One in six deaths worldwide is due to cancer, which makes it the second most common cause of mortality globally. Breast, lung, liver, prostate, and pancreatic cancers are the leading cancers in the world. Because of their aggressive nature and late discovery at advanced stages, lung and liver cancers are one of the leading causes of cancer-related deaths. Malignant cells are generally detected and classified by medical images and laboratory tests. These activities are time-consuming and require a large team. Developing a method is essential to detecting malignant states at the earliest possible stage. Deep-learning methods are one option for supporting doctors in cancer testing. In this paper, a predictive screening method based on deep learning is proposed to detect various cancers, such as lung and liver. Deep learning for cancer detection may enable more cancer types to be diagnosed in a quicker manner. This looks into deep learning algorithms to classify images with malignancy features. A median filter is initially applied as a pre-processing step to the acquired images of different cancers. Then, the thresholding technique is used to segment the obtained pre- processed im-ages for the purpose of detecting tumors. The GLCM is used for feature extraction. The Convolutional Neural Network classifier identifies a tumor and classifies it as benign or malignant. Lastly, a number of measures are calculated and contrasted with the outcomes of earlier methods, including PSNR, accuracy, specificity, sensitivity, and MSE.