Key points are not available for this paper at this time.
Breast cancer is a major worldwide health concern, and patient outcomes are significantly impacted by early detection and accurate prognosis. In this study, we present an integrated machine learning method that uses CT scan images to predict the pathological full response in breast cancer patients. Our methodology leverages a wide range of machine learning models, such as VGG-16, Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), and K-Nearest Neighbours (KNN), to effectively extract image features, analyse sequential data, perform deep learning, and facilitate transfer learning. 1,344 CT scan pictures make up the carefully selected dataset; 80% are used for training and 20% are for testing. To extract useful information from the photos, one must use feature extraction techniques including gradient information, shape descriptors, intensity statistics, and texture characteristics. The investigation's findings show that there are considerable differences in the projected accuracy of the models; CNNs scored the best at 98.7%, followed by VGG-16 at 95.4%, KNN at 93.2%, and RNN at 91.2%. These results demonstrate the potential therapeutic utility of the approach and emphasise the necessity of choosing the best machine learning model in accordance with particular study goals and available datasets. Our study adds to the continuing effort to find more precise and tailored treatment plans for patients with breast cancer, which will ultimately improve patient outcomes and care in the rapidly evolving area of oncology.
Ahmed et al. (Fri,) studied this question.