from 126 patients with head and neck squamous cell carcinoma. The YOLOv7-based B- and D-mode models were optimized using B- and D-mode training and validation images and their detection performance for metastatic LNs was evaluated using B- and D-mode testing images, respectively. The D-mode model's performance was comparable to that of radiologists and superior to that of residents' reading of D-mode images, whereas the B-mode model's performance was higher than that of residents but lower than that of radiologists on B-mode images. Thus, YOLOv7-based B- and D-mode models can assist less experienced residents in ultrasonographic diagnoses. The D-mode model could raise the diagnostic performance of residents to the same level as experienced radiologists.