Deep learning assisted contrast-enhanced CT–based diagnosis of cervical lymph node metastasis of oral cancer: a retrospective study of 1466 cases

Abstract

Objectives Lymph node (LN) metastasis is a common cause of recurrence in oral cancer; however, the accuracy of distinguishing positive and negative LNs is not ideal. Here, we aimed to develop a deep learning model that can identify, locate, and distinguish LNs in contrast-enhanced CT (CECT) images with a higher accuracy. Methods The preoperative CECT images and corresponding postoperative pathological diagnoses of 1466 patients with oral cancer from our hospital were retrospectively collected. In stage I, full-layer images (five common anatomical structures) were labeled; in stage II, negative and positive LNs were separately labeled. The stage I model was innovatively employed for stage II training to improve accuracy with the idea of transfer learning (TL). The Mask R-CNN instance segmentation framework was selected for model construction and training. The accuracy of the model was compared with that of human observers. Results A total of 5412 images and 5601 images were labeled in stage I and II, respectively. The stage I model achieved an excellent segmentation effect in the test set (AP50-0.7249). The positive LN accuracy of the stage II TL model was similar to that of the radiologist and much higher than that of the surgeons and students (0.7042 vs. 0.7647 (p = 0.243), 0.4216 (p < 0.001), and 0.3629 (p < 0.001)). The clinical accuracy of the model was highest (0.8509 vs. 0.8000, 0.5500, 0.4500, and 0.6658 of the Radiology Department). Conclusions The model was constructed using a deep neural network and had high accuracy in LN localization and metastasis discrimination, which could contribute to accurate diagnosis and customized treatment planning. Key Points • Lymph node metastasis is not well recognized with modern medical imaging tools. • Transfer learning can improve the accuracy of deep learning model prediction. • Deep learning can aid the accurate identification of lymph node metastasis.

Deep learning in oral cancer- a systematic review

“… Precision Recall F1-score AUC 76.7% 82.1% 79.3% 0.79 3 Warin K. et al, 2022 [ 43 ] Thailand Clinical oral images (OPMDs) 600 images (70:10:20 for training, validation, and test) Annotated by three oral and maxillofacial surgeons N/A - Faster R-CNN - YOLOv4 100 epochs Batch size = 32 Learning rate = 0.00001 2 of GPU, TitanXP 12GB, Nvidia Driver: 450.102 and CUDA: 11.0. Precision Recall F1-score AUC 79.7% 81.0% 80.3% 0.74 4 Warin K. et al, 2022 [ 46 ] Thailand Clinical oral images (OSCC, OPMDs) 980 images (70:10:20 for training, validation, and test) Annotated by three oral and maxillofacial surgeons N/A - Faster R-CNN - YOLOv5 - RetinaNet - CenterNet2 1882 epochs Batch size = 8, 128 Learning rate = 0.001 15,000 and 20,000 iterations Tesla P100, Nvidia driver: 460.32 and CUDA: 11.2 (Nvidia Corporation, CA, USA) Precision Recall F1-score AUC 98.0% 92.0% 89.0% 0.91 5 Xu X. et al, 2023 [ 60 ] China CT images (OSCC) 5412 images (60:30:10 for training, validation, and testing) Annotated by a radiologist N/A - Mask R-CNN 10, 50, 100 epochs NVIDIA V100 GPU AP50 72.5% AUC Area under the curve …”

Development and validation of convolutional neural network-based model to predict the risk of sentinel or non-sentinel lymph node metastasis in patients with breast cancer: a machine learning study

“…Deep learning-aided lymph node prediction has been enrolled in some solid tumours. 30 , 31 In the present study, we built a CNN model based on DCE-MRI images of primary lesions to predict SLN and NSLN metastasis successively. Firstly, for SLN prediction, the CNN model winded up with an AUC of 0.885 in the external test set 1.…”

Comparing the Diagnostic Accuracy of Ultrasonography, CT, MRI, and PET/CT in Cervical Lymph Node Metastasis of Oral Squamous Cell Carcinoma

“…Although not used in this study, MRI diffusion-weighted imaging and apparent diffusion coefficient may add new diagnostic criteria [42]. In recent years, research on diagnosing lymph nodes using artificial intelligence and radiomics has been reported, and this may be useful in diagnosing lymph node metastasis in the future [43,44].…”