BACKGROUND
The early diagnosis of various gastrointestinal (GI) diseases can lead to effective treatment and also reduce the risk of many life-threatening conditions. Unfortunately, various small GI lesions are undetectable during early-stage examination by medical experts. In previous studies, various deep learning (DL)-based computer-aided diagnosis (CAD) tools were used to make a significant contribution to the effective diagnosis and treatment of GI diseases. However, most of these methods were designed to detect a limited number of GI diseases such as polyps, tumors, or cancers in a specific part of the human GI tract.
OBJECTIVE
Therefore, the aim of this study was to develop a comprehensive CAD tool to assist medical experts in diagnosing various types of GI diseases.
METHODS
Our proposed framework is comprised of a deep learning-based classification network, which is followed by the retrieval method. In the first step, the classification network predicts the disease type for the current medical condition, and then the retrieval part shows the relevant cases (in terms of endoscopic images) from the previous database. In this way, past cases help the medical expert to validate the current prediction by the computer in a subjective way, which ultimately results in better diagnosis and treatment. In the case of a wrong prediction by the computer, the medical expert can check other relevant cases (i.e., second-, third-, or fourth-best matches) which may be more relevant than the first best match.
RESULTS
All the experiments were performed using two endoscopic datasets with a total of 52,471 frames and 37 different classes. The optimal performances obtained by our proposed method in terms of accuracy, F1 score, mean average precision (mAP), and mean average recall (mAR) were 96.19%, 96.99%, 98.18%, and 95.86%, respectively. The overall performance of our proposed diagnostic framework significantly outperforms state-of-the-art methods.
CONCLUSIONS
This study provides a comprehensive computer-aided diagnosis framework for identifying various types of GI diseases. The obtained results show the superiority of our proposed method over the various state-of-the-art methods and illustrate its potential for clinical diagnosis and treatment. In addition, our proposed network can also be applicable to other classifications domains in medical imaging, such as computed tomography (CT) scan, magnetic resonance imaging (MRI), and ultrasound sequences.