Applying CNNs for the Early Detection of Bone Malignant Tumors in X-Ray Images

Bone malignant tumors are cancerous growths that develop in the bones and can be fatal if they remain undiagnosed or untreated for a long time. Since these tumors often grow slowly, early detection is important for successful treatment and high survival rates. Existing medical imaging tools to detect bone malignant tumors, such as X-rays, utilize human expertise and are time-consuming. Thus, recently, computer-aided diagnosis systems have been developed primarily with Convolutional Neural Networks (CNNs). Convolutional neural networks are deep learning models proven to perform highly on pattern recognition tasks, especially in image and video analysis. In this paper, we suggest that CNNs be used in the early finding of bone malignant tumors using X-ray images. It starts by preprocessing the X-ray images before entering them into CNN to perform feature extraction. The resampling data was then used to teach the CNN where one class is “normal,” and another class has an abnormal, namely a bone malignant tumor. The system's performance was tested on a dataset consisting of X-ray images obtained from patients with bone malignant tumors. Results showed high detection accuracy for both these tumors with sensitivity and specificity of < 90% (Table 2). This work demonstrates the potential value of CNNs for early bone malignant tumor detection, and therefore, their use in clinical care will help health professionals make prompt and correct diagnoses.