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.

Keywords: Deep learning; Diagnosis, computer-assisted; Lymphatic metastasis; Mouth neoplasms; Tomography, X-ray computed.

Fig. 1

Process of patient data collection, screening, classification, and data labeling. Flowchart of the study population and assignments for date labeling and machine learning for construction of prediction models. LN, lymph node; LN+, positive lymph node; LN−, negative lymph node

Fig. 2

The process of data labeling. The process of data labeling for the input CECT images of stage I and stage II using our data-labeling tools. The dataset image used for training was generated from labeled data

Fig. 3

Prediction process and PR curves. a Prediction process using the stage I model. Predicted images I and II were generated from the input image with confidence levels of 0.75 and 0.85, respectively. The dataset image was labeled for comparison. b PR curves at different IoU threshold values in stage I. c Prediction process using the stage II model. Predicted image I was generated from the input image in the stage II model. Predicted image II was generated from the input image in the stage II-TL model. The dataset image was labeled for comparison. d PR curves of the model before and after the stage I model was utilized for training of transfer learning at the default IoU threshold value (0.50)

Fig. 4

Model evaluation and prediction results. a LN+ accuracy of the stage II-TL model at different levels at different confidence intervals. b Comparison results of LN+ accuracy and clinical accuracy among the stage II-TL model-0.85, radiologist, surgeons, and students. TL, transfer learning; LN+, positive lymph node