Biological therapy-oriented CBMIR for breast cancer detection via IChOA-CNN-LSTM approach

<p class="MsoNormal" style="line-height: 120%; layout-grid-mode: char; mso-layout-grid-align: none; punctuation-wrap: simple; margin: 12.0pt 0cm 6.0pt 0cm;"><strong style="mso-bidi-font-weight: normal;"><span lang="EN-US" style="font-size: 10.0pt; mso-bidi-font-size: 11.0pt; line-height: 120%; font-family: 'Times New Roman',serif;">Background:</span></strong><span lang="EN-US" style="font-size: 10.0pt; mso-bidi-font-size: 11.0pt; line-height: 120%; font-family: 'Times New Roman',serif;"> Breast cancer is one of the world’s most serious health issues, and early and correct detection is vital for increasing survival rates. Biological therapies, sometimes referred to as immunotherapies or targeted therapies, are used to treat breast cancer in order to control hormone pathways, target certain cancer cells, or strengthen the immune system. These therapies seek to reduce injury to healthy cells when compared to standard treatments such as chemotherapy, potentially leading to fewer side effects. <strong style="mso-bidi-font-weight: normal;">Methods:</strong><span style="mso-bidi-font-weight: bold;"> </span>This research described a novel deep learning-based Content-Based Medical Image Retrieval (CBMIR) method for detecting breast cancer using histological images. It begins with biological regulator BC images, which are input histopathological images of breast tissue. The major input is the BreakHis dataset, with bilateral filtering used as a preprocessing step to decrease noise while retaining important tissue properties. Feature extraction uses the Gray-Level Co-occurrence Matrix (GLCM) and Histogram of Oriented Gradients (HOG), which allow for the effective capture of both textural and spatial information. The Improved Chimp Optimization Algorithm (IChOA) and a cascaded Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) architecture are then coupled to create a hybrid classification model that enhances learning efficiency while also predicting temporal correlations in picture input. To overcome this issue, the proposed IChOA-CNN-LSTM framework employs CNNs for precise image feature extraction, LSTM networks for sequential data analysis, and an IChOA for effective feature fusion. <strong style="mso-bidi-font-weight: normal;">Results:</strong><span style="mso-bidi-font-weight: bold;"> </span>The suggested CBMIR system performed well in both picture classification and retrieval tasks. The system attained an amazing classification accuracy of 97.5%, demonstrating its ability to considerably minimize diagnostic mistakes and processing time in histopathology image analysis.<strong style="mso-bidi-font-weight: normal;"> Conclusion:</strong> The method connects with tailored biological therapy options, including HER2-targeted antibodies and small-molecule inhibitors, by allowing for more reliable early detection of key tumor features. Integrating CBMIR into diagnostic procedures could thus serve as an effective tool for identifying and optimizing tailored therapeutic interventions, thereby boosting precision oncology and patient outcomes.</span></p>