A novel lightweight deep convolutional neural network for early detection of oral cancer

Abstract

Objectives To develop a lightweight deep convolutional neural network (CNN) for binary classification of oral lesions into benign and malignant or potentially malignant using standard real‐time clinical images. Methods A small deep CNN, that uses a pretrained EfficientNet‐B0 as a lightweight transfer learning model, was proposed. A data set of 716 clinical images was used to train and test the proposed model. Accuracy, specificity, sensitivity, receiver operating characteristics (ROC) and area under curve (AUC) were used to evaluate performance. Bootstrapping with 120 repetitions was used to calculate arithmetic means and 95% confidence intervals (CIs). Results The proposed CNN model achieved an accuracy of 85.0% (95% CI: 81.0%–90.0%), a specificity of 84.5% (95% CI: 78.9%–91.5%), a sensitivity of 86.7% (95% CI: 80.4%–93.3%) and an AUC of 0.928 (95% CI: 0.88–0.96). Conclusions Deep CNNs can be an effective method to build low‐budget embedded vision devices with limited computation power and memory capacity for diagnosis of oral cancer. Artificial intelligence (AI) can improve the quality and reach of oral cancer screening and early detection.

Machine-Learning Assisted Discrimination of Precancerous and Cancerous from Healthy Oral Tissue Based on Multispectral Autofluorescence Lifetime Imaging Endoscopy

“…Even the optimal ML models identified during the training stage provided only modest (<80%) levels of specificity ( Table 4 ). It is expected that with a more comprehensive training database and the adoption of more advanced ML models (e.g., deep learning methods) [ 42 , 43 , 44 ], it will be possible to enable automated discrimination of dysplastic and cancerous vs. healthy oral tissue with superior classification performance. Nevertheless, the classification results obtained in the independent maFLIM images used as testing set (ROC-AUC > 0.8, Figure 4 ) strongly support the potentials of an ML-enabled maFLIM-based strategy for automated and unbiased discrimination of dysplastic and cancerous vs. healthy oral tissue.…”

Utilizing Deep Machine Learning for Prognostication of Oral Squamous Cell Carcinoma—A Systematic Review

“…Based on the reported accuracy of the deep machine learning techniques in the included studies, it is evident that the deep machine learning technique can play a significant role toward the improved prognostication of oral cancer and guide clinicians in making informed decisions. The approach of using deep learning for prognostication can provide low-cost screening [19,36], smartphone-based solution [17,23], deep learning-based automatic prognostication [18,27,32], and early detection and prediction of outcomes [15,17,23,24,[37][38][39]42].…”

Clinically Guided Trainable Soft Attention for Early Detection of Oral Cancer

“…Related work reported values of 0.8500 and 0.8875 [10], 0.866 and 0.900 [7], 0.89 and 0.97 [14], and 0.867 and 0.845 [15], for sensitivity and specificity, respectively. Whilst we provide a comparative study of models in Table 2, currently, direct comparisons to related work can be difficult to make because their datasets and consequently their methodologies are designed to tackle different challenges.…”

“…A study [13] demonstrated that simpler CNN architectures are more suitable when fine-tuning on an oral lesion dataset of limited size, VGG-19 [11] performed the best for the classification of referral vs. non-referral. Shamim [14] used VGG-19 for benign vs. pre-cancerous classification and Jubair [15] used a lightweight CNN for benign vs. suspicious classifications; both restricted to tongue lesions only.…”