Dynamic Weight Agnostic Neural Networks and Medical Microwave Radiometry (MWR) for Breast Cancer Diagnostics

Abstract

Background and Objective: Medical microwave radiometry (MWR) is used to capture the thermal properties of internal tissues and has usages in breast cancer detection. Our goal in this paper is to improve classification performance and investigate automated neural architecture search methods. Methods: We investigated extending the weight agnostic neural network by optimizing the weights using the bi-population covariance matrix adaptation evolution strategy (BIPOP-CMA-ES) once the topology was found. We evaluated and compared the model based on the F1 score, accuracy, precision, recall, and the number of connections. Results: The experiments were conducted on a dataset of 4912 patients, classified as low or high risk for breast cancer. The weight agnostic BIPOP-CMA-ES model achieved the best average performance. It obtained an F1-score of 0.933, accuracy of 0.932, precision of 0.929, recall of 0.942, and 163 connections. Conclusions: The results of the model are an indication of the promising potential of MWR utilizing a neural network-based diagnostic tool for cancer detection. By separating the tasks of topology search and weight training, we can improve the overall performance.

Diagnostics of Ovarian Tumors in Postmenopausal Patients

“…The device visualizes electromagnetic radiation in the range of 3400–4200 MHz. The antenna diameter is 39 mm and provides a measurement depth of up to 40–50 mm [ 54 , 55 , 56 , 57 ]. Non-invasive measurements are conducted by registering electromagnetic radiation through the anterior abdominal wall (transabdominal) in the projection of the uterine appendages at nine symmetrical points on both sides.…”

Passive Microwave Radiometry and microRNA Detection for Breast Cancer Diagnostics

“…More recently, artificial intelligencebased approaches have been applied to MWR data for early stage breast cancer prediction to approve diagnostic accuracy. For example, in a study on MWR data from >4000 patients who had been classified by clinicians as either being at low or high risk of developing breast cancer, deep neural networks achieved with an accuracy of breast cancer prediction >0.93 [10,14,15].…”

“…If either 𝑄 𝑚𝑎𝑥 or R exceeded these thresholds, both malignant and benign tumors were possible. Simultaneously with the described method, to determine the risk group according to microwave radiometry, artificial intelligence methods were used, in particular, the weight agnostic neural network, configured by the bi-population covariance matrix adaptation evolution strategy method [10]. The main feature of this neural network is that at the stage of searching for the optimal network architecture for the task, the average weights of the model are used.…”

“…The average internal and skin temperatures of the breast tissue were calculated. The difference between the maximum (7), (8) and average skin and internal temperatures for each gland (9), (10) were also calculated. The obtained differences were summarized for the right and left breasts:…”

Emerging paradigms in microwave imaging technology for biomedical applications: unleashing the power of artificial intelligence

“…Li et al 133 explored new ways to automate the search for optimal neural architecture to boost classification performance. After determining the network architecture, the author explored improving the weights of the weight-agnostic neural network with the bi-population covariance matrix adaptation evolution strategy (BIPOP-CMA-ES).…”